United States Hazardous Waste Engineering EPA/600/9-86/022
Environmental Protection Research Laboratory August 1986
Agency Cincinnati OH 45268
Research and Development
<&EPA Land Disposal,
Remedial Action,
Incineration and
Treatment of
Hazardous Waste
Proceedings of the
Twelfth Annual
Research Symposium
-------
-------
EPA/600/9-86/022
August 1986
LAND DISPOSAL, REMEDIAL ACTION, INCINERATION
AND TREATMENT OF HAZARDOUS WASTE
Proceedings of the Twelfth Annual Research Symposium
at Cincinnati, Ohio, April 21-23, 1986
Sponsored by the U.S. EPA, Office of Research & Development
Hazardous Waste Engineering Research Laboratory
Cincinnati, OH 45268
Edison, NJ 08837
Coordinated by:
JACA Corp.
Fort Washington, PA 19034
Contract No. 68-03-3252
Project Officers:
Harry M. Freeman
Naomi P. Barkley
Cincinnati, OH 45268
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
DISCLAIMER
These Proceedings have been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for presenta-
tion and publication. Mention of trade names or commercial products does not consti-
tute endorsement or recommendation for use.
-------
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 our 1986 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-
-------
PREFACE
These Proceedings are intended to disseminate up-to-date information on
extramural research projects concerning land disposal, remedial action, incineration
and treatment of hazardous waste. These projects are funded by the U.S. Environmental
Protection Agency's Office of Research and Development and have been reviewed in
accordance with the requirements of EPA's Peer and Administrative Review Control
System.
-------
ABSTRACT
The Twelfth Annual Research Symposium on Land Disposal, Remedial Action,
Incineration and Treatment of Hazardous Waste was held in Cincinnati, Ohio, April 21
through April 23, 1986. 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 management.
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 landfill design and operation, waste leaching and
analyses, pollution migration and control, waste modification, surface impoundments,
flexible membrane liners, remedial action techniques, and undergrond mine disposal.
Incineration and treatment subjects include combustion of hazardous wastes 1n
incinerators, boilers and industrial processes; field evaluations, treatment options
and innovative processes for hazardous waste destruction.
-v-
-------
CONTENTS
SESSION A - HAZARDOUS WASTE LAND DISPOSAL PAPERS
Technical Resource Documents and Technical Handbooks for
Hazardous Wastes Management
Norbert Schomaker, U.S. Environmental Protection Agency
Leaking Underground Storage Tank Research Program
Richard Field, U.S. Environmental Protection Agency 11
Hazardous Materials Releases Research Program
Anthony Tafuri, U.S. Environmental Protection Agency 15
Effective Porosity of Geologic Materials
Gary Peyton, Illinois State Water Survey 21
Hydraulic Conductivity of Compacted Clay Soils
Andrew Rogowski, USDA-ARS 29
Geotechnical Analysis for Review of Dike Stability
Mark Meyers, University of Cincinnati ...... 40
The Role of Water Balance in the Long-Term Stability of Hazardous
Waste Site Cover Treatments
Fairley Barnes, Los Alamos National Laboratory 50
Factors Influencing Differential Settlement and the Effects on
Cover Systems
Paul Gilbert, USAE Waterways Experiment Station ... 59
Applicability of the HELP Model in Multilayer Cover Design;
A Field Verification and Modeling Assessment
Nathaniel Peters, University of Kentucky. .. 65
Remediation of an Industrial Dump Site - A Case History
Robert Ahlert, Rutgers University .... 73
Reportable Quantities Guidelines for CERCLA
Richard Field, U.S. Environmental Protection Agency . 90
Predicting Hazardous Waste Leachate Composition
Danny Jackson, Battelle Columbus Laboratory 99
-VI-
-------
Page
SESSION A (Continued)
Chemical Resistance of Synthetic Liner/Hazardous Waste Combinations
Gordon Bellen, National Sanitation Foundation .......... 101
Verification of the Hydrologic Evaluation of Landfill Performance
(HELP) Model Using Field Data
Lee Peyton, University of Missouri 114
Interfacial Stability of Soil Covers on Lined Surface Impoundments
Donald Mitchell, Pacific Northwest Laboratory . . 124
Determination of the Solubility Parameters of FMLs for Use in Assessing
Resistance to Organics
Henry Haxo, Jr., Matrecon Inc.. . . . 132
In-Situ Biological Treatment of Contaminated Groundwater and Soils
Roger Wetzel, Science Applications International Corp. ..... 146
Stringellow Leachate Treatment with Rotating Biological Contactor
Edward Opatken, U.S. Environmental Protection Agency 154
Nondestructive Testing Location of Containers Buried in Soil
Arthur Lord, Jr., Drexel University ..... 161
A Selection Guide for Mobile/Portable Treatment Technologies for the
Removal of Volatile Organics from Water
Jeffrey Fleming, IT Corp 170
A Guidance Manual for the Selection and Use of Sorbents for Liquid
Hazardous Substance Releases
Steven Gibson, Environmental Monitoring & Services, Inc 177
In Situ Treatment of Sodium Hydroxide Spills in Quiescent Water Bodies
Michael Borst, Mason & Hanger-Silas Mason Co., Inc. ....... 185
Chemical and Microbial Stabilization Techniques for Remedial Action
Sites
Janet Rickabaugh, University of Cincinnati. . .......... 193
Treatment of Soils Contaminated with Heavy Metals
William Ellis, Science Applications International Corp. ..... 201
-V11-
-------
Page
SESSION A (Continued)
Field Demonstration of In Situ Washing of Contaminated Soils with
Water/Surfactants
James Nash, Mason & Hanger-Silas Mason Co. Inc. 208
Effect of Freezing on the Level of Contaminants in Uncontrolled
Hazardous Waste Sites - Part II: Preliminary Results
Iskandar Iskandar, U.S. Army Cold Regions
Research & Engineering Laboratory . 218
Feasibility of Using Mined Space for Long-Term Control of Dioxin-
Contaminated Soils
M. Pat Esposito, PEI Associates Inc 226
Evaluation of Cutoff Wall Containment Efficiency Using Aquifer
Stress Analyses - Gil son Road NPL Site
Matthew J. Barvenik, GZA, Inc 491
SESSION B - HAZARDOUS WASTE INCINERATION AND TREATMENT PAPERS
Review of Alternative Treatment Processes for Non-Halogenated Solvent
Wastes
Benjamin Blaney, U.S. Environmental Protection Agency 235
Preliminary Assessment of Aerated Waste Treatment Systems at TSDFs
David Green, Research Triangle Institute 246
An Evaluation of Batch Fractional Distillation for Removing Orgam'cs
from Mixed Aqueous/Organic Wastes
C. Clark Allen, Research Triangle Institute ..... 253
Review of Alternative Treatment Processes for Halogenated Organic
Waste Streams
Ronald Turner, U.S. Environmental Protection Agency 262
Management of Hazardous Wastes Containing Halogenated Organics
Douglas Roeck, GCA/Technology Division, Inc 270
-viii-
-------
Page
SESSION B (Continued)
Waste Minimization Case Studies for Solvents and Metals Waste Streams
Tom Nunno, GCA/Technology Division, Inc 278
A Review of Treatment Alternatives for Dioxin Wastes
Harry Freeman, U.S. Environmental Protection Agency ....... 285
Evaluation of On-Site Incineration for Cleanup of Dioxin-Contamlnated
Materials
Frank Freestone, U.S. Environmental Protection Agency ...... 298
Review of Alternative Treatment Processes for Metal-Bearing Hazardous
Waste Streams
Douglas Grosse, U.S. Environmental Protection Agency 319
Field Evaluation of Treatment Processes for Corrosives and Metal-
Bearing Waste Streams
Neville Chung, Metcalf & Eddy, Inc 329
Recovery of Metal Values from Metal-Finishing Hydroxide Sludges
by Phosphate Precipitation
Larry Twidwell, Montana College of Mineral Science 1 Technology . 338
Treatment Technologies for Corrosive Hazardous Wastes
H. Paul Warner, U.S. Environmental Protection Agency. ...... 352
Engineering Analysis of Hazardous Waste Incineration; Energy and Mass
Balance
W. Randall Seeker, Energy and Environmental Research Corp.. . . . 360
The Formation of Products of Incomplete Combustion in Research
Combustors
George Huffman, U.S. Environmental Protection Agency 369
Analysis of PIC and Total Mass Emissions from an Incinerator
Andrew Trenholm, Midwest Research Institute 376
Concentration and Purification of Dilute Hazardous Wastes by Low
Pressure Composite Membranes
D. Bhattacharyaa, University of Kentucky. .. 382
Trial Burns - Plasma Arc Technology
Nicholas Kolak, New York State Department of Environmental
Conservation 396
-IX-
-------
Page
SESSION B (Continued)
Radio Frequency Enhanced In Situ Decontamination of Soils Contaminated
with Halogenated Hydrocarbons
Harsh Dev, IIT Research Institute 402
The Role of Rogue Droplet Combustion in Hazardous Waste Incineration
J. A. Mulholland, U.S. Environmental Protection Agency 413
Characterization of Hazardous Waste Incineration Residuals
Carlo Castaldini, Acurex Corp 421
Practical Limitation of Waste Characteristics for Effective
Incineration
Edward Martin, PEER Consultants, Inc. 429
Incinerator and Cement Kiln Capacity for Hazardous Waste Treatment
Gregory Vogel, The MITRE Corp 437
Nonsteady-State Testing of Industrial Boilers Burning Hazardous Wastes
Carlo Castaldini, Acurex Corp. ... 445
Biodegradation of Chiorophenoxy Herbicides
Dipak Roy, Louisiana State University .. ..... 454
SESSION C - COMBINED HWERL POSTER PRESENTATIONS
Earthen Liners: A Field Study of Transit Time
K. A. Albrecht, Illinois State Geological Survey. .' 462
Development of Process Monitors for Hazardous Waste Incineration
Sharon Nolen, U.S. Environmental Protection Agency. ........ 463
-X-
-------
SESSION C (Continued)
EPA Evaluation of In-Tank Methods for Detecting Leaks from Underground
Storage Tanks
James Starr, Enviresponse, Inc 464
EPA Hazardous Waste Control Technology Data Base
C.S. Fore, DOE Hazardous Waste Management Program ........ 465
The HWERL Chemical Detoxification Research Laboratory
Alfred Kornel, U.S. Environmental Protection Agency ....... 466
Laboratory Evaluation of Model Slurry Walls
Richard McCandless, University of Cincinnati 467
Environmental Emergency Response Unit Technical Information Exchange
A.C. Gangadharan, Enviresponse, Inc.. . ..... 469
Hydrostatic Testing of FMLs for Determining Critical Failure Factors
David Shultz, Southwest Research Institute 470
Data Requirements for Remedial Action Technology Screening
Paul Exner, GCA/Technology Division . 472
Development of an Environmental Testing Chamber for Soil Reagent Testing
John Glaser, U.S. Environmental Protection Agency ........ 473
Parametric Studies Delineating the Occurrence of Transient Puffs in
a Rotary Kiln Simulator
William Linak, U.S. Environmental Protection Agency 474
Canine 01 faction: A Novel Tool for Investigating Decontaminated
Equipment and Leaking Underground Storage Tanks
Herbert Skovronek, New Jersey Institute of Technology ...... 475
Data Base for Estimating Hazardous Waste VOC Emissions
John Harden, Radian Corporation .... 476
Applications of Catalytic Oxidation to the Treatment of Hazardous
Waste
Howard Greene, University of Akron 477
Thermal Decomposition Characteristics of Chlorinated Methane Mixtures
Philip Taylor, University of Dayton Research Institute. ..... 478
-X.I-
-------
Page
SESSION C (Continued)
Weathering Effects on Exposed Soil Liners
David Anderson, K.W. Brown & Associates, Inc 479
Research Activities at the Hazardous Waste Research Center - An EPA
Center of Excellence
Louis Thibodeaux, Louisiana State University 480
Hazardous Waste Research, Development, and Demonstration Permits
Nancy Pomerleau, U.S. Environmental Protection Agency 481
Center Hill Research Activities
Department of Civil and Environmental Engineering, University
of Cincinnati 482
The U.S. EPA Combustion Research Facility
Larry R. Waterland, Acurex Corp. Combustion Research Facility . . 484
Automated Expert Assistance for RCRA Permit Reviews/Evaluation of
Remedial Alternatives for Superfund Sites
Daniel Greathouse, U.S. Environmental Protection Agency 485
Prediction of Solubility of Hazardous Materials in Water
Frank A. Groves, Jr., Louisiana State University 486
Removal and Recovery of Hazardous Organic Vapors by Adsorption onto
Activated Carbon and XAD4 Resin
Kenneth Noll, Illinois Institute of Technology 487
QA/QC Support During Shakedown and Demonstration of Emergency Response
Equipment
John Koppen, Princeton Aqua Science 488
Determination of Effective Porosity of Soil Materials
Robert Morton, Iowa State University 489
Hazardous Waste Research at the U.S. EPA Test and Evaluation Facility
Douglas Grosse, U.S. Environmental Protection Agency 490
-xii-
-------
TECHNICAL RESOURCE DOCUMENTS AND TECHNICAL HANDBOOKS
FOR HAZARDOUS WASTES MANAGEMENT
Norbert 8. Schomaker and Terry M. BUss
U. S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
The U.S. EPA's Office of Research and Development (ORD) 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 Comprehensive Environmental Response Compensation and Liability Act
(CERCLA) respectively. These documents and handbooks are basically compilations of
research efforts of ORD's Land Pollution Control Division (LPCD) to date. The specific
areas of research being conducted under the RCRA land disposal program relate to
laboratory, pilot and field validation 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 post-closure evaluation. The Technical Resource Documents are intended to
assist both the regulated community and the permitting authorities for development of
new hazardous waste disposal facilities. The Technical Handbooks 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 Re-
covery Act (RCRA) of 1976 and to the
1984 Amendments 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 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 solidifica-
tion, liner systems and disposal facility
evaluation.
-------
0 Cover systems research is developing
and evaluating the effectiveness of
various material components (i.e.,
vegetation, soils, membranes, and
drainage blankets) in relation to the
cover function of minimizing moisture
Ingress and gas egress.
0 Waste leaching research is investi-
gating techniques for predicting
the composition of actual field
leachates from samples of wastes or
mixtures of wastes.
0 Waste solidification research is
evaluating the effectiveness and
performance with time of chemical
stabilization and encapsulation
processes to minimize the release
of pollutants to the environment.
0 Liner systems research is evaluating
the effectiveness and performance of
clay and flexible membrane materials
(as liners) with time to contain
and minimize the release of leachate
and gas pollutants to the environment.
0 Disposal facility evaluation research
is evaluating effective techniques to
ensure that land disposal facilities
are built as designed for either
permanent disposal or short/long-term
storage.
Incorporated throughout the research
for development of control technology
for new waste disposal facilities is a
continuous technology transfer and
assistance program of activity for the
program office and regional offices and
user communities. The TRD's are consi-
dered to be primary documents for trans-
ferring current RCRA control technologies
to the user. The TRD's are being devel-
oped 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. These documents
contain guidance, not regulations 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
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 clean up/containment technology
associated with remedial action at an
existing uncontrolled hazardous waste
site 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.
The development of proper contain-
ment technology to upgrade existing
(CERCLA) waste disposal facilities 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
evaluation.
0 Barrier research is developing and
evaluating in-situ control technolo-
gies to contain or minimize pollutant
releases from uncontrolled sites and
to predict performance with time.
In-situ control technologies such as
slurry walls, grout curtain cut-off
walls, and covers are being evaluated.
0 Waste storage research is evaluating
-------
the cost-effectiveness of placing
wastes from the clean up of uncon-
trolled sites into mines or above-
ground storage facilities. Also,
packaging of hazardous wastes in
various containers is being
investigated.
0 Waste treatment research is evalua-
ting 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.
0 Modeling technology has been
developed for evaluation of remedial
action alternatives and will be
updated to reflect improved perfor-
mance, reliability, and cost informa-
tion from field scale studies, case
studies and other research areas.
0 Post-closure research is evaluating
the criteria for final site usage
once the disposal facility has been
remediated.
Incorporated throughout the
research for.development of containment
technology for existing waste disposal
facilities is a continuous technology
transfer and assistance program of
activity for the proyram office and
user communities. The Technical Hand-
books are considered to be the primary
documents for transferring current
CERCLA containment technology to the
user.
To meet the control technology
aspects of RCRA and CERCLA as relates
to land disposal facilities, the
research program has relied heavily
on the containment aspects of the wastes
at the facility or site. The containment
aspects for waste disposal onto the land
need 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
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 environ-
ment.
TECHNICAL RESOURCE DOCUMENTS
Nine TRD's have been completed to
date. Seven 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 if shown below, along with a
brief description, publication number,
and the project officer's name in paren-
theses.
Evaluating Cover Systems for Solid and
Hazardous Waste (SW-867)
A critical part of the sequence of
designing, construction, and maintain-
ing an effective cover over solid and
hazardous 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 procedures, topographi-
cal data review and climatological data
review procedures, topographical data
review and climatological data review
procedures. The evaluation steps include
cover composition, thickness, placement,
configuration, drainage and vegetation.
The post-closure aspects include mainte-
nance and contingency plan evaluation
procedures. There are 36 specific
steps, regarding the preceding factors,
which are recommended to be followed in
evaluating a permit for a cover for
hazardous waste.
*055-000-0228-2 (R. E. Landreth).
Hydrologic Simulation on Solid Waste
Disposal Sites (SH-"868)
This TRD provides a computer
package to aid planners and designers
by simulating hydrologic characteris-
tics of landfill covers to predict
percolation as a function of cover
design and climate. It has been
updated to include the two-dimensional
aspects of landfill cover systems.
This updated version is entitled
"Hydrologic Evaluation of Landfill
Performance (HELP) Model". The HELP
model TRD will replace the SW-868 TRD.
(D. C. Ammon).
Hydrolpgic Evaluation of Landfill Perfor-
mance (HELP) Model: Vol. I (SW-84-009T
and Vol. II (SW-84-OHJT
the original waste disposal site hydro-
logic model entitled, "Hydrologic Simu-
lation 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. This TRD
was published for public comment in two
volumes (+PB 85-100-840 and PB 85-100-
832, respectively). The two volumes
include the user's guide for version 1
and documentation and description of the
program. Version 2 of the HELP Model is
being developed to incorporate public
comment and results from verification
studies and will be published in late
1986. (D. C. Ammon)
Landfill and Surface Impoundment
Performance Evaluation (SW-869)~
The evaluation of leachate collec-
tion systems using compacted clay of
synthetic liners to determine how much
leachate will be collected and how much
will seep through the liner into under-
lying 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 mathematics involved in reach-
ing his final equations. Thus, any
evaluator can take full advantage of
the TRD up to the level of his own
mathematical proficiency.
*055-000-00233-9 (M. H. Roulier).
The HELP Model is a modification of
* These TRD's have been published
and are available from GPA by
requesting the stock number.
Copies may be purchased from:
Superintendent of Documents
U. S. Government Printing Office
Washington, DC 20402
Phone: (202) 782-3238
+ These documents have been published
and the reports are available from NTIS
by requesting the stock number. Copies
can be purchased from:
National Technical Information
Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: (703) 487-4650
-------
Lining of Waste Impoundment and
Disposal Facilities (SW-870)
This TRD provides information on
performance, selection, and installa-
tion of specific liners and cover
materials for specific disposal situa-
tions, 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-install-
ation and monitoring surveys. A revised
version should be available in late 1986.
*055-000-00231-2 (R. E. Landreth)
Management of Hazardous Waste Leachate
(SN-871)
This TRD has been prepared to
provide guidance for permit officials
and disposal site operators on avail-
able management options for controlling,
treating, and disposing of hazardous
waste leachates. It discusses consid-
erations necessary to develop sound
management plans for leachate generated
at surface impoundments and landfills.
Management may take the form of leachate
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 (Sti-872)
The purpose of this TRD is to
provide 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 TRD addresses the treatment of
hazardous waste for disposal or long
term storage and surveys the current
state and effectiveness of waste treat-
ment technology. A summary of the
major physical and chemical properties
of treated wastes is presented. A
listing of major suppliers of stabili-
zation/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 impound-
ment are discussed and referenced.
Problems related to abandoned impound-
ments causing environmental degradation
are discussed. Closure methods such as
waste removal, consolidating the waste
on site and securing the site as a
landfill are also discussed. It is
written primarily for staff members
in EPA regional offices or state regula-
tory offices, who are charged with
evaluating and approving closure plans
for surface impoundments under regula-
tions 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)
The objectives of "Hazardous Waste
Land Treatment" are to describe current
technology and to provide methods for
evaluating the performance of an appli-
cant's hazardous waste land treatment
facility design. Land treatment is
approached comprehensively from initial
site selection through final closure,
and additional information sources are
referenced 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 TRO's now being developed
or in planning stages by the Office of
Research and Development are shown below,
along with a brief description and the
project officer's name in parentheses.
Soil Properties, Classi'_f_i_c_at1_g_n and
Hydraulic Conductivity Testing
This TRO is a compilation of availa-
ble laboratory and field testing methods
for the measurement of hydraulic conduc-
tivity (permeability) of soils. Back-
ground information on soil classifica-
tion, soil water, and soil compaction are
included along with descriptions of six-
teen methods for determination of
saturated or unsaturated hydraulic
conductivity. This TRD (Sw-925) was
published in March 1984 for public
comment. It is being revised to incor-
porate public comments that were
received. The revised version will be
published through the Government Printing
Office in late 1986. (M. H. Roulier)
Solid Waste Leaching Procedures Manual
This is a TRD on laboratory batch
procedures for extracting or leaching
a sample of solid waste so that the
composition of the lab leachate is
similar to the composition of leachate
from waste under field conditions. This
TRD (SW-924) was published in March,
1984 for public comment. It is being
revised to incorporate public comments
that were received. The revised version
will be available through the Govern-
ment Printing Office in late 1986.
(M. H. Roulier)
This project is studying the
feasibility of using soil/leachate batch
adsorption procedures for designing
compacted clay liners to remove/retain
known amounts of pollutants, and will
contribute to information on designing
and evaluating clay liners to limit
pollutant release from landfills. The
project will be completed in late 1986.
(M. H. Roulier)
Methodsforthe Prediction of Leachate
Plume Migration and Mixfng"
This project has developed a variety
of computer programs for hand-held cal-
culators, microcomputers, and macro-
computers. 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 will be completed in early
1987. (M. H. Roulier)
Design, Construction, Maintenance, and
Eva!uation of' C1ay Liners for Hazardous
waste Facilities
This TRD has gathered and assembled
information on construction and use of
compacted soil liners, including test
methods, chemical compatibility, failure
mechanisms, performance of existing
facilities, and field quality assurance
and inspection to insure that the liner
performs as designed. The project is
scheduled for completion in summer 1986.
(M. H. Roulier)
TECHNICAL HANDBOOKS
Nine technical handbooks have been
completed to date. These handbooks
cover a variety of techniques from
general remedial action guidance; to
in-situ treatment; to barriers; to
decontamination and modeling. A listing
of these handbooks is shown below,
along with a brief description, publica-
tion number, and the project officer's
name in parenthesis.
Guidance Manual for Minimizing Pollution
from Maste Disposal Sites (EPA OTO/2-;
78-142)'
-------
This Handbook provides guidance
in the selection on available engineer-
ing technology, to reduce or eliminate
leachate generation at existing dumps
and landfills. Most of the techniques
discussed in the report deal with the
reduction or elimination of infiltration
into landfills in one of five
categories, active groundwater or plume
management, chemical immobilization of
wastes, and excavation and reburial.
The manual emphasizes passive remedial
measueres, requiring little or no
maintenance once emplaced. Other
measures are active and require a contin-
uing input of manpower or electricity.
+PB 286 905/AS (D. E. Sanning)
Review of In-Pi ace Treatment Techniques
for Contaminated Surface Soli's" Vo1_7T
(EPA 540/2-84-0033) and Vol. II (EPA
540/2-84-005bJ~~~
This two-volume Handbook presents
information on in-place treatment
technologies applicable to contaminated
soils less than 2 feet in depth. Volume
I discusses the selection of the appro-
priate 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, degrada-
tion, attenuation, or reduction of
volatiles) is determined on the basis
of information available from the
remedial investigation. Selection of
a specific technology involves
assessment of waste, soil, and site-
specific variables. The technology is
implemented if it is considered more
cost-effective in comparison with the
other alternatives.
+PB 85-124881 Vol. 1 (N. P. Barkley)
+PB 85-124889 Vol. 2 (N. P. Barkley)
Remedial Action at Waste Pisposal Sites
(revised) (EPA 62S/6-85-006)
This handbook is intended for use
as a basic reference tool on remedial
action. It will assist government and
industrial officials and technical
persons with the range of remedial
technologies; their applications and
limitations; major design and construc-
tion considerations; technology
selection/evaluation; operation, mainte-
nance and monitoring; and approximate
costs. It provides sufficient infor-
mation on remedial technologies to
enable the user to select potentially
applicable technologies for a given
site, and understand what is involved
in designing and implementing these
technologies. While emphasis is placed
on those technologies which have been
demonstrated for hazardous waste sites,
information is also included on emerging
technologies. No order number is yet
available. (D. E. Sanning)
Handbook for Evaluating Remedial Action
Technology Plans(EPA 600/2-83-076)
This Remedial Action Handbook des-
cribes how the technologies and methods
for evaluating proposed new RCRA
hazardous waste disposal sites can be
applied to site-specific remedial
response activities for uncontrolled
hazardous waste sites. The Remedial
Action Handbook is based on the state-
of-the-art technical and cost information
in eight TRDs for design and evaluation
of new hazardous waste disposal sites
under RCRA. That information will be
reviewed for relevance to remedial
response at uncontrolled hazardous waste
disposal sites, and will be edited and
refocused to address the needs of per-
sonnel involved in response and remedial
action planning under CERCLA.
+PB 84-1182149 (H. R. Pahren)
-------
Slurry Trench Construction for Pollution
Migration Control (EPA 540/2-84-OOlT
A Handbook for slurry trench cut-
off wall design, construction, and
performance evaluation provides recom-
mendations on a variety of scientific
and technical parameters relevant to
using this approach to isolate
hazardous chemicals in near-surface
groundwater regimes. The accomplish-
ment of this effort required extensive
information gathering and integration
of technical data gathered from a
diverse array of experience and autho-
rities. +PB 84-177831 (W. E. Grube)
Guide for Decontaminating Buildings,
Structures and Equipment at Superfund
Sites (EPA 600/2-85-02irr
A decontamination Handbook was
designed for EPA Headquarters program
offices and regional Superfund programs
as part of the restoration 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) decontamina-
tion methods, 3) economics, 4) health
hazards and 5) availability of
equipment/personnel for the detoxifica-
tion procedures. Specific waste types
found in contaminated buildings,
structures and equipment at Superfund
sites are identified. Potential secon-
dary impacts from the available and
potential decontamination treatment
methods are addressed in this study.
Costs versus risk and projected
ultimate site usage are also addressed.
Methods for monitoring the success of
the various procedures are defined.
+PB 85-201234 (N. P. Barkley)
Modeling Remedial Actions at Uncontrolled
Hazardous Haste Sites (EPA 540/2-85-001)
The objective of this Handbook is
to provide technical guidance on the
selection and application of models for
evaluating remedial action alternatives
at uncontrolled hazardous waste sites.
The volumes cover selection of models,
simplified methods for subsurface and
waste control actions, and analytical
and numerical models for evaluation of
surface water remedial actions.
+PB 85-211357 (D. C. Ammon)
Leachate PI ume Management (EPA540/2-85-
OTJ
A Study on leachate plume manage-
ment techniques has been conducted.
This handbook describes the factors
that affect leachate plume movement,
key considerations in delineating the
current and future extent of the
leachate plume technologies for con-
toning the migration of plumes, and
criteria for evaluating and selecting
plume management alternatives.
+PB 122330 (N. P. Barkley).
DustControlat Hazardous Waste Sites
(EPA 540/2-85-003)
Three field studies were performed
to determine the effectiveness of dust
control technologies at hazardous sites.
In the first field study, dust suppres-
sants were tested to determine the
effectiveness of fugitive dust control
against wind erosion from exposed areas.
The second field study was an evaluation
of the effectiveness of windscreens and
windscreen/dust suppressant combinations
in controlling fugitive dust from storage
piles. The third field study, investi-
gating active cleanup emissions, consisted
of testing fugitive dust control measures
applicable to the use of a front-end
loader to load dirt into a truck. No
order number is yet available.
(S. C. James)
-------
FUTURE TECHNICAL HANDBOOKS
Additional technical handbooks
are now being developed or are in
planning stages by the Office of
Research and Development. A listing
of these handbooks is shown below,
along with a brief description and
the project officer's name in paren-
thesis.
Evaluation of Systems to Accelerate the
Stabilization of Waste Piles of Deposit's
This Handbook investigates in situ
systems which accelerate the stabiliza-
tion of waste deposits. In situ
applications involve three essential
elements: selection 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
reactants 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
well systems. This Handbook will be
available in the fall of 1986.
(y. E. Grube)
Covers for Uncontrolled Hazardous
Waste Sites
A handbook is being developed
which can be used as a guidance
document for the selection, design,
installation, and long-term mainte-
nance of covers as remedial actions.
This handbook will provide technical
information for regulatory personnel
as well as guidance for cover-system
designers and construction engineers,
and will be available in late 1986.
(d. M. Houthoofd)
Stabi1i zation/Sol i di fi cati on Al ternati ves
for Remedial Action at Uncontrolled
Waste Sites
Another handbook is being 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 application of solidification/
stabilization is divided into two
phases: process selection and scenario
selection. Process selection is con-
cerned with the chemistry of the
Stabilization/solidification processes
in the identification of the composition
of the waste. Presented in the handbook
are testing and analysis techniques for
characterizing waste as a basis for
selection of pretreatment and stabili-
zation/solidification processes. Also
data are being developed on the compati-
bility of additives and specific classes
of waste, and testing systems for the
evaluation of stabilized/solidified
wastes are being reviewed. Scenario
selection is concerned with the develop-
ment of equipment requirements, construc-
tion sequencing, and cost estimating
for the chosen solidification/stabili-
zation process. The handbook will pre-
sent, based on field surveys, four
basic field scenarios that have been
used successfully, and will be published
in the fall of 1986.
(J. M. Houthoofd)
CONCLUSION
The Technical Resource Documents
and the Technical Handbooks that are
being prepared and updated by the ORD's
-------
Land Pollution Control Division
(LPCD) are a series of documents which
provide best engineering control
technology to meet the needs of RCRA and
CERCLA, respectively. The TRD's provide
design operation, and evaluation informa-
tion 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 facilities to
assist the user community and on-scene
coordinators. These documents and
handbooks present the sura 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 TRD's
and Technical Handbooks so that they
reflect the latest state-of-the art
Information.
More information about a specific
project or study can be obtained by
contacting the Project Officer
referenced in the text.
The 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
Telephone (513) 569-7871.
10
-------
LEAKING UNDERGROUND STORAGE TANK
RESEARCH PROGRAM
Richard Field, Anthony N. Tafuri
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, New Jersey 08837
ABSTRACT
The U.S. Environmental Protection Agency's (EPA) Office of Research and Development
is undertaking research to control leaks from underground storage tanks (UST) containing
petroleum products and hazardous substances regulated under the Comprehensive Environmen-
tal Response, Compensation, and Liability Act of 1980 (CERCLA). The principal purpose of
this effort is to protect the nation's groundwater supplies from UST leak contamination
as mandated by the Resource, Conservation, and Recovery Act reauthorization of 1984. The
EPA Office of Underground Storage Tanks is responsible for the promulgation of rulemaking
for UST leak control and the Hazardous Waste Engineering Research Laboratory (HWERL) to-
gether with the Environmental Monitoring and Support Laboratory-Las Vegas is responsible
for the development of technical background information for the mandated rulemaking. The
HWERL research program covers the areas of UST leak prevention, detection, and corrective
action.
INTRODUCTION
During the 1950s and 1960s, the
consturction of many gasoline stations,
chemical manufacturing and processing
facilities and other facilities storing
hazardous (toxic, flammable, etc.)
materials in bulk led to the installation
of millions of underground storage tanks
(UST). The United States Environmental
Protecton Agency (EPA) has estimated that
for petroleum products alone there are
approximately 2.5 million UST currently
in use. In addition, an unknown number
of abandoned tanks exists and undoubt-
edly not all have been appropriately
drained of their contents. Many of these
tanks have exceeded or are currently
close to the end of their useful life,
and because of this or other reasons,
are now (or will soon be) leaking, posing
a serious threat to the nation's ground-
water supplies as well as to the public
health and welfare. Leaks can be viewed
as a potential threat to human life and a
potential source of large economic losses
because they can also damage building
foundations, present fire or explosion
hazards, and damage crops, livestock and
wildlife. An immediate need exists for
cost-effective techniques and equipment
that can be applied to the prevention
and control of releases from leaking UST
so as to minimize or eliminate unreasonable
risks to health and the environment.
PROBLEM
Billions of gallons of petroleum prod-
ucts and other hazardous substances are
stored in underground tanks. A basic prob-
lem associated with this mode of storage
is leaks and their associated costs,
11
-------
dangers and potential environmental
threats. The primary causes of leaks from
UST are corrosion, poor installation prac-
tices and poor operating practices.
Unprotected steel tanks begin to leak
anywhere from 2 to 20 years after installa-
tion. Even steel tanks with corrosion
protection can corrode if that protection
1s not correctly designed, installed and
maintained. Fiberglass tanks will not
"corrode" but these tanks are not as
strong as steel tanks and may begin to
leak If they are not properly installed
or if incompatible materials are stored
in them. Better information is needed to
characterize existing hazardous material
storage practices adequately and to
evaluate the economic impacts of preven-
tion and control techniques.
Currently, over 100,000 petroleum and
other chemical storage tanks are estimated
to be leaking. Leak rates are often slow
and unnoticed; large quantities of haz-
ardous substances may be leaked into slow-
moving groundwater over many years before
being detected. Available information
regarding storage tank leaks is largely
anecdotal and is insufficient to accur-
ately assess environmental impacts.
Better techniques are needed to detect
leaks and measure leak rates from under-
ground tanks and associated piping.
RATIONALE
The Resource, Conservation, and
Recovery Act reauthorization legislation
of 1984 (RCRA) requires EPA to promul-
gate regulations to control underground
storage tanks containing "regulated sub-
stances": petroleum products and hazardous
substances as Indicated by the Comprehen-
sive Environmental Response, Compensation,
and Liability Act (CERCLA). According to
the law, EPA must set final standards for
existing tanks that cover leak detection
and tank testing, recordkeeping and re-
porting corrective action, financial
reponsibility (as deemed necessary) and
closure. Regulations for new tanks must
include design, construction, installation,
release detection and compatibility
standards.
Short-term EPA research efforts are
required to establish, in a timely manner,
a strong technical foundation for these
regulations. Deadlines mandated in the
RCRA amendments of 1984 for UST require
new/existing petroleum tank standards
by 2/87 and new/existing chemical tank
standards by 8/88. In the longer term,
EPA research efforts will support the
revision of regulations promulgated under
RCRA for UST.
PROGRESS TO DATE
Hazardous Waste Engineering Research
Laboratory (HWERL) activities to date have
included a state-of-the-art (SOTA) review
of existing leak detection methods, and
initiation of controlled-condition field
studies to evaluate and verify detection
methods and develop information for im-
proved methods for application to under-
ground fuel and chemical storage tanks.
Construction of a controlled-condition
test site at HWERL's Edison, New Jersey
facility is estimated to be completed
during May 1986.
Considerable experience has been
obtained in spill prevention, control, and
countermeasures for aboveground storage
facilities. Much of this experience is
contained in two reports prepared by HWERL
for EPA's Office of Emergency and Remedial
Response; a third report on the prevention
of hazardous spills has recently been
completed.
In addition, under support from the
Superfund program, HWERL has developed, or
is currently developing, the following
technologies that, with further research,
can be adapted to the unique problems asso-
ciated with UST: containment methods
(e.g., slurry walls, plume control, and
covers); in situ methods (e.g.,-soil flush-
ing, soil freezing, electrokinetics, grout-
ing, and soil detoxification) to clean up
hazardous waste sites; and on-site, mobile
soil- and water-treatment systems to decon-
taminate soil, recover contaminants and
treat contaminated waters.
12
-------
UST related projects underway at
the Environmental Monitoring and Support
Laboratory-Las Vegas Include: (a) the
evaluation of soil-gas monitoring techni-
ques; (b) an inventory of UST's using
aerial photography; and (c) geophysical
technical support to the U.S. Air Force
for mapping hydrocarbon contamination
from UST.
A national survey of motor fuel tanks
performed by EPA's Office of Toxic Sub-
stances will supply preliminary data for
evaluating the useful life of new and
existing UST's. In addition, EPA's
Office of Solid Waste is involved in a
joint effort with the Department of
Defense to evaluate and verify leak detec-
tion in large UST; to conduct a study of
pressurized vessels and current practices
as they pertain to underground storage of
regulated substances; and to perform a
SOTA study on compatibility and current
standards used by tank manufacturers and
testing institutes.
RESEARCH PROGRAM
HWERL's research program is being
directed toward satisfying the needs of
three subcategories: release prevention,
leak detection, and corrective action.
Release Prevention
The major problems associated with
this area involve the identification and
improvement of design, construction, and
installation practices (in relation to
locational factors) for effective preven-
tion of releases from underground storage
systems. Guidelines to estimate the
remaining life of a tank at a given age
over a range of environmental conditions
are another high priority research need
under this category. These outputs are
required to support the development of
performance standards for existing and
new tanks.
An extensive body of knowledge exists
on the causes and prevention of corrosion.
Similarly, much information and experience
has been obtained on methods and materials
used for secondary containment of hazard-
ous wastes. Much of this is applicable to
underground storage tanks. The feasibil-
ity of retrofitting tanks with cathodic
protection, however, is not known. A
value or rating system to anticipate tank
life under various soil conditions is
needed. Limited information exists on
the compatibility of storage tanks with
their contents. Current tank testing
practices do not include procedures to
determine the long-term (greater than one
year) performance of tank materials with
various contents.
A background SOTA report on structural
and non-structural prevention practices and
developing techniques to evaluate and ex-
tend the useful life of both new and exist-
ing underground storage tanks has nearly
been completed in draft form.
Leak Detection
The major problems associated with
this area involve the identification and
verification of techniques to detect leaks
and measure leak rates from underground
tanks and associated piping. An immediate
need in this area and one that is currently
being addressed is the identification and
verification of existing leak detection
methods. Of the 36 leak detection methods
known, many were derived from techniques
developed for aboveground tanks. Methods
involve examining the environment that is
contiguous to the tank system for traces
of the contents of the tank (environmental
methods); measuring the contents of the
tank system for changes in the amount of
material stored (volumetric methods); or
checking the integrity of the tank system
to identify the existence of orifices that
may leak (non-volumetric methods). Little
data exist on the accuracy, precision, and
sensitivity of these methods when applied
to underground tank systems.
The previously mentioned controlled-
condition UST test site has been partially
constructed to evaluate existing techno-
logical approaches (and associated per-
formance data) and develop improved ones
for detecting leaks in UST. It is anti-
cipated that the transfer of information
to practitioners and potential users
will enable detection equipment refine-
ment and better operation. The program
will start with gasoline and diesel
13
-------
detection equipment and in the near
future evaluate tank testing for other
hazardous substances.
Corrective Action
The major problems associated with
this area involve stopping further
leaks from a leaking tank system and
correcting environmental damage.
Corrective actions may involve several
approaches: (1) stopping the leak by
merely emptying the tank or by adding
polymers or in-place coatings that will
seal the leak; (2) application of emer-
gency response techniques to alleviate
an immediate threat to life and/or
the environment; (3) long-term cleanup
and repair or replacement. As a first
priority, current practices in emergency
and remedial response should be investi-
gated for application to UST. Research
results and field experience are avail-
able from the cleanup of spills of oil
and hazardous materials, and guidance
has been developed on containment methods.
However, further research is required to
adapt and evaluate the application of
these technologies to the unique problems
associated with UST.
A project has been initiated to adapt
existing emergency response and remedial
response technological approaches to
correcting releases from leaking under-
ground storage tanks by reviewing avail-
able emergency response and remedial tech-
niques and selecting the most promising
ones for application to UST releases.
14
-------
HAZARDOUS MATERIALS RELEASES
RESEARCH PROGRAM
Anthony N. Tafuri
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, New Jersey 08837
ABSTRACT
As mandated by the Clean Water Act and the Comprehensive Environmental Response,
Compensation and Liability Act of 1980 (CERCLA), the U.S. Environmental Protection
Agency's (EPA) Office of Research and Development 1s undertaking research to prevent,
control, and eliminate pollution caused by the releases of hazardous materials into
the environment. Such releases can result front transportation accidents, inplant
discharges where the hazardous material leaves the boundary of the plant, and uncon-
trolled hazardous waste disposal sites. These releases pose a problem of national
significance; of greatest concern is the irreversible contamination of the nation's
drinking water aquifers. What is needed, and what is being pursued under the Hazardous
Waste Engineering Research Laboratory's (HWERL) research program is a systematic and
thorough examination/documentation of cost-effective technologies that can be applied
as countermeasure operations to mitigate risk to public health and to the environment
caused by these release situations.
INTRODUCTION
Hazardous material emergencies
involve acute releases (or threats of
acute releases) of chemicals or mixtures
of chemicals into the environment that
present an actual or potential detrimental
effect on the public health, welfare
and/or ecosystem. Such emergencies include
releases from transportation accidents
(rail cars, tank trucks, vessels, and
pipelines); inplant releases, where the
hazardous material leaves the boundary of
the plant; and releases from uncontrolled
hazardous waste disposal sites. These
releases pose a problem of national signi-
ficance and have resulted in loss of
lives, and, when viewed over a period of
years, measurable contamination of our
natural resources. Of greatest concern is
the irreversible contamination of the
nation's drinking water aquifers, since
one-half of the population of the U.S.
depends upon groundwater for water
supply. Accordingly, there is an imme-
diate and continuous need to develop
cost-effective technologies and scientific
techniques that can be applied to the
prevention and control of hazardous
material releases so that on-scene coor-
dinators and other emergency response
personnel can direct countermeasure
operations to mitigate risk to public
health and to the environment.
PROBLEM
Each year, there are hundreds of
reported releases of chemicals which have
been designated by The Clean Water Act and
CERCLA as hazardous. Some of these
15
-------
incidents Involve carcinogens/chronic
toxic substances such as arsenic, benzi-
dlne, and asbestos, which pose significant
cleanup problems because of their extreme
hazardous nature and required extent of
removal for safe conditions. The Clean
Water Act and CERCLA lists of hazardous
substances also include persistent bio-
refractory (and often bioaccumulative)
substances, such as polychlorinated bi-
phenyls (PCBs), which have been reportedly
spilled at a rate of several hundred times
per year.
In addition, releases from uncon-
trolled hazardous waste sites continue to
significantly impact the environment.
There are approximately 22,000 potentially
hazardous waste sites identified through-
out the U.S., of which many are expected
to require some form of removal and/or
remedial action.
By solving this problem, health risks
and the irreversable loss of our natural
resources from hazardous material releases
can be prevented or reduced. Because
these releases can contaminate land, air,
and water, the technical problems asso-
ciated with response are notably diverse
and involve many possible combinations of
thousands of substances and mixtures,
volumes spilled or released, human popu-
lations Impacted, and other local factors
such as geology and weather. What is
needed, and what is being conducted under
HWERL's research program, is a systematic
and thorough examination and documentation
of cost-effective technologies and tech-
niques for the prevention and control of
hazardous material releases.
PROGRESS TO DATE
As mandated by the Clean Water Act
and CERCLA, EPA is engaged in such
research development and demonstrations
as appropriate to prevent, control, and
eliminate pollution caused by the release
of hazardous materials into the environ-
ment.
Progress to date has included the
development of hardware and research
reports and methodologies associated with
such areas as: prevention (reduction of
release frequency and severity); assess-
ment (decision-making information for
first-responders and site cleanup mana-
gers); containment (limiting the spread of
hazardous substance release in all
environmental media); concentration and
separation (collecting and concentrating
a dispersed hazardous material into a
minimum practical volume and separating
the material from co-collected air, water,
soil or sediments); and destruction and
disposal (providing for the destruction of
residuals from hazardous material releases
cleanup operations with emphasis on tech-
nologies useable on-site).
Significant accomplishments in these
areas include development of the following
technologies:
0 Mobile physical/chemical treatment
system consisting of carbon adsorp-
tion, mixed media filtration, floccu-
lation, precipitation, neutralization,
etc., for field use in removing in-
organic and organic hazardous sub-
stances from wastewaters (commer-
cially available);
0 Hazardous materials detection kit to
detect and monitor the location of
chemical releases in water when the
identity of the chemical is known
(commercially available);
0 Enzyme-based system for detecting the
presence of spilled organophosphate
and carbamate pesticides in water
(commercially available);
0 Foam dike system which provides an
"instant dike" from a portable back-
pack apparatus for the emergency
containment of released hazardous
materials (commercially available);
0 Mobile Spill Alarm System - an 1 fi-
st ream warning system consisting of
a number of individual probes and
sensors (TOC, conductivity, UV
absorption, etc.) for the continuous
detection of a broad variety of
hazardous materials in water;
16
-------
0 Emergency Response Test Kit - a port-
able device for measuring oxidation
and reduction (REDOX) potential of
unknown mixed chemicals at abandoned
waste sites for the purpose of pre-
venting explosive reactions between
oxidizing and reducing chemicals
during drum compositing operations;
0 Oil Dispersant Application System -
a system of port and starboard spray
booms and associated pumps, hoses,
and fittings, designed to be installed
on a vessel for the application of
dispersants to oil spills;
0 Mobile Decontamination Station - a
semi-trailer van designed to support
the personnel decontamination needs
of cleanup activities involving high-
ly toxic materials;
0 Mobile Carbon Regenerator - a system
designed for field use in reactiva-
ting spent activated carbon used in
spill or waste site cleanup opera-
tions; system consists of a pyrolytic
kiln, a secondary combustion chamber
and air pollution control devices;
0 Mobile In-Situ Treatment System - a
system designed to clean and detoxify
subsurface contaminated soils in place
by water flushing with additives, or
by chemical reaction;
0 Mobile Incinerator - a system consist-
ing of four semi-trailers equipped with
specialized combustion equipment, air
pollution control devices and monitor-
ing equipment for the purpose of
on-site thermal destruction of refrac-
tory organic compounds (e.g., PCBs,
Kepone, Dioxins);
0 Mobile Soils Washing System - a system
designed to scrub contaminants from
excavated soil at a spill or hazardous
waste site so that the treated soil
can be returned to the site.
RESEARCH NEEDS
In general, the most urgent needs are
for techniques and equipment for the safe,
rapid assessment of a hazardous material
release situation; the evaluation of clean-
up alternatives; and the development of on-
site control techniques which are normally
limited to containment of the released ma-
terial to prevent spreading or concentra-
tion of the material prior to removal of
the collected residuals to an off-site
location for disposal. Since on-site dis-
posal or destruction is now only emerging
as a technology, long-term needs are for
on-site permanent disposal or destruction
of the residuals collected from cleanup
operations, and restoration of the impacted
area to its pre-contarninated condition.
Situation assessment, alternatives
evaluation, and decision-making activities
are high on the priority list of needs.
There is a very clear immediate need for
standards for hazardous materials in air,
water, and groundwater to assist in the
determinations of "how clean is clean" and
the extent of "significant" contamination
at a release episode. Procedures currently
under development to assist in making
cleanup priority decisions must be accel-
erated. Because analytical information
plays such an important role in decision-
making, there is a distinct need for rapid
analytical methods development for precise,
quantitative determinations of the amount
of hazardous material in air, water, soils,
sediments, and groundwater. Furthermore,
there is a need for on-site kits for
screening a variety of samples to determine
which are suitable for more precise, but
more time consuming and expensive analysis.
Mathematical modeling must be made field-
worthy to take advantage of the essentially
untapped capability of automatic data
processing for releases control. Models
usable in the field are needed for predict-
ing movement of contaminants in"all media.
Automatic data processing should also be
applied to "lessons learned" from past
cleanup operations, since these exper-
iences come at a significant cost and
are worthy of recording in a retrievable
manner to minimize recurrence. Specific
emphasis must also be placed on performing
rapid pilot studies on emerging RSD tech-
nologies as well as on tests for evalua-
tion of alternative treatment options for
carcinogenic or unknown materials. (The
cost of treatability studies conducted at
17
-------
pilot scale is small when compared to the
excess cost of choosing an unnecessarily
expensive cleanup alternative.)
On-s1te control actions for hazardous
material releases are currently limited
to containment activities to limit the
spread of the released material, and
concentration and collection actions
taken prior to transportation of the
collected residuals off-site. Increased
emphasis must be placed upon the develop-
ment of a complete assortment of on-site
control actions that can be used by OSCs
for the many differing release circum-
stances. Rapidly deployable means to
limit the spread of spilled hazardous
materials must be developed for releases
affecting air, surface waters, sediments,
soils, and groundwaters. Furthermore,
once the spread is confined, if possible,
a wide variety of treatment options must
be available, including an increased
emphasis on the use of dispersants to
control floating materials as well as
other techniques to treat streams and
rivers—including both flowing water and
sediments—that have become contaminated
with hazardous materials. (The Oil and
Hazardous Materials Simulated Environ-
mental Test Tank [OHMSETT] facility can
provide for much-needed research into
the treatment of large bodies of water
contaminated by hazardous material
releases.)
There must be an increased emphasis
on treating entire spill situations on-site
to the maximum extent possible, because
of the significant difficulties associated
with transportation of hazardous materials
off-site, and the problems associated with
using landfill sites for "final" disposal
of the cleanup residuals. There is a long-
term need for research into alternative
means for disposing of, or otherwise
destroying, spilled hazardous materials
where they are spilled. New approaches
must be pursued for the destruction of
PCBs and other refractory compounds,
particularly in soils and sediments.
These approaches should include both
In-sltu and process-type techniques
employing physical, chemical, and biolo-
gical methods. Furthermore, an acceler-
ated emphasis must be placed on restoring
spill-damaged areas to their pre-spill
condition.
RESEARCH PROGRAM
Activities associated with HWERL's
Hazardous Materials Releases Research
Program have been divided into specific
technological areas.
Prevent1 on
The overall objective of the Preven-
tion area is to develop technological
approaches to the reduction of release
frequency and severity. Activities to date
have included evaluating pollutant release
data and transferring information on re-
leases control technology from major chem-
ical handlers and manufacturers to smaller
facility operators. A technical handbook
and training program on release prevention
technology have been prepared; based on the
state-of-the-art review associated with
this work, future efforts will pursue the
development of prototype spill prevention
techniques and equipment.
Assessment
The Assessment category addresses
decision-making information for first-re-
sponders and site cleanup managers. Work
has been directed toward: preparing a
manual of practice for selecting removal
(countermeasure) actions for hazardous
material release situations; developing
and demonstrating non-destructive testing
techniques to locate sub-surface releases
and chemical containers, and canine
olfaction techniques to locate residual
hazardous or toxic material contamination
and underground storage tank leaks and
monitor cleanup equipment decontamination;
developing and demonstrating rapid
analytical methods and techniques to
assess levels of pollution before and
after a release situation and to support
mobile and central laboratory operations
during cleanup episodes; and developing
and demonstrating predictive models to
determine the flow, dispersion and f
settling of insoluble hazardous materials
in streams, and the performance of
physical-chemical treatment technologies
utilized during hazardous material
release situations.
Future activities will demonstrate
the field applicability of such non-destruc-
tive testing techniques as ground-probing
18
-------
radar, microwaves, sonics/ultrasonics and
magnetometer metal detectors and will
investigate additional more innovative
methodologies such as dye penetrants,
lasers, and nuclear radiation.
Predictive models to determine the
movement of hazardous materials in all
environmental media and the performance
of treatment technologies utilized
during hazardous release situations
are in the early stages of development.
These models will require extensive
field verification and refinement
before they can be considered field-
worthy. Future work will address this
situation and will pursue the generation
of a "library" of models for the
variety of treatment technologies that
are available.
Containment
Containment technology is concerned
with limiting the spread of a hazardous
substance release in all environmental
media (air, soil, groundwater, surface
water, and sediments). Activities to
date have included developing techniques
for controlling spillage from impoundments
and waste lagoons, including rapidly
deployable covers to prevent overtopping
from rainwater, and stabilization and
reinforcement techniques for impoundment
walls and dikes; and evaluating the
performance of containment devices for
spilIs in water.
Few objective evaluation data exist
on containment equipment for floating
spills. Such data are required to be
specified in exploration and development
plans for offshore oil and gas production
platforms (to demonstrate capability
of coping with pollution accidents).
HWERL's QHMSETT facility is currently
addressing this situation as part of
an Interagency Technical Committee
(co-supported by, the Minerals Management
Service, Coast Guard, Navy, and
Environment Canada). Typically,
equipment is evaluated first in OHMSETT1s
main tank with oil and then offshore
without oil. Equipment most commonly
specified are being evaluated first.
It is anticipated that this evaluation
process will continue into the near
future.
Concentration and Separation
Research in the Concentration and
Separation area is directed toward
technology to collect and concentrate
a dispersed hazardous material into a
minimum practical volume and separate
the material from co-collected air, water,
soil or sediments. Current activities are
involved with: developing treatment tech-
niques for processing high-strength wastes
from spill and waste storage sites; devel-
oping a technical handbook on mobile treat-
ment technologies for removing volatile
organics from water and air; developing
new and improved approaches for handling
toxics-contaminated sediments; identifying
and evaluating commercially available
sorbents for congealing spilled hazardous
liquids and contents of damaged drums and
preparing a user's guide for field selec-
tion; and developing technical handbooks
on state-of-the-art spill cleanup equipment
and technology (evaluated at the OHMSETT
facility) with specific emphasis on
improved techniques for cold climate
removal operations and shoreline cleanup
methods.
Future activities will continue to
investigate techniques for processing
high-strength wastes and will include
combinations of processes such as liming,
membrane separation, and recirculated
aerobic/anaerobic processing tailored to
attain greater removal efficiencies and
evaluated for actual on-site waste treat-
ment. Pilot and field verification
studies will be conducted on such techni-
ques as unaided evaporation, evaporation
enhanced by mechanical agitation, air
stripping columns and distillation systems
for a variety of volatile and gaseous
materials of Interest. Other areas that
will be investigated include: high-energy
scrubbing and countercurrent chemical
extraction techniques for removal of
chlorinated organics from excavated soils
and recovered sediments; and techniques
and equipment for handling spill situa-
tions involving emulsions.
19
-------
Destructionand Disposal
The objective of this area is to
provide for the destruction of residuals
from hazardous materials releases cleanup
operations with emphasis on technologies
useable on-site. Activities have involved
developing novel destruction/detoxifica-
tion techniques with the ultimate objec-
tive of mobile system design for on-site
treatment; evaluating the application of
thermal desorption/destruction technology
to CERCLA-designated high-hazard chemicals;
developing a computer model for evaluating
the use of dispersants at a release situa-
tion; evaluating National Contingency
Plan dispersants and application techni-
ques; evaluating and Improving test proce-
dures for determining dispersant effec-
tiveness; and evaluating the effectiveness
of existing dispersants for releases under
varying temperature and salinity condi-
tions and in specific inland waterway
situations.
Future activities will pursue pilot
scale evaluations of novel methods such as
high-temperature fluid wall reactor pyrol-
ysis, in situ radio-frequency heating, and
wet-air oxidation for destruction/detoxifi-
cation of residuals of hazardous materials
from cleanup operations. Initial work in
thermal destruction/desorption of high-
hazard chemicals involved dloxin; future
work will address other CERCLA-designated
high-hazard chemicals. The potential for
dispersants to cope successfully with
difficult release situations remains un-
proven; accordingly, additional work will
be required to determine and improve upon
dispersant effectiveness in mitigating
spills and protecting shorelines.
20
-------
EFFECTIVE POROSITY OF GEOLOGIC MATERIALS
Gary R. Peyton
James P. Gibb
Mary H. LeFaivre
Joseph D. Ritchey
Stephen L. Bureh
Michael J. Barcelona
Illinois State Water Survey
Champaign, Illinois 61820
ABSTRACT
Estimation of travel times for pollutant solutes through fine-grained, saturated
porous media is of interest in a number of areas. In all cases, such estimation requires
a knowledge of the porosity of the material. Evidence from the chroinatographic
literature suggests that the "effective porosity" of a porous medium may not necessarily
be the same for all solutes and/or conditions. The object of this work was to define,
understand and measure effective porosity. Diffusion is seen to play a major role in
determining breakthrough times In clayey materials. Approximate breakthrough times may
be simply calculated using chromatographie data analysis techniques. Differences in
effective porosity of a medium for different solutes may be very important under the
conditions of facilitated solute transport.
INTRODUCTION
It is of interest to be able to
predict travel times and velocities of
pollutant solutes moving through
saturated porous media. In particular,
understanding the movement of solutes
through fine-grained materials, such as
the clays from which landfill liners and
caps are constructed, is important for
the purpose of predicting the "break-
through" time of the solute, that is, the
time required for that solute to appear
at a specified distance downgradient from
the source. While Darcy's law can be
used to calculate bulk flow rates in
coarser materials, its application to
solute transport in clayey materials is
questionable. Even in situations where
Darcy's law can be used to estimate to
movement of the center of mass of a
pollutant plume, a value for the porosity
of the material must be known. These
values are usually obtained from tables
of porosities in the geologic literature.
Furthermore, at sufficiently low
velocity, dispersion makes an important
contribution to the breakthrough of a
solute. Experience from the field of
chromatography indicates that a) the
"effective" porosity, that is the void
volume "available" to the solute might
vary with the nature of the solute and
the actual flow conditions, as well as
being different from the "total"
porosity, and b) that commonly used
chromatographie techniques might be of
value in not only evaluating the data
obtained from laboratory measurements of
effective porosity, but also in providing
a simple means of estimating breakthrough
times in solute transport situations.
21
-------
Sterie exclusion chromatography
(SEC) is used extensively In the polymer
industry to characterize products
according to their molecular size. In
this technique a column is packed with a
porous medium having both inter- and
intrapartlculate porosity, such that the
dead end pprea in the particles are of a
dimension comparable to or slightly
larger than the molecules of interest.
When a sample containing the polymer is
passed through the column, the smaller
molecules can diffuse into the dead end
pores, retarding their passage through
the column, while larger ones are con-
fined to the liquid moving in the inter-
particulate (interstitial) space. Thus,
the larger molecules elute from the
column first, having a smaller pore
volume available to them than did the
smaller molecules. In other words, the
effectiveporosity of the column is less
for the larger molecules than for the
smaller ones. The existence of this
branch of chromatography is proof that
the importance of the concept of effec-
tive porosity in solute transport needs
to be evaluated. This paper describes
early work in the application of chroma-
tographic theory to soil column experi-
ments with the object of measuring as
well as better understanding effective
porosity.
the injection of a small volume of sample
onto a chromatographic column. This is
analogous to a discrete contaminant spill
whereas a step function input would, for
example, simulate continuous input to a
clay liner from collected leachate. For
a given level of experimental precision,
information is more easily extracted from
a "peak" elution curve than a "front",
due to the existence of the peak maximum
and negative slope on the back side of
the peak. Once the desired parameters of
interest are obtained in the laboratory,
they can easily be used in a step func-
tion-input model to describe an actual
field situation.
The solution to the one-dimensional
transport model for pulse input was given
by Biesenberger and Ouano (1) for an
unretained conservative tracer. At the
column exit, the solute concentration, C,
as a function of volume of effluent col-
lected, ¥, is given by
AP1/2
c = «£H^ « (1)
where Vo is the void volume of the
column, A is the amount of solute in-
jected, P is a sort of Peclet number,
defined as
THEORETICAL BACKGROUND
The one-dimensional transport model
(ODTM) arises out of the theory of com-
bined diffusion and advection, usually
formulated in terms of solute flux
(Gillham and Cherry CO). Historically,
the chromatographic model grew out of the
random walk model of diffusion (Giddings
(3))- Although the distinction is arti-
ficial since both draw on the same physi-
cal principles from the engineering
literature, usage has diverged to the
point that investigators in the two dis-
ciplines of hydrogeology and chromatogra-
phy can hardly converse with one another
on this topic common to both fields. As
a result, both fields miss out on the
cross-fertilization process which is
essential to interdisciplinary science.
Because of the close analogy between
soil column experiments and liquid chro-
matography, the following discussion will
consider a pulse input function, such as
and
LV/DL
V/Vr
(2)
(3)
In these equations, L is the column
length, v is the average linear velocity,
i.e. the distance traveled divided by the
travel time, t
L/t
CO
and 0^ is the longitudinal dispersion . •
coefficient. Use of this equation for
evaluation of the dispersion coefficient
and effective porosity requires a curve
fit of equation (1) to the experimental
data.
22
-------
In the chromatographic approach
(Giddings (3)} the number of theoretical
plates, N, and the reduced velocity, vr,
are characteristic parameters of the
system. Historically, the theoretical
plate terminology arose from a comparison
between gas chromatography and distilla-
tion for separation of components of a
mixture. The reduced velocity is defined
as
vdp/D
(5)
where dp is the particle diameter and D
the bulK-phase diffusion coefficient of
the solute. The number of theoretical
plates for a particular column is given
by
5.54 (L/w)2
(6)
porosity as the linear velocity of the
fluid becomes very low. Under these
conditions molecular diffusion dominates
longitudinal dispersion, all pores suffi-
ciently large to allow entry by solute
molecules are explored, and a
single-porosity model should be adequate
to describe the system. Defined in this
manner, the effective porosity of a
column may .vary from one solute to
another for the same conditions, or from
one set of conditions to another for a
given solute. The utility of, such a
definition is that effective porosity is
a) directly identifiable with
average solute travel time in
saturated porous media,
b) easily conceived, and
c) it is directly measurable by
experiment.
with w = the peak width at half height.
The reduced plate height (dimensionless)
is defined by
h = L/Ndr
(7)
The quantity h is proportional to the
square of the ratio of peak width to
distance of travel, and is thus a measure
of dispersion. The so-called-
"van Deemter plot" (9) is a plot of h
versus vr or log h versus log vr and will
be discussed in more detail later.
In steric exclusion chromatography,
Janca (6) identified the first statisti-
cal moment or "center-of-mass" (CM) of an
elution peak with the average retention
volume, which is the average pore volume
"seen" by the component comprising the
peak. From this interpretation and the
identification of the volume in equation
1 with the void volume, we define the
dynamic porosity of a column, either soil
or chromatographic, to be the average
pore'-volume experienced by the solute
molecules as they travel through the
column. In a column experiment, this
volume porosity will be given by the CM
of the elution peak produced by an un~
retained tracer. We define the effective
porosity as the limit of the dynamic
METHODOLOGY
Samples of lacustrine clay, were
taken directly in 1,25" i.d. polycar.bo- -
nate.tubing- segments mounted in the
center of a modified split-spoon sampler.
The relatively undisturbed vertical
samples were taken to the laboratory and
sawn into 7" segments. A small amount of
soil material was removed from each end
in a glove bag under nitrogen atmosphere,
and end fittings containing a fritted
disk and tube fitting were inserted. The
column was placed in a frame to hold the
end fittings in place under pressure,-and
ground water from a well near the sample
collection site was forced through the
columns under 30 psig of pressure.
Flow was moni.tored using a glass
pipet and two electric eyes, monitored by
a computer. When the liquid level
reached the lower eye, a valve was opened
by computer to refill the pi.pet until the
top eye detected the meniscus of the
liquid. The computer used the time
elapsed between refills to calculate the
liquid flow rate, which was generally
between 0.04 and 0.11 mL per hour.
23
-------
Tritiated water was used as the
tracer in most experiments. Fractions of
column effluent were collected In weighed
vials, weighed again, scintillation
cocktail added and the samples counted in
the usual manner. A plot of activity per
unit volume versus elution volume pro-
vided the "chromatogram" which was then
analyzed using the above theoretical
approaches. At the end of the experimen-
tation the columns were weighed in their
(assumed) saturated state, then opened,
the contents dried to constant weight at
105°C, and weighed again. The weight
loss was corrected by the density of
water to give the total volume porosity,
for comparison with the effective
porosity.
Several experiments were also run on
columns which were packed with sand or
glass beads. Five different size ranges
were used, with the average diameter
ranging from 12 to ^62 \im between the
various samples. Flow rate was varied
over almost two orders of magnitude,
using an HPLC pump. Rhodamine WT and
tritiated water were used as the tracers,
the former being monitored by an HPLC
fluorescence detector. These experiments
were used to investigate the suitability
of chromatographlc interpretation of the
data.
PiOBLEMS BNCOUNTEBED
Only about one out of every four of
the olay columns continued to maintain
constant enough flow to warrant tracer
injection. Because the duration of an
experiment at these flow rates was two to
six months, it was difficult to keep
everything running throughout the entire
experiment, and data collection was very
slow. .Attempts to build a "minicolumn"
system (2" x 1/1" i.d. soil columns)' from
HPLC components to shorten the experiment
time essentially failed for reasons which
are not presently understood. Car-
bon1 ''-labeled bicarbonate was tried as a
tracer in a few experiments, but experi-
enced severe retardation on the clay
columns, making it unsuitable as a
porosity tracer.
RESULTS AND DISCUSSION
Figure 1 shows a typical elution
curve for tritiated water (HTO) from a
clay column. The center of mass of the
peak was determined to be the 50% isotope
recovery point. Recovery was 99 ± 3% for
column 1. The elution curve was fit by
the one-dimensional transport model using
the dispersion coefficient and the void
volume as fitting parameters. Table 1
shows the results of three different
experiments, and lists the value of
effective porosity determined by
curve-fitting and by center-of-mass de-
termination, compared with the gravi-
metrically-determined total porosity.
Table 1. Experimentally-iDetermined
'_ Porosities
Effective
Porosity*
Center
Column Test Curve of Total
no. no. fit mass porosity
1
1
2
1
2
1
0.11
O.M
n.d.
0.43
0.42
0.40
0,44
O.il4
n.d.
*Given as a decimal fraction. See text
for methods of determination.
Shown in Figure 2 is the van Deemter
plot for 105 experiments we have con-
ducted using sand and glass bead column
packings as well as four clay column
experiments. The characteristic shape is
seen, with the slope approximately equal
to -1 at low values of the reduced veloc-
ity, vp. The region below log vr - -1 is
that in which dispersion is dominated by
molecular diffusion. Under those condi-
tions, tracer molecules should be able to
explore essentially all of the pores
which are large enough to be accessible,
and thus is the region in which dynamic
porosity becomes essentially equal to
effective porosity. Although the sand
and glass bead data are spread over a
range of h (they are actually ordered
from largest to smallest particle
diameter, as also observed by Blackwell
(2)) the clay column data are reasonably
well clustered and lie at a point
24
-------
consistent with extrapolation of the data
(given by Blackwell (2) and Pfannkuch
(7)) obtained at higher vr, implying some
generality of the parameter h in describ-
ing dispersion at low reduced velocities.
Another manner of data representa-
tion, also used by Blackwell (2) and
Pfannkuch (7) is shown in Figure 3, where
log DL/D is plotted versus log vr. The
quantity DL/D has been calculated using
equation 8, derived using equations 9 and
10 from Horvath and Lin (5) and some
relationships from Giddings (3).
and
DL/D = hvr/2
DL = YD + AdpV/Z
h = 2Y/vr + 2A/Z
Z = (1 + yvr-1/3)
(8)
(9)
(10)
(11)
where Y is an "obstruction factor", A is
the "flow inequality parameter," and y is
an insensitive function of the porosity,
having a value in the range of 2.6-2.9.
Below vr s-1 log DL/D becomes roughly
linear, extrapolating to the log of the
diffusive part of dispersion at low vr.
DISCUSSION
At the low linear velocities used in
these experiments or encountered in
ground-water flow through similar
fine-grained media, the diffusional con-
tribution to advancement of the sol-ute
front is comparable to that of advection.
This is illustrated in Figure 1 where
breakthrough is seen five days after
injection, while the center of mass (cm)
emerges after 15.3 days. The implication
of the results obtained so far is that
under these conditions it is more impor-
tant to be able to predict dispersion
than to accurately know the effective
porosity. A simple method for predicting
breakthrough is suggested by equations 6,
7, and 10 in view of the reasonably dem-
onstrated validity of using the chromato-
graphic interpretation of the column
data.
To a first approximation, the elu-
tion curve in Figure 1 can be considered
to be a triangle or a gaussian distribu-
tion. In either case, a reasonable ap-
proximation of its width at the base is
twice the width at half height, so that
while the CM moves a distance L, the
advancement ahead of the CM due to dis-
persion is w, given by
(11YDAaL/Q)l/2
(12)
where a is the effective porosity, A is
the cross-sectional area, and Q is the
volumetric flow rate.
The arrival time tf at a point a distance
downgradient, relative to the arrival
time of the CM, tcm, is given by
tf.
t
cm
F ((1 + l)d/F)1/2-1)
2d
where
F = (1TYDaA/Q)l/2
For the example illustrated in Figure 1,
the calculated value for tf/tcm is 0.33,
corresponding to 5.1 days. At the end of
five days, the experimentally determined
effluent concentration of solute was
about one percent of the maximum value
which would elute later. On the previous
day the effluent concentration of tracer
was undetectable, while on the sixth and
seventh days it was 3% and 8%, respec-
tively, so the calculation yields a quite
good estimate of the breakthrough time.
In that calculation, a value of 0.7 was
used for Y (references 2, 1) and 7) and
the value 2.1) x 10~5 cm2/sec was used for
the bulk phase diffusion coefficient for
tritiated water (10). In cases where the
effective porosity is not known, the
total porosity appears to be a reasonable
approximation for small unretained
solutes at values of vr (see Table 1).
Where movement of the solute is known to
be retarded by interaction with the soil
material, a retardation correction will
be needed, such as that suggested by
Schwarzenbach and Westall (8).
25
-------
One situation in which knowledge of
the effective porosity would be espe-
cially important is that of facilitated
transport, in which a small molecule is
adsorbed to and carried "piggyback" on a
larger molecule or particle. There may
be considerable difference between the
pore volume available to a small molecule
and that accessible to a macromolecule or
small colloidal particle. In addition,
small organic solutes which normally
experience retardation by the soil may be
strongly enough adsorbed to the vehicle
molecule/particle that retardation does
not occur. Under these conditions break-
through (as indicated by analytical de-
tection of the solute in the effluent
stream) may be considerably more rapid
than even that of a small unretarded
molecule. This and the other ideas pre-
sented above require further study and
verification in order to be used in the
reliable estimation of solute break-
through times for fine-grained soils.
CONCLUSIONS
1) At low average linear velocity,
such as that encountered in
fine-grained sediments, the
effective porosity is identified
with the average void volume
explored by a tracer probe
injected onto a soil column.
This is represented by the first
statistical moment or
"center-of-mass" of the elution
peak.
2) For the soils tested, the
effective porosity with respect
to the trltiated water molecule
was within experimental error of
the gravimetrically-determined
total porosity.
3) Under the conditions of 1) it is
essential to include the effect
of dispersion in any estimate of
solute breakthrough times.
1) Chromatographio data analysis
techniques may be used to
estimate with a simple calcula-
tion the importance of disper-
sion in reducing breakthrough
times, without explicitly per-
forming a one-dimensional trans-
port calculation.
5) The effective porosity can
differ considerably from total
porosity in the case of facili-
tated transport.
ACKNOWLEDGEMENT
This work was supported in part by
the Hazardous Waste Engineering Research
Laboratory, USEPA, Cincinnati, OH,
Dr. Walter E. Grube, Jr., Project
Officer.
REFERENCES
1. Biesenberger, J. A. and A. Ouano,
1970. Extraparticle Diffusional
Effects in Gel Permeation Chromatog-
raphy. I. Theory., J. Appl.Poly-
mer Sci., v. 11, pp. H71-W2.
2. Blackwell, R. J., 1962. Laboratory
Studies of Microscopic Dispersion
Phenomena. Soo. Pet. Eng. J., 2,
pp. 1-8.
3. Giddings, J. C., 1965. Dynamics of
Chromatography, Part 1, Principles
and Theory, Dekker, New York.
«. Gillham, R. W,, and J. A. Cherry,
1982. - Contaminant Migration in
Saturated Unconsolidated Geologic
Deposits, in Recent Trends in Hydro-
geology, edited by T. N. Narasimhan.
GSA Special Paper 189, Geological
Society of American, Boulder, CO,
pp. 31-62.
5. Horvath, C., and H. J. Lin, 1976.
Movement and Band Spreading of
Unsorbed Solutes in Liquid Chroma-
tography. Journal of Chromatogra-
phy, 126, p. 401-420.
6. Janca, J., 1984. Sterio Exclusion
Liquid Chromatography of Polymers,
Dekker, New York.
7. Pfannkuch, H. 0., 1963. Contribu-
tion a L1etude des Deplacements de
Fluides Miscibles dans un Milieu
Poreux., Rev. Inst. Fr. Pet., v, 18,
p. 215-260.
26
-------
8. Schwarzenbach, R. P. and J. Westall,
1981. Transport of Nonpolar Organic
Compounds from Surface Water to
Ground Water. Laboratory Sorption
Studies. Environ. Sci. Technol.,
v. 15, pp. 1360-1367.
9. van Deemter, J. J., F. J. Zuiderweg,
and A. Klinkenberg, 1956. Longitu-
dinal Diffusion and Resistance to
Mass Transfer as Causes of
Nonideality in Chromatography,
Chem. Eng. Sci., v. 5, pp. 271-289.
10. Wang, J. H., C. V. Robinson, and I.
S. Edelman, 1953- Self-Diffusion
and Structure of Liquid Water. III.
J. Am. Chem. Soc., 75, pp.
466-470.
Center of Mass
a.
u
•H
>
3
r-t
U-l
W
34000
27200
20400
13600 -
6800 -
8/6/85 8/21/85 9/6/85 9/21/85 10/7/85 10/22/85
Date
Figure 1. Elution curve for tritiated water tracer, column 1, test 2.
27
-------
00
o
Figure 2.
5
4
3 .
Average
Column particle
Symbol no. diameter Material Tracer
R
R
R
R
R
T
T
T
T
T
-5
-4
Van Deemter plot of column dispersion data, using reduced variables.
Region I is Che area spanned "by the data of Pfannkuch (1963), while
Region II is calculated from the data of Blackwell (1962).
T » Tritiated water, R = Rhodamine WT.
Average
Column particle
nbol no. diameter Material 'Tracer
o
ec
-1
log v
Figure 3. Effect of mobile phase velocity on dispersion in column experiments.
Shaded area is the region spanned by the data of Pfannkuch (1963) and
Blackwell (1962). T = Tritiated water, R = Rhodamine WT.
28
-------
HYDRAULIC CONDUCTIVITY OF COMPACTED CLAY SOILS
Andrew S. Rogowski
USDA-ARS, Northeast Watershed Research Center
University Park, Pennsylvania 16802
ABSTRACT
To measure hydraulic conductivity a field-»scale research facility was constructed,
consisting of a 30 x 75 foot (9.1 x 22.9 m) area of clay soil compacted in three lifts to
specifications and with equipment used in constructing clay liners. Analysis of bulk
density, infiltration and leachate data showed that zones where infiltration and leachate
percolation were high were also the zones that were not fully saturated. The analysis
implied that flow may be taking place through preferential pathways in areas of lower
compaction density. The flow rates were proportional to the hydraulic gradient, in part
verifying the assumption that Darcy's Law was valid,, Changes in density following ponding
were expressed as volume of water needed to saturate the liner. Measured swelling of the
liner was very slight while overall average density increased. Infiltration and outflow
rates which have increased consistently following ponding have lately begun to decline.
INTRODUCTION
Compacted clay liners, installed in
landfills and other waste disposal sites to
prevent, or at least to minimize ground
water contamination by percolating
leachates, have been shown to have perme-
ability higher than demonstracted by labo-
ratory tests of the same soils (Daniel,
1984; Day and Daniel, 1985). The study
described here has been initiated to mea-
sure the distribution of hydraulic conduc-
tivity in a simulated liner constructed of
compacted clay soil using standard industry
procedure and equipment (Rogowski and
Richie, 1984). The facility. (Figure 1) was
instrumented to measure infiltration,
drainage and density. Results from proto-
type studies have shown that perforations
of compacted soil may result in preferen-
tial water movement paths. To avoid this
situation, aluminum access tubes were
placed horizontally to measure soil
density. Underdrains were used to collect
outflow, and infiltration cylinders were
installed to monitor infiltration rates.
Data obtained during construction showed
that although water content and density of
compacted clay met design specifications,
spatial variability of values was large.
Infiltration and drainage following ponding
were poorly predicted by the prototype
data, although initial observations showed
breakthrough of percolate near the confin-
ing walls. In this presentation results of
the first nine months of this study will be
discussed. The primary objective of this
research is to discover the relationship
between integrity of the liner as a whole,
and hydraulic conductivity of point,
field, and laboratory samples derived from
the same materials.
METHODS
Testing Facility
Figure 1 shows a field-scale facility
constructed to support the compacted clay
soil. The facility consists of a elevated
bridge-like reinforced concrete platform
supported by concrete beams which rest on
compacted level subgrade (Jacoby and
Rogowski, 1985). A 3 x 3 foot (ft) (0.91 x
0=91 M) grid of collection drains underly-
ing the compacted clay soil is complemented
by a 3 x 3 grid of 11 inch (in) (0.28 m)
diameter buffered infiltration cylinders
at the surface. Imbedded in the floor of
the platform horizontally across the
29
-------
Figure 1. Clay liner facility.
facility are the lower of the 24 aluminum
access tubes for the measurement of density
using Troxler^- dual gamma probe. Position-
ed on the clay surface and situated exactly
1 ft (0.30 m) above the lower ones are the
upper access tubes. The attenuation mea- .
surements are made on a 3 x 3 ft grid with
a gamma source (Cs^^) £n the lower tube
and the detector in the upper tube. The
degree of attenuation is used to measure
wet density in a pyramid-like slice of soil
between the source and the detector.
Figure 2 shows experimental measuring grids
used In the study. The origin is the
southwest corner of the platform.
Clay Liner
Clay soil used as the liner material is
a commercially available cherty silt loam
from central Pennsylvania. The soil is
predominantly illite and kaolinite with
some montmorillonite. It is classified as
a CL type with a laboratory permeability
of 1.10 to 1.76 x 10~° cm/sec determined in
a falling head permeameter at 1701 kg/rn^
compacted density, and a gradient of about
20. Details of installation, literature
review and preliminary results are given in
Sogowski and Richie (1984); Rogowski (1985);
and Rogowski et al. (1985),
Density
Following construction but prior to
ponding, 13 sets of wet bulk density were
taken With the dual gamma probe on a grid
shown in Figure 2a, and average available
pore space prior to ponding was calculated.
The comparison of changes in density fol-
lowing ponding, to the available pore space
prior to ponding if expressed in the same
units, should be indicative of the degree
of saturation attained in time by the liner
as a whole.
Clay Swelling
Since the liner material contains some
montmorillonlte changes in density could
occur as a result of swelling thereby
obscuring effects of infiltration. Swell-
Ing therefore was measured routinely. At
each site (Figure 2c) a double faced 6x6
inch (0.15 x 0.15 m) pedestal was placed
flat on the soil surface prior to ponding
The mention of trade names in. this publication does not constitute an endorsement of
the product by the U.S. Department of Agriculture over other products not mentioned.
30
-------
0123456769
0 1 23 456789
A
B
C
D
E
F
• G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
w
X
• 240 DUAL PROBE DENSITY
READING SITES
A
B
C
D
" F
G
H
1
J
K
L
M
N
O
P
Q
R
T
U
W
X
1 2 3 45 6789
A
B
C
D
E
F
G
H
I
j
K
L
M
N
O
P
Q
R
S
T
U
V
W
i
i
9
1 1
( 1
o
• 250 INFILTRATION CYLINDERS
AND UNDERDRAWS
® 35 CYLINDERS STARTED
INITIALLY
• 35 SITES TO MEASURE SWELLING
°35 SMALL EVAPORATION PANS
O 1 LARGE EVAPORATION PAN
Figure 2. Experimental .3 x 3 ft (0.91 x 0.91 m) grid'giving'locations
for measuring (a) bulk density, (b) infiltration and
drainage, (c) swelling and evaporation.
and changes in elevation to 0.01 mm were
monitored using an Optocator system.
Optocator system measures distances by
bouncing a laser beam off the surfaces.. In
our case a mobile scaffold was constructed
traversing on railings the entire length of
the platform. The laser beam gauge was
positioned.in predrtiled slots and distance
to 35 pedestals and 5 standard sites mount-
ed permanently on a solid backwall was
measured monthly.
Inflow
Conceptually clay liner facility is a
double ring infiltrometer. Three large 275
gallon (1041 £) tanks set up in series as a
giant Mariotte bottle maintain the water
level constant over body of the liner.
Grid of.individual 613 cm2 infiltration
cylinders (Figure 2b) is. served by 1 liter
(£) washbottles set up also as constant
head Mariotte bottles and adjusted to the
level of the water outside the cylinder.
The bottles are refilled to the mark daily
or weekly as needed. Cumulated amounts are
corrected for evaporation obtained by. aver-
aging readings from the 35 pans (Figure 2c)
which have an identical size, configuration
and geometry as the infiltration cylinders.
Evaporation correction for the liner as a
whole outside the rings is obtained from the
class A evaporation pan situated on the west
side of the platform (Figure. 2c) .
Outflow
The grid of individual leachate drains
(Figure 2c) stoppered with air vents, .is
connected to the outside of the platform
with 1/4 in (6.3 mm) ID rigid plastic tubes.
The outside drains empty into a collection
pipe. The pipe carries the effluent to a
sump 'which flushes everytime a 30 gallon
(114 £) mark is reached recirculating the
leachate for later application to the liner.
Leachate is measured by volume and cumulated
weekly.
Data Analysis
In the traditional analysis of vari-
ability, properties measured at a number of
sampling points within an area are assumed
spatially independent and are represented
by mean, standard deviation and an assumed
(or estimated) probability density function.
However, the assumption of spatial indepen-
dence in soils, at least for sampling
points close together seldom holds (Webster,
1985). Geostatiatical techniques (Journel
and Huijbregts, 1978), can then be employed
31
-------
to evaluate the extent of spatial depen-
dence. The spatial dependence of neighbor-
ing observations of a property measured as
a function of a distance vector h Is
expressed in geostatlstics by a semivario-
gram Y(h). A semivariogram describes the
average rate of change of Y(h) with
distance and shows the variance structure
of observations, and these observations
may or may not follow any type of probabil-
ity density function. Once the spatial
structure is defined by a semivariogram,
the sampling points can be kriged
(Matheron, 1971) to describe the distribu-
tion of a property over an area, in space
or in time. By using information on
spatial dependence from a semivariogram,
kriging produces a minimum variance (best),
unbiased linear estimate of a property at a
point and an associated value of estimation
variance. Kriging can thus be defined as
an estimation of an average value of a
property over an area (or volume) based on
geometry and spatial variability (informa-
tion contained in the semivariogram) of
samples, and subject to a minimum estima-
tion variance constraint. Geostatistical
estimation techniques such as kriging were
used by us here because they permit calcu-
lation of error associated with the
estimates, they consider spatial variabil-
ity of data through the use of semivario—
grams, and they assign weighting coeffi-
cients for a given distribution of sample
values that minimize estimation errors.
RESULTS AHD DISCUSSION
Childs (1969) observed that the choice
between laboratory or field methods of mea-
suring hydraulic conductivity is not merely
a choice between differing levels of
attainable precision, but rather matching
of a suitable method to a particular scale
of interest. To ensure the integrity of a
clay liner, measurements of conductivity
need to be made in the field. One reason
for this is that sampling, no matter how
careful, invariably alters structure, and
since conductivity is very strongly affect-
ed by structure, ensuing laboratory values
cannot be applied with confidence to the
field. The second reason is that Darcy's
Law (Darcy, 1856) applies to statistically
averaged materials, while laboratory K
values are those of a point sample. One
recommended solution is to Increase sample
size, or to take more samples. If the
structural units are large, sample size
may have to be considerable, and although
many small samples can provide a reproduc-
ible average, such average may miss
altogether important features and be gross-
ly in error. To investigate this dilemma
we have been monitoring changes in available
pore space, infiltration volumes and
leachate outflow from different sized areas
of compacted clay material.
Changes in Density
Changes in bulk density of a compacted
clay liner particularly if measured verti-
cally down over time are indicative of the
wetting front advance into the clay matrix.
Figure 3 shows the average bulk density of
a liner as a whole from start, through
ponding and at nine months of operation.
Initial preponding values represent 13
calibration runs between December and
March. The two months of data which follow
(April and May) appeared somewhat erratic
and have, led to probe repair (6/19/85) and
recalibration (8/23/85). Rise in density
following ponding reflects the dual probe
bias towards the surface as the wetting
front first penetrated the clay liner. The
data' show a continuing increase in density
suggesting that after nine months, ponded
water is still moving into the clay matrix,
It is expected that whenever effective
porosity becomes fully charged density
values will stabilize at a reasonably con-
stant value similar to the initial three
months data. When interpreting "stabilize"
and "reasonably constant" we need to
remember that although average density
fluctuations in Figure 3 appear large, in
reality they barely exceed statistical
precision of measurement. For example,
reading from left to right on 1/18/85,
average density was a low 2181 kg/m and on
1/27/85 it was a high 2202 kg/m3, this
translates to 6805 and 6581 counts per
minute (cpm) respectively or somewhat more
than the 200 cpm maximum difference
required between consecutive standards.
Consequently, when interpreting results
such as those shown in Figure 3 it is
necessary to look at the overall trend
rather than attach special significance to
Individual fluctuations.
Swelling
One possible source of fluctuations in
density could be the swelling of randomly
distributed tnontmorillonite clay minerals
within the liner matrix on wetting and
possibly moderate shrinkage as a result of
32
-------
12/84 1/85
Figure 3. Average bulk density (wet) of compacted clay liner in kg/m for
three and a half months before and nine months after ponding.
consolidation and piping in zones of lower
compaction. The mineral composition of the
clay liner material is primarily illite and
kaolinite with some montmorillonite. Thus,
although swelling could not be ruled out
(Grim, 1962, p. 248) little was expected.
It is also conceivable that in pockets of
inadequately compacted clay some consolida-
tion may occur. Figure 4 shows the kriged
(Journel and Huijbregts, 1978) changes in
surface elevation (mm) as measured with the
Selcoml Optocator unit. In general, the
changes at nine months amount to between
0.5 and 1.50 ram. There are however, two
zones of distinctly different behavior. On
the north side of the liner in the middle
there appears a small zone of consolidation
(-1,50 to -2.50 mm), while on the south
side to the left of center there is a cor-
responding small zone of expansion (2,00 to
3.50 mm). Although expansion appears to be
associated with a less permeable zone and
consolidation with a more permeable one,
there does not appear to be any significant
correlation between the degree of expansion
and observed conductivity.
Extent of Saturation
Comparison of changes in density after
ponding with available pore space prior to
ponding may be indicative of the degree of
saturation attained in time by the liner as
a whole. In Figure 5 this is expressed as
the amount of water needed to saturate the
liner at ponding, at 2 days, 3 months and
9 months after ponding. Inspection of
Figure 5 reveals a seemlingly rapid satura-
tion within a centrally located zone and
appearance of negative contours within the
same area after 9 months of ponding (d).
To understand what is taking place we need
to recall how the contours in Figure 5 were
generated. Figure 5a represents total
available pore space before ponding correct-
ed by the soil water content which ranged
between 300 and 330 kg/m3 (Rogowski et al.,
1985) at the time of compaction. The liner
was constructed in October 1984 and ponded
in late March 1985. Despite precautions,
certain amount of drying would have
certainly occurred during the intervening
five months, and 'the actual amount of water
in the clay matrix at the time of ponding
would have been lower than at the start.
This would lead to higher initial values of
the "amount of water required to saturate
the liner" and eliminate negative contours
in Figure 5d. Since water content distri-
bution at the time of compaction is based
on 108 regularized and kriged (Journel and
Huijbregts, 1978) observations of the three
lifts comprising the liner, it constitutes
a reasonably reliable primary data base.
Ultimately, when there is no further change
in density and assuming complete saturation
CO-contour), negative contours (such as the
ones in Figure 5d) will indicate the
33
-------
FFFF 5D BS
FFF 2 C B
FFP! » C BB
l«11HII«in« KHHtlH
HH8HH3HH3HHBHHHH8
RHHHKHRBillfHHIfHKHHnmi
KflHHHHHHHttKUHHHHHHPHH
BBSS
KffHKWHH
SGQS'IGGSG
UK NUBHHfl HH HUHHUHWR
II iUHMIIHH HHHIUf
Ell HKHH if mini?
HII HH HHHHrm
HHMflHHtllf3tH'1HMN!r
IIHHHHHJMIKH
SISSS™
ISSS,
""5??
IQ-A
itRHHHli
nnno'jG
G,-.»«U
HKIEK
HIWHR
CCCG'SC nilHHHEtHHKBH fiGGGS ^i~,GG
GfiaCRG HnKMHKHHHHIIH CCGGff fifi
S RimHKHtlHHHHilHKHHHK
£ HHH
TZ HHHHH
HFtHIt HUB II IXIIXIZ1IXX Witt
G 1C 11 I JJJ JJJ I FtH 1GG'*.?C^E
HUH IIHH WUCCCGGS
HHifHH HHM III Htt GGSGG3GGG
MBHH1I 8HHS 2J IIS GSSCGS-GCa
r
iGf.:: i
oS.wieG6csP
f-.GGGI
51
»
Figure 4. Elevation changes In milimeters (nun) for the compacted clay
liner 9 months after ponding; positive symbols indicate
swelling,, negative symbols are indicative of shrinkage,,
MM. -2.50 to -2.00 mm
BBBB -2.00 to -1.50 mm
CCCC -1.50 to -1.00 mm
DDDD -1.00 to -0.50 mm
EEEE -0.50 to 0.00 mm
MTF 0.00 to 0.50 mm
GGGG 0.50 to 1.00 nun
HHHH 1.00 to 1.50 mm
I1II 1.50 to 2.00 mm
JJJJ 2.00 to 2.50 mm
KKKK 2.50 to 3.00 mm
LLLL 3.00 to 3.50 mm
location and amount of water lost from the
liner by evaporation between the time it
was constructed and five months later when
it was ponded.
Outflow
The average outflow from the liner in
Figure 6 is divided into two parts. The
upper curve represents the average of con-
tributions from the outer 66 drains around
the perimeter of the facility, while lower
curve gives the average for the inner 184
drains. The outer drains are in the area
next to the sides and endwalls that have
received less then adequate compaction,
while the Inner 184 drains represent well
compacted zones. The outflow rate both
from the Inner and outer drains rose
steadily during the first seven to eight
months after ponding and then began to
decline. Vertical dashed line on the left
Indicates the time when automatic sump was
installed to cumulate outer drain outflow
and recirculate the water, while the right
dashed line gives the time when all drains
were vented. Installation of the sump
smoothed out the outer drain data, while
venting of the drains had no apparent
effect on the average rate of outflow.
Neither action seemed to initiate or aid
the decline of observed outflow rate.
Inflow
A similar situation appears to have
prevailed in Figure 7 among the cylinder
inflltrometers. Initially, only 35 out of
250 cylinders were activated, seven of
those were in poorly compacted material
next to the sidewalls (Figure 2) and 28
were within a well compacted zone. In
June '85 all 184 cylinders within a well
compacted zone were activated while perim-
eter cylinders were measured only
occasionally. Although installation of
the sump and recirculation of water had no
direct apparent effect on drains it sub-
stantially reduced the infiltration rate
In the cylinders, primarily because of
increased concentration of salts in recy-
cled water. However, neither implementa-
tion of the balance weighing technique
(which replaced volumetric measurements)
nor venting of the drains had any effect on
the observed infiltration rate. After
reaching a plateau in 7 to 8 months infil-
tration rate also begun to decline.
The decline in infiltration rates co-
.incides with the time when approximately
one pore volume of water on the average has
passed through the liner,, This suggests
that before effective porosity is filled, a
more rapid rate represents combined Inflow
34
-------
0 12 24 36 48 60 72
i i i i i i i 1
I I 1 1 ! 1 1 I 1 1 I I
N
0 IS 24 36 48 60 72
0 12 24 36 48 60 72
DISTANCE, feet
Figure 5« Kriged (Jcmrnel and Huijbregts", 1978) distributions of
the amount of water needed (in kg/m ) to saturate the
liner before ponding (a), and 2 days (b), 3 months
(c), and 9 months (d) after ponding.
35
-------
Outwr Drains (66)
Sump
\
LJ_
8/85
TIME
Figure 6. Average measured outflow rate from inner 184 and outer 66 leachate
drains during the nine months after ponding (3/26/85) ; times of
occurrence for installation of the sump to collect outflow from
outer drains, use of balance to weigh small amounts of percolate,
and venting of the lines are given "by vertical dashed lines.
into the day matrix and flow through the
liner. Once the effective porosity is
filled the infiltration is confined pri-
marily to large pores that exit at the
bottom of the liner. While this may be a
plausible explanation for decline in infil-
tration rate it does not satisfactorily-
explain the observed decline in outflow,
the cause of which may be physical—
temperature effects, or clogging of
pores—, chemical—localized swelling--!-., or
microbiological reactions under ponded
anaerobic conditions.
Hydraulic Conductivity
Based on the infiltration rate after
nine months and measured gradient values,
Figure 8 shows the kriged (Journel and
Huijbregts, 1978) distribution of hydraulic
conductivity in the compacted clay liner.
The values range over three orders of
magnitude from 10"-' to 10~8 cm/sec. While
lowest values of K are observed on the
'south and east side, highest values prevail
at the western end. There appears to be no
particular correlation with the distribution
of density, the amount of water required to
saturate the liner, nor the distribution of
swelling and shrinkage locations„ It is
expected that post-ponding dissection and
analysis of the clay liner will explain
causes of these differences.
There are several obvious possibili-
ties. Hydraulic conductivity (K) is
affected by temperature, ionic composition
of water, and presence of entrapped air
(Bouwer, 1978)„ The effects of temperature
on K are primarily due to the effect of
temperature on water viscosity. The lower
the temperature the higher the viscosity of
water and the more difficult it would be
for water to move through pores of the clay
36
-------
I
I 60
Initial (7)
/Sump
3/85
6/85
8/85
TIME
9/85
Figure 7. Average infiltration rate measured with cylinder infiltrometers for
nine months after ponding; dashed lines indicate times when sump
was installed, balance weighing was implemented, and drains were
vented; initial (28) and (7) pertain to 28 inner and 7 outer
cylinders started initially (Figure 2b), inner (184) pertain to all
184 inner cylinder infiltrometers started in June 19850
matrix, this would be reflected in depress-
ed K, infiltration, and outflow. A 50 per-
cent increase in viscosity will reduce the
value of K, infiltration and outflow by
1/3. Summer and early fall temperatures
were often in excess of 20 to 25°C, while
during the winter months the temperature
was maintained at 5°C0 This decrease in
ambient temperature could in part account
for observed decreases in infiltration and
outflow rates.
The effect on hydraulic conductivity
of cations in solution passing through a
soil could be significant. In general, for
a given salt concentration (Na+, Ca++ and
Mg"*^" expressed as SAR—sodium adsorption
ratio) hydraulic conductivity will decrease
with increasing Na+ content, but increase
for a given SAR with increasing concentra-
tion of solution. The SAR values of
leachate effluent and ponded water are
quite low (2 to 3) and little effect is
expected. However, we have observed,
.particularly after the sump was installed,
differences in Na+ and Ca~"~ concentrations
in ponded water and in leachate from
selected drains<> Increased Na+ saturation
of the clay matrix could increase disper-
sion of clay decreasing hydraulic
conductivity., while increased Ca"*"1" adsorp-
tion could lead to better flocculation and
increased hydraulic conductivity values.
Superimposed on the system appears to be
also an anaerobic denitrification with
reduction of Fe' serving as an energy
source for microorganisms. More soluble
Fe is then readily leached out,
37
-------
1CT5
Figure 8. Distribution of hydraulic conductivity in compacted
clay liner nine months after ponding.
oxidizing, precipitating and clogging pores
when it comes in contact with air.
Clay liner compacted to engineering
(Proctor) specifications constitutes an
unsaturated system that contains much en-
trapped air when, ponded. As the water
moves into the clay matrix entrapped air
will become either displaced or reach an
equilibrium with pore water. The amount of
air dissolved in pore water and ambient
pressures developed, will vary and be high-
ly dependent on temperature. In general,
cold water can hold more air than warm,
consequently a drop of temperature could
lead to shrinkage of entrapped air vacuoles
and localized Increase In hydraulic
conductivity. If, however, cold water,
such as ambient water added to the liner in
winter, enters warmer clay matrix, air may
go out of solution leading to air blocking
or binding and decreased hydraulic
conductivity.
SUMMARY AND CONCLUSIONS
One—foot thick clay liner constructed
on an elevated platform to specifications
and with the equipment used by the industry
was ponded, and infiltration, leachate,
density, and swelling were monitored at a
large number of locations over a period of
nine months. The data were analyzed using
geostatistical techniques.. Preliminary
density data suggest that despite precau-
tions some drying of the clay liner has
occurred between the time it was compacted
and ponded. While in general little or no
change of surface elevation was apparent
over the liner as a whole, at two small
locations modest swelling and shrinkage
were observed. Inflow and outflow rates
increased for seven to eight months after
ponding and then begun to decline. Measured
distribution of hydraulic conductivity over
an area ranged from 10~8 to 10~5 cm/sec«
ACKNOWLEDGMENTS
The author acknowledges the support of
this study by the Land Pollution Control
Division, Hazardous Waste Engineering
Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio,
through Interagency Agreement No. DM129-
303-03-01-0 with the Northeast Watershed
Research Center, Agricultural Research
Service, U.S. Department of Agriculture,
University Park, Pennsylvania; Dr. Walter
E. Grube,.Jr. is the U.S. EPA Project
Officer. Special thanks and recognition
go to the USDA-ARS personnel involved in
different aspects of this project, E. L.
Jacoby, Jr., D. E. Simmons, W. M. Yazujian,
J. E. Donley, B«, J. Chamfaerlin, C. W. Artz,
38
-------
P. J. Dockey and students D. E. Gedon,
R. M. Petery and J. D. Dietz.
REFERENCES
1. Bouwer, H., 1978. Groundwater
Hydrology. McGraw-Hill Book
Company, New York, 480pp.
2. Childs, E. C., 1968. The Physical
B_asis of Soil Water Phenoinea.
Wiley Interscience, John Wiley &
Sons Ltd., New York, 493pp.
3. Daniel, D. E., 1984. Predicting
hydraulic conductivity of clay -
liners. Journal ,°f_Gept_echnical
Engineering, 110(2), pp285-3QO.
4. Darcy, H,, 1856. Les Fontaines
Publiques de LaVille de Dijon.
Paris, Dalmont, 647pp.
5. Day, S. R. and D. E. Daniel, 1985.
Hydraulic conductivity of two
prototype clay liners. Journal of
Geqtechnlea1 Engineering,111(8),
pp957-970.
6. Grim, R. E., 1962. Applied, JTLay
Mineralogy^. McGraw-Hill Book
Company, New York, 422pp.
7. Jacoby, E. L., Jr. and A. S.
Rogowski, 1985. Clay Liner
Fa c i 11ty Lo_gbo ok . 0SDA-ARSt.
Northeast Watershed Research
Center, University Park,
Pennsylvania 16802.
8. Journal, A. G. and Ch. J. Huijbregts,
1978. Mining Geostatisti.cs.
Academic Press, New York.
9. "Matheron, G., 1971. The Theory of
jRegionalized Variables and its
Application,, Ecole Nat.ionale
Superiere des Mines de' Paris, Paris,
France, 211pp.
10. Rogowski, A. S. and E. B» Richie,
1984. Relationship of laboratory
and field determined hydraulic
conductivity in compacted clay
soils. Proceedings of the Mid-
Atlantic Industrial Waste
_Con_ference, University Park,
Pennsylvania, pp520-533. ,
11. .Rogowski, A, S., 1985. Effectiveness
of a compacted clay liner in prevent—
ing--grouiid water contamination.
Proceedings_ of the 5th National
SytBposium and Exposition on Aguifer
^Restoration and Ground Water
"Monitoring, Columbus, Ohio, pp412-
429.
12. Rogowski, A. S., B. E. Weinrich, and
D. E. Simmons, 1985. Permeability
assessment in a compacted clay liner.
PrQceedings^of the 8th Annual Madison
Waste Conference on Municipal and
jjftdusjErjLal Wa_stet University of
Wisconsin, Madison, pp315-336.
13. Webster, H., 1985. Quantitative
spatial analysis of soil in the'
field. _In B0 A. Stewart (ed.)
Advances in SoilScience, Springer- .
Verlag, New York.
39
-------
GEOTECHNICAL ANALYSIS FOR REVIEW OF DIKE STABILITY
Mark. S. Meyers, Richard M. McCandless, and Andrew Bodocsi
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
The structure and capabilities of a user-friendly, interactive computer program
developed for the stability analysis of dikes (GARDS) are described. The GAROS program
is designed to guide a geotechnical nonspecialist (EPA regulatory personnel) through the
customary steps of earth dike analysis considering slope stability, settlement, lique-
faction, hydraulic flow and pressure conditions, and piping. The program was developed
under the sponsorship of the U. S. Environmental Protection Agency and therefore empha-
sizes Hazardous Haste applications although it is suitable for general use. The GARDS
package is designed for use on the IBM-PC/XT microcomputer.
The significant differences between GARDS and other slope stability programs
commonly available are the opportunity to perform all of the above analyses in a single
user-friendly, Interactive package; an internal automatic search routine to determine the
critical failure surface for both rotational (slip circle) and translational (wedge)
stability analyses; an internal, automatic search routine to locate zones of greatest
liquefaction potential and to compute total and differential settlements of foundation
soils; an internal finite element hydraulic analysis to determine the steady state
piezoraetric surface through the section, including the case of an impermeable barrier
such as a clay liner; the ability to model excess pore pressure conditions produced by
confined steady state flow and evaluate slope stability and uplift conditions resulting
therefrom; and the ability to determine the critical exit gradient and the potential for
piping failure.
User documentation consists of a combined Handbook/User's Manual (under development)
which presents basic theory, program operational procedures, and example long hand and
computer solutions for each analysis.
INTRODUCTION
Geotechnical Analysis and Review of
Dike Stability (GARDS) is a user-friendly,
interactive computer software package
developed under the sponsorship of the U.S.
Environmental Protection Agency, Land
Pollution Control Division, for the geo-
technical analysis of dikes at hazardous
waste sites. The program is designed to
guide a geotechnical non-specialist through
the customary steps of earth dike analysis,
considering slope stability, settlement,
liquefaction, hydraulic flow and pressure
conditions, and piping.
GARDS was developed to meet an
expressed need for a geotechnical software
tool to evaluate existing and planned
earth dike structures at hazardous waste
facilities. Moreover, the tool was to be
designed for use by regulatory personnel
with a technical, but not necessarily
geotechnical, background. The GARDS
package is not an "expert system," but is
suitable for use as a design review tool
where data describing site conditions and
soil parameters have been supplied by a
geotechnical professional.
The program was developed to meet the
40
-------
following specific technical and logistical
objectives:
1. It was to be user-friendly, inter-
active, and take a step-by-step
approach in explaining the required
data for analysis.
2. It was to have internal data check
capabilities.
3. It was to have user-friendly error
diagnostic features.
4. The adapted programs for the various
analytical tasks should be written in a
common programming language, compatible
in terms of input/output functions, or
at least amenable to efficient modifi-
cation.
After an extensive software search,
it was determined that no programs which
were ideal in all respects existed in the
public domain. Two available programs
which were judged to meet the technical
requirements and which were sufficiently
compatible were REAME (Rotational Equili-
brium Analysis of Multilayered Embankments;
Y.H. Huang, University of Kentucky) and
SEEP (hydraulics analysis; K.S. Wong and
J.M. Duncan, Virginia Polytechnic Insti-
tute). REAME was selected as a base
program to which SEEP and all other program
blocks (developed in-house) were adapted.
PROGRAM FEATURES
As developed, the GARDS software
package consists of several distinct
blocks. Each block is a program in itself
and is called by the user through the use
of menu displays. The main command struc-
ture of the GARDS program is as follows:
1. Control Block
2. Input Block
3. Input Data Check and Edit Block
4. Hydraulics Analysis Block
5. Slope Stability Analysis Block
6. Settlement Analysis Block
7. Liquefaction Analysis Block
8. Summary Output Block
The following paragraphs present a
brief description of each of the major
program blocks.
Control Block
The control block is the first block
encountered by the user. It gives an
introduction to the program and other pro-
gram blocks. Its primary function is to
call the appropriate analysis block
selected by the user from a "command menu"
displayed on the screen. The menu is a
list of the input/output and analytical
options available to the user.
The Input Block •
The input block is the first block to
be selected by the user, and offers three
user-friendly options:
1. Initial data entry with text explana-
tions and graphical examples
2. Initial data entry without text
explanations
3. Data entry using an existing input data
file
The first input option is designed to
guide the occasional user through the
initial data entry procedure. Explanations
and graphical examples of the types of
required input data are displayed on the
screen prior to data entry. The option to
review any or all explanations prior to
data entry is available to enable the user
to proceed at his own pace.
The second input option contains all
of the features of option one, except that
the text explanations and graphical
examples are not included. This option
would be selected for initial data entry by
the user who is experienced in the use of
GARDS and is therefore familiar with the
definitions of the required input data.
The third option is selected by the
user who has previously entered all data
using options 1 or 2, has saved it on a
data file, and elects to use the analysis
features of the program at a later time.
A listing of the data files available to
the user is displayed. The data is
assigned to the appropriate program vari-
ables internally. The use of this option
eliminates the repeated entry of basic soil
property and geometric data if the user
wishes to re-analyze a particular dike
section under different conditions.
An internal data check is performed
automatically with all input options to
41
-------
verify that values of all input variables
fall within specified ranges. If the value
of a data item is not within the specified
range, an error message will prompt the
user to enter a value within the specified
range. Once notified, the user has the
option to override this error trap routine
if warranted.
Input Data Check and Edit Block
After having used one of the options
of the input block, the input data check
and edit block provides for general verifi-
cation and modification of input data. The
user may check or edit all or selected
portions of the input data in any existing
data file. Since, however, input and
internally generated data routinely are
output as part of the Summary Output Block
(described later), the primary function of
this block is to edit portions of existing
data files in order to model different
conditions for subsequent runs. This block
is also routinely used to conduct a visual
check of geometric continuity of the
section before proceeding with an analysis.
Hydraulics Block
The hydraulics block automatically
determines the coordinates of the piezome-
tric surface for any of six idealized
hydraulic boundary conditions. The compu-
tations for the hydraulics block are made
at the end of the input block so that the
piezometric surface data is available for
subsequent analyses. This block utilizes
SEEP, a finite element based program for
solving steady-state problems of free
surface or confined flow of water in a
two-dimensional porous region. A support
program was developed to Internally
generate an input data file for SEEP from
the general user-input data file. Another
support program was developed to utilize
the output from SEEP to determine the
location of the critical exit gradient and
to relay the amount of seepage through the
dike section to the summary output block.
The potential for an uplift pressure
failure of an impermeable barrier such as
clay liner is also determined and noted in
the summary output listing.
The six hydraulic boundary conditions
comprise four typical unconfined steady
state seepage conditions and two cases of
confined seepage involving an impermeable
barrier (clay liner) and excess pore
pressure on the barrier. The six cases are
illustrated in Figures I through 6 and are
briefly described in the following para-
graphs. The illustrations and narrative
descriptions for each condition describe
the internal analysis which is performed
automatically within the hydraulics block.
In all cases, a user-defined piezometric
surface is available as an option to model
unusual hydraulic conditions.
Figure 1 illustrates the Deep Static
hydraulic condition for the case of an
existing or proposed dike with or without
an adjacent excavation (landfill cell). In
either case the controlling criterion is
that the groundwater table is everywhere at
or below the existing or modified ground
surface. As such, the condition does not
represent a seepage case per se, and there-
fore does not involve rigorous hydraulic
analysis (use of the SEEP program).
The second hydraulic boundary condi-
tion involves shallow unconfined steady-
state seepage which daylights on a natural
or adjacent cut slope (into a landfill cell
for example). Figure 2 illustrates the
range of the Shallow Seepage condition.
For this analysis, the piezometric surface
is assumed to be coincident with the ground
surface below the seepage daylight point on
the slope. Since drawdown of the ground-
water table in the vicinity of the daylight
point is not assumed, a "worst case" slope
condition is modeled.
The Free Pool Seepage condition illus-
trated in Figure 3 is the third unconfined
seepage model. It is similar in principal
to the Shallow Seepage case (Figure 2),
but assumes a stationary upstream pool and
development of steady-state free flow
through the dike and into an adjacent
excavation (if any). Using this hydraulic
model, the user can investigate the sta-
bility of the dike under different pool
conditions, with or without an adjacent
excavation.
Figures 4 and 5 represent the two
confined seepage conditions involving the
development of excess pore pressures within
the slope and uplift pressure on the base
of an impermeable confining layer (clay
liner). In both cases, it is assumed that
the stability of the section under Shallow
Seepage (Figure 2) or Free Pool Seepage
conditions (Figure 3) is satisfactory (pre-
viously determined). A seepage analysis
42
-------
Dike C e n t e r 11 n e
C t y ple a I)
117
-y*.
i
JU,
~-wl
o r
wit
' o r
X
thout 8 d J a
natural
h a d J a o • n
natural
I \
•% %
o o n t cell \%vs.
slope ^ \
t o * II — -4
slops \
i
__2L
-JL
— Y.
V-
™Jr
Bedrock Surface (typical)
Figure 1. Range of Deep Static Condition.
i
I
I • d J « c • n t u n I I n • d coll
or natural slope
JZ_
Figure 2. Range of Shallow Seepage Condition.
43
-------
static
upstream pool
^34
without
adjacent cell
"" X. ° r natural
.^r •"••__,")
"^. ^«"»\ r"~~" seepaae surface ^ 1 /
X ^L<7 ^-X\l
V'^ rTT
with adjacent unllned call T seepage
or natural slope — H surface '~*~
Figure 3. Range of Free Pool Seepage Condition.
plezomotrlc
pressure surface
over lined cell
uplift pressure
Figure 4. Range of Shallow Confined Seepage Condition
44
-------
static
upstream pool
plezometrlc
pressure surface
over lined cell
.. —L... —-J..—
seepage surface
uplift pressure
Figure 5. Range of Confined Pool Condition.
is performed within the SEEP subroutine
to define the coordinates of the steady
state piezometric surface through the
section. An uplift check is then performed
within the Hydraulics Block to determine
the adequacy of the liner either alone or
with a waste surcharge. If results indi-
cate that the liner is stable under the
prevailing piezometric conditions, the user
may then use the piezometric surface data
for analysis of slope stabilty. If liner
check results are unsatisfactory, slope
stability analysis is considered irrelevant
for the hydraulic conditions modeled.
The final boundary condition automa-
tically modeled in the Hydraulics Block is
the Drawdown Pool condition shown in Figure
6. This condition is equivalent to the
conventional "rapid drawdown" analysis on
the upstream slope of a dike. The analysis
assumes undrained slope conditions immedi-
ately after the abrupt drawdown of an
adjacent pool which has existed long enough
to establish a steady-state free flow
surface through the dike. The SEEP program
is used to compute the coordinates of the
piezometric surface through the dike.
Results are then automatically modified to
reflect pool drawdown to the ground surface
(as shown) before the hydraulics data is
transferred to the slope stability block
for stability analysis.
Slope Stability Analysis Block
The slope stability block uses data
defining the piezometric surface from the
Hydraulics Block to perform either of two
conventional slope stability analyses. The
first is a rotational (circular slip sur-
face) analysis which uses the Simplified
Bishop Method of slices. The second is a
translational (plane slip surface) analysis
based upon analysis procedures published in
an engineering design manual developed by
the U.S. Department of the Navy (NAVFAC
DM-7.1, May 1982).
The major features of the rotational
analysis include:
Dike slopes of any configuration and
made up of up to 19 soil layers may be
analyzed;
Seepage conditions are modeled in the
form of the steady state piezometric
surface through the section as
determined in the Hydraulics Block;
Seismic effects may be included;
The program has a radius control
feature which produces trial circles
that fall between specified limits
defined by the soil layer geometry;
45
-------
lowered
Downstream Condition*
Confined pool
seepage .surface
t t f t T f
-A ^^/JAVVV^ „ . „ „ l~ „ ^^ „ ~~^^
f «w^^«5'>^NN^5"OO^w^X>\y55'\\NXyfS'XV^
-------
EXAMPLE MOTIONAL SIABILltV ANALVSIS File! EXAMPLE
Hydraulic Condition 5: Confined Pool Seepage
Consolidated Drained Soil Parameters
Factor of Sifets : 1,58, Failure Center: ( 483 , 188 ), Radius! 35
. SCftU ,; 1 ineh = 28 fee.t
Figure 7. Typical hard copy output for Rotational Stability.Analysis.
EXAMPLE lEAiLAIIOHAL SIMIUIV AHAUSIS
Hydraulic Condition 5: Confined fool Seepage
Consolidated Brained Soil Parameters
Faotor of Safetg = 1,55
File; EXAWLE
.« 41B ,
. SCALE.: 1 inch - 28 fee.t
Figure 8. Typical hard copy output for Translational Stability Analysis.
47
-------
soils due to stresses caused by the weight
of an overlying dike. The settlements are
calculated using Boussinesq's Theory and
Terzaghi's One-Dimensional Consolidation
Settlement Theorem. Required soil para-
meters for each soil include unit weight,
initial void ratio, compression and recom-
pression indices, and the over-consolidation
ratio.
Settlements are calculated at the
toes, crest points, and center-line of the
dike. The consolidation of each soil is
calculated for each layer and summed up for
all soils to determine the total settlement
at each point. Differential settlements
are calculated between each toe and crest,
toe and center!ine, and crest and center-
line.
Liquefaction Block
The liquefaction block determines the
potential for liquefaction of the dike and
foundation soils. The procedure used,
developed by Seed and Idress (1983),
applies mainly to site conditions Involving
sand or silty sand strata occuring below a
level ground surface. The analysis uses
soil parameters routinely obtained during
site exploration and subsequent laboratory
testing. Site seismic parameters may be
estimated using published guidelines or may
be obtained from site-specific investiga-
tion (preferred). The analysis uses an
automatic search routine controlled by the
section geometry to locate the area of
highest liquefaction potential. Results
are reported as a tabulation of soil layers
and points having a factor of safety which
is less than E.50.
Summary Output Block
The summary output block allows the
user to obtain a hard copy of the input
data and the results of all analyses run
for the dike section under study. The
critical factors of safety, failure circle
center coordinates, radius, and plane
failure line segment coordinates are all
highlighted in the output listing, along
with the critical differential settlements,
liquefaction potential, and critical exit
gradient. If an analysis was not run, such
1s indicated in a summary table at the end
of the output listing.
SOFTWARE SUPPORT
Software support for GARDS consists of
a combined Handbook/User's Manual (now
under development). The Handbook section
is intended to give the user a background
sufficient enough to understand the basic
principals underlying the various analyses
that are being performed. A separate
chapter is devoted to each analysis in-
cluded in the program. After a brief
introduction stating the purpose of the
analysis, explaining when the analysis
should be used, and listing the capabili-
ties and limitations of the program, a
section of the basic theory and the method
of solution are presented. Following this,
a complete step-by-step long-hand example
solution is given. To provide continuity,
the same dike section is used for all
analyses. A complete computer solution of
the analysis of the dike section is pre-
sented in the User's Manual section for
comparison.
The User's Manual section includes a
discussion of the capabilities, limita-
tions, and essential details of each of the
different blocks, along with a complete
interactive example for the user to study.
A sketch of the example section and all
required data are provided, along with a
summary output listing for the example
section. Block flow charts, a program
listing, and other pertinent material are
appended to the Handbook/User's Manual.
When completed, the GARDS package
will be available pending EPA review and
approval.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the
assistance and contributions of Mr. Doug
Ammon, EPA Technical Project Monitor, Mr.
Philip Cluxton, Research Associate, and
graduate research assistants James Albertz,
Brian Barney, Douglas Keller, and Steve
Scott.
BIBLIOGRAPHY
Bishop, A.W. The Use of Slip Circle in the
Stability Analysis of Slopes. In:
Geotechnique. 1955. Vol. 5, No. 1,
pp. 7-17.
48
-------
Bureau of Reclamation. Design of Small
Dams, 2nd Ed. United States Government
Printing Office, Washington DC, 1973.
Garstens, M.R., and G.D. May. Graphs for
Locating the Line of Seepage in an
Earth Dam. In: Civil Engineering,
ASCE, August 1967.
Cedergren, H.R.
Flow Nets,
York, 1977.
Seepage, Drainage, and
John Wiley & Sons, New
Department of Navy. Design Manual, Soil
Mechanics, Foundations, and Earth
Structures, NAVFAC DM-7, Naval
Facilities Engineering Command,
Philadelphia, 1971.
An Engineering Manual for the Evaluation of
Dike Stability. EPA Permit Writers'
Training Course. 1983.
GEOSLOPE, Slope Stability Analysis Program.
GEOCOMP Corporation, Concord, MA,
1984.
Hazardous Waste Containment Areas. EPA
Dike Stability Workshop. US Army
Engineers, WES, Vicksburg, MS, 1983.
Huang, Y.H. Stability Analysis of Earth
Slopes. Van Nostrand Reinhold Co.,
New York, 1983.
Huang, Y.H. User's Manual: REAMES, A
Simplified Version of REAME in Both
BASIC and FORTRAN for the Stability
Analysis of Slopes. Institute for
Mining and Mineral's Research,
University of Kentucky, Lexington, KY,
1982.
Permit Applicant's Guidance Manual for
Hazardous Waste Land Treatment,
Storage, and Disposal Facilities,
Final Draft. EPA Office of Solid
Waste and Emergency Response. 1984.
EPA 530 SW-84-004.
Seed, H. Bolton, et al. Evaluation of
Liquefaction Potential Using Field
Performance Data. Journal of Geo-
technical Engineering, Vol. 109,
No. 3, March, 1983.
Seed, H. Bolton, et al. Simplified Proce-
dure for Evaluating Soil Liquefaction
Potential. Journal of the Soil
Mechanics and Foundations Division of
the American Society of Civil Engi-
neers, Vol. 97, No. SM9, September,
1971.
Wong, K.S., and J.M. Duncan. SEEP: A Com-
puter Program for Seepage Analysis of
Saturated Free Surface or Confined
Steady Flow. Department of Civil
Engineering, Virginia Polytechnic
Institute, Blacksburg, VA. February
1985.
49
-------
THE ROLE OF WATER BALANCE IN THE LONG-TERM STABILITY
OF HAZARDOUS WASTE SITE COVER TREATMENTS
Fairley J, Barnes, John C. Rodgers, and George Trujillo
Los Alamos National Laboratory
Los Alamos, New Mexico 87545
ABSTRACT
After the 30-year post-closure maintenance period at hazardous waste landfills,
long-term stability must be assured without continued intervention. Understanding water
balance in the established vegetative cover system is central to predicting such stability.
As part of a US EPA funded research project a series of experimental cover treatment
plots were established on a closed waste disposal site. The effects of such critical
parameters as soil cover design, leaf area index, and rooting characteristics on water
balance were determined under varied conditions. The results show consistent differences
fn soil moisture storage between soil profiles and between vegetation cover treatments.
Data from these experiments are being used to examine the utility of water balance models
(such as CREAMS and HELP) for predicting soil moisture storage in landfill covers.
INTRODUCTION
Long-term stability of hazardous
waste landfill covers depends on design
and construction factors, and also on both
biotic and abiotic factors in the period
following closure and post-closure mainte-
nance. The process of designating a region
as suitable for land disposal of hazardous
waste, and predicting the criteria neces-
sary for successful operation and post-
closure stability can be greatly facili-
tated by the use of simulation modeling.
Ideally, models can predict the effects of
average environmental conditions as well
as climatic extremes on the stability of
the cover system.
Strategies to maximize the stability
of landfill covers must consider the charac-
teristics of the applied soil profile and
the vegetative cover. Modeling the hydro-
logic balance of the cover system is an
important tool in the design process.
Such models facilitate estimating erosion
rates and seepage of water into the under-
lying hazardous waste or into the drainage
layer associated with a cover system that
incorporates an impermeable membrane liner.
However, typical models now in use {CREAMS,
Knisel 1980; HELP, Schroeder et al. 1984)
have been developed using data from agri-
cultural systems consisting of deep fertile
soils, monocultures of cultivated annual
crops, and controlled irrigation inputs.
The evaluation of such models using
data from field scale studies on landfill
covers has rarely been addressed. As part
of a Los Alamos National Laboratory pro-
ject, the validity of using hydrologic
models to predict some components of site
water balance on actual landfill cover
systems is being investigated. Several
combinations of soil and vegetation have
been established as experimental cover
treatments on a closed waste disposal
site. Two of the treatments have been in
place for three years as part of a DOE
funded project (Hakonson 1986) and soil
moisture data at different depths are
available over this time period. Addi-
tional cover treatments were planted for
the purposes the current EPA funded pro-
ject (Rodgers et al. 1984) with soil mois-
ture and vegetation data available for 1985.
50
-------
In this portion of the study, the aim
was to determine if a close correlation
between model predictions of average soil
moisture and observed soil moisture could
be achieved. Water balance models have
typically been evaluated in terms of their
ability to predict runoff volume, which
along with deep percolation, are critical
factors in the performance of landfill
cover systems. However, in this project
the evaluations are based on the pre-
diction of soil moisture and evapotranspi-
ration, which are strongly dependent on
the vegetative component of the cover
treatment.
FIELD EXPERIMENTS
The experimental site ("Area B") is a
2.4 ha landfill at Los Alamos, NM. Los
Alamos is at an elevation of 2200 m, with
an average annual precipitation of 46 cm.
Over 60% of the annual precipitation results
from summer storms of high intensity and
short duration, with the winter precipita-
tion occurring mainly as snow.
Area B slopes towards the southeast
with a grade of about 5%. It is one of
the oldest hazardous and radioactive waste
sites at Los Alamos and has had several
past and recent remedial treatments. It
presents the widest range of closure treat-
ments of any disposal site at Los Alamos.
In 1982, vegetation which had invaded the
site was removed, and a new standard cover
treatment installed in two areas, consisting
mainly of crushed tuff (85 cm) and top
soil (15 cm). A small section of the
site, between the two standard profile
areas, has a cobble-gravel biobarrier
installed as an experimental profile (Figure
1).
The topsoil used in the construction
of the Area B landfill cover remedial
treatment is a local Hackroy sandy clay
loam (Nyhan et al. 1978). The crushed
Bandelier tuff can be classified as a
sandy loam. From soil pits excavated
into the standard cover treatment, it was
found that the soil profiles differ
significantly between the two standard
profile areas ("east" and "west") and
varied slightly between the plots estab-
lished on each area (Figure 2). Compared
to the west plots, the east plots have a
much higher proportion of soil relative
to crushed tuff, and thus the east soil
profiles have significantly higher water
holding capacity.
In this report, data and modeling
results from only two vegetation treat-
ments on the two standard soil profile
areas (i.e. four vegetation/soil combi-
nations) is presented. Data and analysis
for the biobarrier soil profile and addi-
tional vegetative cover treatments will
be presented in our final report.
In 1984 Area B was disked to remove
existing vegetation which consisted of a
mixture of native grasses and forbs which
is the usual revegetation treatment used
at Los Alamos. Of the experimental plots
(24 x 8 m) plots established, two were
maintained free of vegetation ("bare")
(west plot 13 and east plot 2) and two
vegetated with shrubs (west plot 11 and
east plot 3) (see Figure 2). The shrub
plots were planted densely with 2 year
old rabbitbrush (Chrysothamnus nauseosus)
in a regular design (1 plant/0.
In 1985, high spring precipitation
resulted in a heavy growth of weedy grass-
es and forbs on all plots. Vegetation
was sampled for biomass and leaf area
indices (LAI) in June 1985. LAI is the
total projected leaf area of vegetation
per unit area of ground and is an im-
portant parameter for the determination
of evapotranspiration by vegetated surfaces.
WEST STANDARD PROFILE
BIOBARRIER
EAST STANDARD PROFILE
15
14
I
10
DATA FOR THESE PLOTS ANALYSED IN TEXT
Figure 1. Plan view of "Area B" waste disposal site with
experimental plots. ,. .
51
-------
After sampling, all plots were weeded and
maintained as either bare soil or dense
rabbitbrush plantings. Estimations of LAI
and biomass measurements for forbs and
grasses from 1983 to June 1985 are based
on the June 1985 measurements and on obser-
vations of the seasonal growth patterns of
the mix of herbaceous plants at Area B.
To estimate LAI and biomass of the
shrub cover nondestructively, regression
relationships between crown canopy parame-
ters (diameter, height, volume) and total
crown biomass, leaf biomass and leaf area
are being developed using shrubs harvested
from designated areas at Area B and else-
where on Los Alamos National Laboratory
lands. The size of the shrubs on the
plots is measured periodically, and the
regression relationships used to estimate
biomass and LAI on each plot. Estimated
LAI of the vegetative cover (both shrubs
and weeds) of the plots for 1985 is shown
in Figure 3. Note that the growth and
subsequent removal of weeds resulted in a
high LAI in spring and early summer with a
subsequent sharp drop in LAI after the
plots were weeded.
Each plot has 3 or 4 neutron moisture
probe access tubes installed to a depth of
100 cm or more. Soil moisture distribution
with depth has been measured periodically
with a neutron moisture probe (Campbell
Pacific Model 503) throughout the year.
The measurement frequency was not great
enough to follow each precipitation event
and thus represents average trends in soil
moisture.
Volumetric soil moistures averaged
over the assumed rooting depth of 76 cm on
the four soil/vegetation combinations
during the year showed pronounced differ-
ences between plots as well as a consist-
ent declining soil moisture content (Fig-
ure 4). Soil profiles with a high soil
content (east plots 2 and 3) had higher
soil moisture than west plots 11 and 13,
which have a distinctly layered profile
with a thin layer of sandy clay loam as
top soil. From May to August, plots with
dense rabbitbrush planting have total
profile soil moisture decreasing at a rate
of 0.073 cm/day (plot 3) and 0.069 cm/day
(plot 11), a more rapid change in soil
moisture than on the bare plots where the
rates were 0.042.cm/day (plot 2) and
0.054 cm/day (plot 13). This suggests that
the shrub cover was more efficient than
the weedy herbaceous cover in removing
water from the soil profile after the
shrubs started their annual growth period
with increasing temperatures in May. Such
high transpiration rates even under con-
ditions of low water availability is con-
sistent with preliminary measurements we
have made on the water use physiology of
rabbitbrush. This is an indication of a
limitation in utilizing LAI as a simple
predictor of transpiration rates.
HYDROLOGIC MODELING
The USDA model CREAMS (Knisel 1980)
and the EPA model HELP (Schroeder et al.
1983) are being assessed as predictive
management tools for calculating the
WEST
PLOT 13
BARE
WEST
PLOT 11
RABBITBRUSH
SCL
MIXTURE
(SL : SCL
2: 1)
DEPTH
OF
PROFILE
LAYERS
»0 CM
'•15 CM
•30 CM
EAST
PLOT 3
RABBITBRUSH
SCL
SL
MIXTURE
(SL : SCL
1:4)
EAST
PLOT 2
BARE
SCL
SL
MIXTURE
(SL : SCL
1:5)
SCL - SANDY CLAY LOAU SL - SANDY LOAU
Figure 2. Soil profiles of the four experimental plots at Area B.
52
-------
05-
0.4-
x
W
02-
0.1-
0.0
SHRUB EAST
- SHRUB WEST
- BARE EAST
•• BARE WIST
WEED REMOVAL
JFMAMJJASOND
MONTH
30-
O
25-
i
20-
15-
10
•EAST BARE
EAST SHRUB
WEST BARE
WEST SHRUB
JFMAMJJASOND
MONTH
Figure 3. Estimated leaf area indices (LA!)
on four experimental plots in
1985.
water balance of the vegetation/soil cover
treatments. Initially, a series of simul-
ations have been conducted to explore the
capability of the models to reproduce
observed profile-averaged soil moisture
trends from 1983 to 1985 on the west con-
trol plot 11 which has a dense rabbitbrush
cover. Plot 11 is the only plot on a
standard soil profile which has a three-
year record of soil moisture measurements.
Preliminary simulations with both
models showed that although the water
storage routing algorithms are very simi-
lar in the two models, the seasonal dynam-
ics in soil water storage seemed to be
better represented by CREAMS simulations,
possibly due to differences in the calcu-
lations of evapotranspiration between the
two models. Consequently, the values of
key parameters were optimized using the
CREAMS model, and then the simulations
were repeated using these optimized param-
eter values in HELP simulations.
Water balance of a site is represented
one-dimensionally in these models by the
simplified water balance equation:
%•'-
- ET - L
ds
where gr = time rate of change in soil mois-
ture, p = precipitation, Q = runoff, ET =
evapotranspiration, and L = seepage or per-
colation.
Figure 4. Volumetric soil moisture av-
eraged over the rooting depth
on four plots at Area B dur-
ing 1985.
Soil water fluxes into and out of
the soil profile are calculated using a
seven layer representation of the profile
from the surface extending through the
rooting zone of the vegetative cover.
Initial responses to precipitation are
calculated on a daily time-step using a
modification of the SCS curve number
model (Knisel 1980).
In this study, the basic structure
and initialization of the model was estab-
lished using known site specific data,
actual daily precipitation totals for
1983-1985, and 20 year averages of air
temperature and solar radiation. LAI in
1983 and 1984 was estimated so as to
agree with observed seasonality in growth
patterns and LAI levels attained by herba-
ceous weeds in 1985. For 1985, estimates
of LAI were based on field measurements
on shrubs and herbaceous weeds. Instead
of using recommended values for soil
hydrologic parameters of the appropriate
soil texture classes represented on our
site (Knisel 1980, Lane 1984, Schroeder
et al. 1984) several simulations tested
the effects of varying the parameters
over the reported ranges on predicted
average soil moisture content.
First, average values of certain soil
parameters (saturated hydraulic conduc-
tivity, field capacity and wilt point)
were used as recommended by Lane (1984)
53
-------
and Shroeder et al. (1984) for the soil
texture classes represented in the plot 11
soil profile. Second, the ranges sug-
gested for the soil texture classes (Lane
1984) were tested. Third, the effect of
using the saturated hydraulic conductivities
of the specific soils on the site (Abeele
1984) was tested. Finally, after deter-
mining which parameter values maximized
the fit between observed and predicted
retention and drainage of soil water, the
curve number was adjusted to obtain the
best possible representation of infil-
tration of precipitation from summer storm
events.
The performance of the model in pre-
dicting water balance on plot 11 at Area B
was then assessed. Predicted average soil
moisture was compared with observed
average soil moisture. The observed
values were averaged over three meas-
urement depths (20, 40 and 60 cm) for each
neutron access tube, and over four tube
locations on the plot to give one mean
soil moisture value for each measurement
time. Best fit of the predicted to the
observed soil moisture patterns was as-
sessed by the standard technique of regres-
sion analysis with aim of maximizing the
correlation coefficient (r2) and opti-
mizing the slope and intercept of the
regression line to approach the equal
values line (Pathak et al. 1984). The
parameters that produced the optimum
CREAMS simulation were the minimum satu-
rated hydraulic conductivity of any layer
(RC=2.5 x 10~8 m/s for Hackroy sandy clay
loam); the maximum field capacity (34%)
and average wilting point (15%) for sandy
clay loam; and the average field capacity
(19%) and average wilting point (8%) for
sandy loam. The curve number used was
95. These same parameters were then used
to initialize a 3 year simulation of plot
11 using the HELP model (version 1),
The optimum CREAMS simulation and
the corresponding HELP simulation are
shown in Figure 5. When the predicted
and observed soil moisture were compared
point for point by regression analysis,
large differences (under or over-predic-
tions) were observed (Figure 6). Using
OBSERVED AND PREDICTED SOIL MOISTURE 1983-1985
OBSERVED
CREAMS PREDICTED
360 540 720
TIME (JULIAN DATE)
900
1080
Figure 5. Volumetric soil moisture on west plot 11 from 1983-1985, showing both observed
data, and predicted values from HELP and CREAMS simulations.
54
-------
Y =
CREAMS PREDICTED VERSUS OBSERVED
SOIL MOISTURE 1983-65
EQUAL VALUES LINE
0.154 + OJ269X
r* - 023
0.30
u
0!
O
£0,20
8
Io.l5
0.15 050 025 0.30
OBSERVED SOIL MOISTURE (cm/cm)
E0.10
HELP PREDICTED VERSUS OBSERVED
SOIL MOISTURE 1963-65
0.10
EQUAL VALUES LIKE-
Y = 0,073 4- 0.362X
r* = 0.34 '
0,15 OJ30 0,25 0.30
OBSERVED SOIL MOISTURE (em/cm)
Relationship between observed soil moisture data and soil moisture as predicted
by CREAMS and HELP water balance modeling, on west plot 11 from 1983 to.1985.
the correlation coefficient as a measure
of overall goodness of fit, only a por-
tion of the variation in soil moisture is
explainable by the algorithms in the
models (23% for CREAMS, 34% for HELP).
This implies that a large proportion of
the variability in soil moisture results
from processes either poorly represented
or absent in the models. It is very
clear from both Figures 5 and 6 that HELP
grossly underestimates soil moisture
throughout the 3 year simulations. This
is due in part to the inability to ini-
tialize soil moisture in the standard
version of HELP to field measured values,
and in part to the apparently excessive
ET estimates predicted by HELP which
result in soil moisture depletion to a
very low level.
In the process of testing the effects
of parameter values in CREAMS, it was
found that with a higher value of hydrau-
lic conductivity (at more usual recom-
mended levels) soil moisture stored in
the profile at the start of the simulation
rapidly drained, resulting in a steep
drop in average soil moisture far below
the observed values. It is possible this
results from the model's initialization
procedure which allows one to specify a
profile averaged soil moisture, but not a
layer-by-layer soil moisture profile.
This short coming in the model is a prob-
lem only when a very short (1 or 2 year)
simulation is used (Lane, 1984) as in the
present report.
With a low hydraulic conductivity as
in Figure 5, initialization of soil mois-
tures is quite accurate, but the low
hydraulic conductivity results in dimin-
ished dynamics of predicted soil moisture,
with predicted soil moisture variations
which are very flat compared to the ob-
served pattern. This is particularly
obvious in fall 1983 (days 335-360),
summer 1984 (days 540-600), and winter
1984-85 (days 700-800). In the former two
cases, greater runoff or ET must have
occurred than was predicted by CREAMS. In
the latter case, lateral flow of water
into the site may have occurred under
conditions of high precipitation and poor
drainage.
PREDICTION OF SOIL MOISTURE
The overall predictive power of the
models was tested by assessing what per-
centage of the variability in the observed
data was explained by model predictions
using the optimized values of soil hydro-
logic parameters in one year simulations
on the remaining three soil/vegetation
combinations (plots 2, 3 and 13). Each
plot had soil profiles with varying per-
centages of sandy clay loam and sandy loam
soils (Figure 2). For each plot, profile
values for hydrologic parameters were
calculated as weighted averages of the
optimum field capacities and wilting
points already determined. LAI data spe-
cific for each plot were used (Figure 3).
The driving variable was 1985 daily pre-
-------
cipitation totals. The first field meas-
urement of soil moisture in 1985 was used
to initialize profile soil moisture in
CREAMS simulations. The HELP model ini-
tializes soil moisture to an intermediate
value appropriate to each profile.
The overall predictive power of the
two models as measured by the correlation
between predicted and observed soil
moisture values for the three simulations
(Figure 7) shows that 63% of the variation
in soil moisture was explained by water
balance modeling by CREAMS and 40% by
HELP. The slope of the CREAMS regression
line (0.558) shows that CREAMS underesti-
mates high soil moisture and overestimates
low soil moisture, thus underestimating
the seasonal range in soil moisture. How-
ever, the CREAMS modeling represents the
actual soil moisture conditions below the
three soil/vegetation covers much more
accurately than the HELP modeling.
The agreement between observed and
predicted soil moisture was highly vari-
able (e.g. Figure 8). The dynamic pattern
and range of soil moisture variation in
the field was best represented by CREAMS
simulations on plot 3. The soil profile
on this plot most closely resembles the
uniform agricultural soils with high
field capacities that were used in the
development of the model. In contrast, the
plot most poorly represented by CREAMS
modeling (plot 13} has a shallow surface
layer of soil over a thick sandy layer
that may well allow lateral drainage. If
lateral drainage does occur, it violates
the assumptions of one dimensional vertical
flow of soil water inherent in both models.
DISCUSSION
The results presented raise many
issues over the use of these water balance
models for the representation of long-term
soil moisture trends. Estimation of soil
parameters as recommended by the documen-
tation for these models may be inappropri-
ate for soils at landfill sites. The
variability inherent in natural soils may
be increased by the construction process.
Uncertainties and variation in the degree
of compaction of soils can contribute
greatly to changes in the rate of infiltra-
tion of precipitation, and probably also
in the effective saturated hydraulic conduc-
tivity of the profile.
The hydraulic conductivity is a sensi-
tive parameter for controlling retention
of water in the profile. The recommended
values (Lane 1984, Schroeder et al. 1984)
allowed excessive drainage to occur at
high moisture contents. Decreasing the
hydraulic conductivity improved the dynamics
of drying cycles. However, the recommended
curve numbers resulted in excessive infil-
tration into the profile with consequent
excessive soil moisture after a precipi-
tation event. Choosing a high curve number
3050
=0.25
CREAMS PREDICTED VERSUS OBSERVED
SOIL MOISTURE 1885
- EQUAL VALUES LINE
0.16 O20 025 0.30
OBSERVED SOIL MOISTURE (cm/cm)
0.30
i
925
S a20
•J
g 0.18
0.10
a.
HELP PREDICTED VERSUS OBSERVED
SOIL MOISTURE 19B5
0.10 0.15 0.20 Q2S 0.30
OBSERVED SOIL MOISTURE (cm/cm)
7. Relationship between observed soil moisture and soil moisture as predicted by
CREAMS and HELP water balance modeling, on plots 2, 3 and 13 in 1985.
56
-------
AREA B SHRUB PLOT 3
AREA B BARE PLOT 13
35'
30-
O
25-
u
K
1
O
20-
15-
10
CREAMS PREDICTED
OBSERVED
35-
30-
O
25-
20-
15-
10-"
OBSERVED
^CREAMS PREDICTS
JFMAMJJASOND
MONTH
JFMAMJJASOND
MONTH
Figure 8. Observed soil moisture on plots 3 and 13 in 1985, and predicted soil moisture
from CREAMS modeling of the plots.
forced higher runoff, and further in-
creased the fit between observed and
CREAMS predicted values. These extreme
values of curve number and hydraulic
conductivity can only be validated by
further instrumenting the field site to
measure runoff, as well as the determi-
nation of hydraulic conductivities of the
different soil profile layers.
Surprisingly, CREAMS and HELP are
not constructed so as to allow.alteration
of curve number (CN) during simulation
runs. Establishment of a vegetative
cover is a dynamic process, and changes
in the surface conditions may continue
for many years as a result of succes-
sional processes of both vegetation and
soil microorganisms, and maturation of
the vegetation. A high CN was chosen to
force high runoff. However, data from
successive years may suggest that the use
of a lower CN would be more appropriate
as vegetation matures.
Calculation of evapotranspiration in
CREAMS and HELP depends on LAI, soil
moisture, and rooting depth. Leaf area
indices are time consuming to determine
in the field. Since successional pro-
cesses will almost certainly result in
gradual establishment of native plants on
landfill covers, it is necessary to have
data on both LAI and seasonality of growth
patterns of native plants. Yet such data
are not readily available and the docu-
mentation for the models gives no guidance
in this area.
Native species, particularly from
semi-arid and arid regions, may have very
different transpirational responses to
decreasing soil moisture (Lane and Barnes
1986). The rate of decline in soil mois-
ture in the spring and summer (Figure 4)
clearly demonstrates that the shrub plots
had higher canopy transpiration than a
comparable LAI of the weedy mixtures on
the bare plots. Thus LAI may prove to be
an inadequate representation of ET from
diverse vegetation covers.
Assessing the utility of water bal-
ance models for predicting the performance
of hazardous waste sites will require
further validation studies. In addition,
the development of data bases on hydro-
logic characteristics of compacted soils,
and on the transpiration characteristics
of native plant covers in various climatic
regimes will be required for successful
field applications.
ACKNOWLEDGEMENTS
This project was funded by the U. S.
Environmental Protection Agency under IA6
No. DW 89930235-01-1. Ms. Naomi P.
Barkley was the Project Officer. We thank
B. Drennon, W. Herrera, and L. Martinez
57
-------
for expert field assistance.
REFERENCES
1. Abeele, W. V, 1984. Geotechnical
aspects of Hackroy sandy loam and
crushed tuff. Los Alamos National
Laboratory Report LA-9916-MS. Los
Alamos, New Mexico.
2. Devaurs, M. 1985. Use of CREAMS
model in experimental designs for
shallow land burial of low level
wastes. Waste Management '85, March
27, 1985. Tucson, AZ.
3. Hakonson, T. E. 1986. Evaluation of
geologic materials to limit biological
intrusion into low-level radioactive
waste disposal sites. Los Alamos
National Laboratory Report LA-10286-
MS. Los Alamos, NM.
4. Knisel, W. G., Ed. 1980. CREAMS: A
field scale model for chemicals,
runoff, and erosion from agricultural
management systems. US Department
of Agriculture, Conservation Research
Report No. 26. Washington, DC.
5. Lane, L. J. and V. A. Ferriera,
1980. Sensitivity Analysis. Pp
113-158 in W. G. Knisel, Ed. CREAMS:
A field scale model for chemicals,
runoff, and erosion from agricultural
management systems. US Department
of Agriculture, Conservation Research
Report No. 26, Washington, DC.
6. Lane, L. J. 1984. Surface Water
Management: A user's guide to calcu-
late a water balance using the CREAMS
model. Los Alamos National Laboratory
Report LA-10177-M. Los Alamos, NM.
7. Lane, L. J. and F. J. Barnes, 1986.
Water balance calculations in south-
western woodlands. Pinyon-Juniper
Conference, Reno, NV, January 1986.
USDA Forest Service, in press.
8. Nyhan, J. W., L. W. Hacker, T. E.
Calhoun, and D. L. Young. 1978. Soil
survey of Los Alamos County, New
Mexico. Los Alamos Scientific Labora-
tory Report LA-6779-MS. Los Alamos,
9. Pathak, C. S., F. R. Crow, and R. L.
Bengtson. 1984. Comparative perform-
ance of two runoff models on grassland
watersheds. Transactions of the ASAE
27(2): 397-406.
10. Rodgers, J., T. S. Foxx, G. D. Tierney,
F. J. Barnes. 1984. Maintenance free
vegetative systems for landfill covers.
Interim Report EPA Project DW 89930235-
01-1.
11. Schroeder, P. R., A. C. Gibson, and M.
D. Smolen. 1984. The hydrologic
evaluation of landfill performance
(HELP) Model. US Environmental Pro-
tection Agency 530-SW-84-010, Cincin-
nati, OH.
58
-------
FACTORS INFLUENCING DIFFERENTIAL SETTLEMENT AND THE
EFFECTS ON COVER SYSTEMS
Paul A. Gilbert
USAE Waterways Experiment Station
Vicksburg, MS 39180
ABSTRACT
Differential settlement can cause cracking in the cover layer of a hazardous waste ;
landfill. If such cracking occurs, there is significant danger that the cover will be
breached, releasing toxic material into the environment. Thus, there is a need for a
reliable method to identify conditions which will result in excessive differential
settlement, cover system cracking, and the ensuing threat to the environment.
The present work will utilize beam theory to identify material and geometric parameters
which significantly influence differential settlement. It will be shown how the identified
parameters may be used to control and minimize the effect of differential settlement.
An analytical procedure will be presented which allows estimation of tensile strain
developed during differential settlement. Additionally, a design procedure based on this
estimated tensile will be presented.
INTRODUCTION
Settlement analyses are generally made
for projected construction- to predetermine
whether excessive settlement is likely to
occur and cause damage. Generally permis-
sible deformation is controlled by the type
of construction rather than by soil
behavior. In many cases, the criteria for
establishing allowable settlement are out-
side the realm of soil mechanics. For
example the stresses induced in a structure
by the deformation of the foundation con-
stitutes one of the criteria for deter-
mining allowable settlement in a founda-
tion. Since foundation settlement is
usually nonuniform, it is important to
distinguish between maximum settlement
and differential settlement. Differen-
tial settlement is often more important
since it produces distortion and stress
concentrations in surface installa-
tions and structures.
The rigidity of a structure will deter-
mine the amount of differential settlement
which it may withstand without distress.
Generally, soil structures are much more
flexible and therefore less susceptible
to damage than the more rigid frame,
masonry or concrete structures. In the
present study, the structures of interest
are the cover systems of hazardous waste
landfills and the foundations on which they
rest are the hazardous wastes contained in
those landfills. By understanding the
nature of these two interacting elements
and the mechanics involved in differential
settlement, cover systems may be designed
to minimize the effects of cover cracking
and all the associated dangerous
consequences.
BACKGROUND
Settlement by consolidation is impor-
tant in clays. It is usually not a problem
in cohesionless soils (sands and gravels)
because the compressibility of such soils
is relatively small. Differential settle-
ment occurs in a soil structure when an
element of soil consolidates or compresses
differently than that of a neighboring
element, resulting in nonuniform vertical
59
-------
deformation within that structure. Such
nonunifoua deformation occurs as a result
of inhomogeneity within the foundation
soil. Because of the nature of the con-
tents and the construction of a hazardous
waste landfill, it is probable that the
landfill will be highly inhomogeneous and
therefore susceptible to differential
settletmnt. Although differential settle-
ment is virtually impossible to entirely
eliminate, its consequences may be mini-
mized by first, taking actions to minimize
total settlement, then designing flexi-
bility into the cover system to minimize
damage due to differential settlement.
THEORETICAL CONSIDERATION
Damage and distortion can occur in
tie cover system of a hazardous waste
landfill if foundation support is lost as
a result of internal differential settle—
pent. In such a case, the cover system
would be required to "beam over" the zone
of lost support. For this reason it was
decided to formulate a model to determine
the important factors involved in differ-
ential settlement using elementary beam
theory. The model assumes that the cover
system will lose support over a length,
L , and as a result will undergo a dif-
ferential settlement, A . The model
representation is therefore a beam with
fixed supports at either end and is dis-
torted when one support settles an amount
A . See fig. 1.
Expressions for the vertical shear,
moment, slope and deflection of the
idealized cover may be determined using
elementary beam theory (1). the maximum
stresses due to moment and shear are
determined from theory to be
shear
•(«(*) <•>(!)' <»
and
where
moment
(3)
(A)
0 , .= maximum stress due
shear, moment . .
to shear, moment
E = Youngs modulus of
the cover material
h » thickness of the
cover
Although these expressions were
developed using small deflection beam theory
and may not be entirely appropriate for the
large deflections observed in soil struc-
tures, the expressions identify parameters
which quantify distress caused by differen-
tial settlement. For example equations (1)
and (2) suggest that stress is minimized if
d/L , E and h/L are minimized. Obviously
A/L is minimized if the differential
settlement, & , is minimized. This may be
accomplished by minimizing total settlement
and involves compacting wastes during place-
ment, eliminating void space within the
landfill, stabilizing liquids before drum
disposal, etc. (2). Additionally A/L may
be minimized by maximizing L . This will
reduce cover stress by spreading the distor-
tion over a greater distance. L may be
maximized by placing the landfill wastes as
homogeneously as possible to give uniform
support to the cover. Minimizing the Youngs
Modulus, E , of the cover material can be
accomplished by compacting the cover soil
wet of the optimum water content and will
result in a cover with lower strength but
with greater pliability and capacity to
distort without rupture. An additional
"free" benefit of the wet of optimum com-
paction is a lower cover permeability.
Finally, the equations suggest that the
ratio h/L should be minimized. This
should be done by maximizing L . A
thick cover is necessary to control diffu-
sion as well as to prevent the intrusion
of animals and plant roots Into the land-
fill (2). A thick cover also offers the
advantage that it is more resistant to
desiccation due to its large mass and
thickness.
To conclude, the effects of differen-
tial settlement on a cover system may be
minimized by compacting the waste fill dur-
ing placement to eliminate voids and weak
compressible zones. Additionally, the cover
should be compacted met of the optimum
water content to maximize pliability and
make the cover less susceptible to rupture
under the differential settlement which
does occur.
TENSILE STRAIN IN THE COVER
The cover system is stretched and
undergoes tensile strain as differential
settlement occurs in a landfill and the
60
-------
cover will crack if this tensile strain
becomes excessive. Soils are, in general,
not able to sustain high levels of tensile
strains without cracking. The average
tensile strain developed within the cover
may be computed from the beam model
presented above. Mathematical details
of this derivation will not be presented
in this work, however the procedure
involves integrating over the deflected
beam shape and requires numerical evalu-
ation since closed form integration of
the required mathematical expressions is
not possible. The result of this inte-
gration is seen in fig. 2 which shows
how the average tensile strain increases
as the parameter A/L increases. If
the maximum tensile strain which can be
resisted by a given soil is determined
then the maximum £/L which can be
tolerated in a cover system of that soil
may be estimated from fig 2.
Figure 3 is a plot of maximum
tensile strain reported by several
investigators (3,4) versus soil plastic-
ity index. It shows roughly that the
capacity for tensile strain .increases as
the plasticity index of the soil.
Increases. Obviously for completeness,
more research is needed on the tensile
capacity of soils, but the trend of
fig. 3 Is clear. From the treatment
described above, it may be reasonable to
conclude that a soil with a high tensile
strain capacity is preferable to one with
a lower capacity and therefore the selec-
tion of a soil with a high plasticity
index for a cover system is indicated.
It should be stated that if a soil is to
be used for a cover system, it may be
desirable to perform tensile tests on
that soil in the laboratory at several
water contents to determine the water
content which offers the best combina-
tion of high tensile strain and ease of
placement.
RECOMMENDATIONS
Based on analyses presented above,
the following general recommendation .
to minimize the effects of differential
settlement are believed warranted:
1. Wastes should be placed and
compacted in a landfill so as to achieve
maximum reasonable density and homo-
geneity. Soft zones and zones of high
void concentration should not be placed
in close proximity to well compacted
stiff aones.
2. The landfill cover should be
compacted at a water content wet of the
optimum water content.
3. If a choice of materials for a
cover soil is possible, the soil having
the highest plasticity index should 'be
selected if other factors permit.
4. If possible, a laboratory study
should be performed on the material
selected as a cover soil to determine the
water content which will facilitate cover
placement and which will sustain adequate
tensile strain for the differential settle-
ment anticipated.
REFERENCES
1. Crandall, S. H., and Dahl, N. C. 1959.
An Introduction to the Mechanics of
Solids, McGraw-Hill, Inc., New York.
2. Murphy, W. L., and Gilbert, P. A. 1985.
Settlement and Cover Subsidence of
Hazardous Waste Landfills, EPA-600/2-85/
035, Hazardous Waste Engineering Office
of Research and Development, USEPA,
Cincinnati, Ohio.
3. Leonards, G. A., and Narin, J. 1963.
Flexibility of Clay and Cracking of
Dams, Journal of the Soil Mechanics
and Foundations Division, American
Society of Civil Engineers, Vol 89,
No. SM2.
4. Tschebotarioff, G. P., Ward, E. R., and
DePhilippe,. A. A. 1953. The Tensile
Strength of Disturbed and Recompacted
Soils, Proceedings of the Third Inter-
national Conference on Soil Mechanics
and Foundations Engineering, Vol 1,
Zurich, Switzerland.
51
-------
I
/• BEFORE
SETTLEMENT
AFTER
SETTLEMENT
Figurt \. Btwn rtprtsantatlon of a cov«r systtm-
62
-------
o
d
o
9
o"
at
z
at
5
e
*>
OJ
O
O
o
«
d
CJ
d
63
-------
3.5
3.0
I I 1 I I I 11 I I I I I I I I
LEGEND
i i r
TTTl
A LEONARD ( BEAM FLEXURE TESTS }
O TSCHEBOTARLOFF ( DIRECT TENSION TESTS )
Q WES DATA ( DIRECT TENSION TESTS )
10 100
PLASTICITY INDEX, %
1000
Figure 3. Ttnsils strain virtus plasticity Indtx
-------
APPLICABILITY OF THE HELP MODEL IN
MULTILAYER DOVER DESIGN: A FIELD
VERIFICATION AND MODELING ASSESSMENT.
Nat Peters, II, Richard C. Warner,
Anna L. Coates and David S. Logs don
Dept. of Agricultural Engineering
University of EentwsJy
Lexington, ET 40546-0075
Walter E. Grube, Jr.
Hazardous Waste Eng. Research Lab
0. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Data obtained from two field-scale experimental landfill covers are used to verify
the ability of the Hydrologic Evaluation of Landfill Performance (HELP) model to predict
the performance of a multilayer soil cover system. The data used for HELP model testing
were obtained from experimental covers. The covers, which were 90 ft x 20 ft with a 3%
slope in the longitudinal direction, were large enough to require field-scale compaction
equipment. Extensive density and moisture testing was conducted daring construction of
the soil barrier layers.
A thorough description of the procedure used for estimating soil parameters is
given. These initial parameters were employed in the HELP model to predict values of
runoff, drainage, and percolation volumes for the covers, which ace compared with actual
field data. The input parameters were then optimized within each parameter's expected
range, and those calibrated values were utilized to predict subsequent flows. The HELP
Model is shown to be capable of estimating drainage from and percolation through, a
multilayer soil cover for the specific geometric conditions of this site. Optimized
parameters give an even better match between predicted and observed results.
OVERVIEW
This paper presents an assessment of
the HELP (Bydrologic Evaluation of Land-
fill Performance) model based on in-
tensely instrumented and monitored
field-scale multilayer landfill covers.
Components of this analysis include: (1)
summary results of the multilayered cover
experiment; (2) procedures and results of
the HELP model analysis, including a
thorough discussion of input parameters;
(3) assessment of the HELP aodel
performance based upon calibration and
testing; (4) advantages and limitations
of the HELP model, and (5) conclusions.
The HELP model was developed under an
Interagency Agreement between the 0.S.EPA
Municipal Environmental Research Labo-
ratory, Cincinnati, Ohio and the U.S.
Army Waterways Experiment Station at
Vicksburg, Mississippi (Ref. 7). The
program is a quasi—two-dimensional
hydrologic model of water movement into.
through and out of a landfill. The HELP
model was intended for use as a design
and review tool so that alternate land-
fill designs could be rapidly evaluated
with regard to expected amounts of
surface runoff, subsurface drainage and
leachate. The program models the hydro-
logic processes occurring in a landfill
system including surface runoff, infiltra-
tion, percolation, evapotranspiration,
soil moisture storage, lateral drainage
and leachate production (Ref. 7). The
experimental site used in this analysis
concentrates on the cover portion of a
landfill. The cover system, if properly
planned, designed and constructed, is the
primary line of defense against the pro-
duction of leachate and the associated
need for leachate containment and/or
treatment.
Three multilayered, field-scale
covers were constructed (Ref. 9), and
have been comprehensively monitored for
the past IS months. The multilayered
65
-------
cover experiment consists of three 6
meter (m) [20 foot (ft)] by 27 m (90 ft)
cells. Each cell has three layers: (1) a
lower clay barrier, (2) a middle drainage
layer of sand, and (3) an upper topsoil
layer. All layers are 0.6 m (2ft) in
depth.
A nearly automatic data acquisition
system has been installed. Data ports
are scanned every second and summed for
five minute to one hour intervals, depend-
ing upon predetermined criteria which
define significant changes. Instrumenta-
tion provides data on in—situ moisture at
48 locations and temperature at 28
locations throughout the multilayered
cover profile. Complete hydrographs of
surface runoff, subsurface lateral
drainage, and leachate are automatically
recorded. Standard climatological data
are also collected. Thus, all hydrologic
information necessary for assessing the
predictive capability of the HELP model
has boon provided by sensors within this
field-scale, multilayered cover system.
SUMMARY FIELD RESULTS
The data used herein for evaluating
the HELP model are for the twelve month
period from October 1984 through
September 1985. The HELP model compari-
sons are based on the second and third
constructed covers. The first cover was
developed as a construction demonstration
cell and has only minimal instrumenta-
tion. A synoptic summary of hydrologic
data from covers 2 and 3 is presented in
Table 1. Three time frames are consider-
ed: (1) October 1984 to June 1985 (2)
July 1985 to September 1985, and (3) the
total period of October 1984 to September
1985.
A cursory review of the data
contained in Table 1 shows that there is
a significant difference in the overall
performance between covers.
Although the construction of the clay
barriers in covers 2 and. 3 was quite
similar, the method and sequence of
topsoil placement varied between covers.
These differences resulted in signifi-
cantly different amounts of surface
runoff and evapotranspiration. See Ref.
10 for further details on construction
effects on soil moisture movement.
HELP MODEL ANALYSIS
This section provides brief back-
ground material regarding the methodology
of the HELP model. Input parameters of
the HELP model encompass the areas of
geometry, soil characteristics and
climate. Initial estimates of input
parameters are applied to the HELP model
and computer predictions are compared to
measured data for covers 2 and 3. Para-
meters are optimized for the October 1984
through June 1985 time frame and opti-
mized results are compared to observed
data. Then the performance of the HELP
model, using the optimized values i.s
TABLE 1. Hydrologic Summary of Multilayer Cover Performance
Quantity
Oct. 84 - June 85
Cover 2 Cover 3
July 85 - Sept. 85
Cover 2 Cover 3
Total
Cover 2 Cover 3
1)
2)
3)
4)
5)
6)
Preoip. (in)
Surf. Runoff (in)
Drainage (in)
Leachate (in)
(2)+(3)+(4)
(l)-(S)1
30.22
1.66
16.64
1.80
20.10
10.12
30
0
9
1
11
18
.22
.89
.54
.31
.74
.48
14
1
5
0
7
6
.15*
.16
.44
.70
.30
.85
15.16
0.09
5.33
0.54
5.96
9.20
44
2
22
2
27
16
.37*
.82
.08
.50
.40
.97
45
0
14
1
17
27
.38*
.98
.87
.85
.70.
.68
i
i
(1)— (5) = evapotranspiration
Different amounts
rainfall simulator
+ net change
of artificial
•
rainfall
in
were
soil moisture
applied
using
the modified
Kentucky
66
-------
described foe the July through September
1985 time frame.
Input Parameters
The HELP model requires three types
of input parameters: (1) geometric (2)
soil characteristics, and (3) climatic.
The geometric parameters are defined by
design and construction criteria.
Required input geometry encompasses: (1)
number of layers, (2) layer type, (3)
thickness, (4) slope at the base of all
drainage layers, and (5) coyer area. The
" layer type" is a numeric designation
that indicates whether a layer is
represented by: (1) vertical percolation,
(2) lateral drainage, (3) soil barrier,
(4) waste, or (5) soil barrier with an
impermeable flexible membrane liner (Ref.
7). The covers constructed for this
research all have three layers which are
from the surface downward, a vertical
percolation layer, a lateral drainage
layer and a barrier soil layer. All
layers are approximately 61 cm (24 in)
thick. The cover is sloped 3%, therefore
the base of the drain layer in the cover
is sloped 3%. Cover area is nominally
167.2 m2 (1800 ft2). Actual cover area
ranges from 167 m2 to 177.5 m2 (1911
ft2).
The soil parameters required by the
HELP model include porosity, field
capacity (FC), wilting point (WP),
hydraulic conductivity (K), and eva-
poration coefficient (EVAPC). Refer to
Table 2.
Field capacity and wilting point are
terms used, mostly in agriculture, to
define levels of soil water available to
plants. Field capacity is defined as the
water content remaining after free
drainage from an initially saturated soil
has practically ceased (Ref. 1) . Wilting
point is defined as the water content of
a soil at which plants wilt and fail to
recover their turgidity when placed in a
dark humid atmosphere (Ref. 1). Since
these values are not easily determined,
they have been somewhat standardized by
assuming that field capacity is the water
content corresponding to 33.5k Pa (= .33
ate) soil suction and that wilting point
corresponds to 1520k Pa (14.8 atm) soil
suction (Refs. 2,5). A pressure—plate
apparatus was used according to the pro-
cedure presented in Ref. 6, to determine
the needed field capacities and wilting
points.
Porosity values were determined as a
first step in the pressure-plate pro-
cedure. The saturated water content, on
a weight/weight basis, was determined
grav metrical ly. Knowing the saturated
water content and the specific gravity of
the soil, it was then possible to
calculate the volume/volume saturated
water content, which is approximately
equal to the porosity. The bulk density
could then be obtained, based on the
saturated sample, and used to convert
later (nnsaturated) gravimetric water
contents to volumetric values. A limita-
tion of the pressure-plate method, for
characterizing the barrier layer soil is
that the remolded sample is typically not
compacted to maximum density. Therefore
considerable consolidatiou may take place
as the pressure is increased to various
levels. Thus the estimated value of
TABLE 2. Required HELP Model Soil Parameters/Layer
Paramater
Vertical Percolation
Layer
Lateral Drainage
Layer
Barrier Soil
Layer
Porosity
Field Capacity
Wilting Point
K (in/hr)
Evap. Coeff.
Required
Required
Required
Required
Required
Required
Required
*
Required
*
Required
*
Required
* Required input but not used in modelling.
67
-------
field capacity must be somewhat in error.
Fortunately, a sensitivity analysis
showed that the HELP model is not very
sensitive to the field capacity of the
barrier layer.
The results of the presure-plate
analysis are more important for the
topsoil (vertical percolation layer)
because the HELP model uses all para-
meters for that layer. Since the entire
topsoil layer is normally not highly
compacted in the field, the pressure-
plate procedure is more representative of
field conditions. For purposes of
initial parameter estimation, the pre-
viously mentioned differences between the
topsoil layer of covers 2 and 3 were
initially disregarded. Once a soil layer
is placed in the field, samples can be
taken to detemiae in-situ porosity, field
capacity, and wilting point. But a
designer attempting to use the HELP model
to compare alternate designs prior to
construction would not be afforded this
luxury. The estimated topsoil para-
meters, from the pressure-plate analyses,
were very similar to typical values given
under the USDA Soil classification
system. See Table 3 for this comparison.
Because the three parameters'for the top-
soil are so similar to loam, the values
of K and EVAPC corresponding to loam
(Ref. 7) were used as estimated para-
meters to describe the vertical percola-
tion layer.
Parameters required in the lateral
drainage layer (sand in this case) ace
porosity, FC, and K. Again, the porosity
based upon, a saturated sample should
relate well to the field case, because
sand will not consolidate much under
pressure. In fact, the porosity was
estimated to be 0.340 while the typical
value listed for coarse sand (TJSDA
textural class) is 0.351 (lef. 7).
However, the estimated value of field
capacity using the pressure-plate
appeared very low. It was 0.041 vol/vol
as compared with a published value of
0.174 vol/vol for coarse sand. The very
low number indicates that the sand has
almost no capillary suction. Hydraulic
conductivity in sand reportedly ranges
from 6.65 in/hr for uniform fine sand to
666.5 in/hr for uniform coarse sand (Ref.
8). Ref. 7 gives rather different
values, corresponding to the USDA soil
teztnral classes, which range from 5.4
in/hr for fine sand to 11.95 in/hr for
coarse sand. Obviously, the " coarse
sand" referred to by Ref. 8 is much
coarser than that referred to in the OSDA
textural classes. As a first estimate,
6.62 in/hr (which corresponds to the USDA
teztnral class " sand'5 was selected for
K in the drainage layer.
The one parameter in Table 2 not yet
discussed is the K of the barrier soil
layer. As with many of the previous para-
meters, it is impossible to know what
that value is prior to construction of
the layer. The clay soil which was used
in the experimental covers was tested and
classified as a OR soil according to the
Unified Soil Classification system. Such
soils are expected to be " practically
impermeable" when compacted, which
generally means having a. K less than 1 x
10~7 cm/s (1.417 x 10~4 ia/hr) (Ref.
4) , Therefore, 1 x. 10"' cm/s was used
as the input parameter for the HELP
model. Table 5 gives a summary of the
design soil data initially used as input
to the HELP model.
HELP Model Performance with Initial
Parameter Selection.
The capabilities of the HELP model
were first tested using the initial
estimated and/or laboratory-estimated
TABLE 3. Comparison of Soil Parameters
Parameter
Estimate from
Pressure Plate
For USDA Texture
Class Loam (Ref. 17)
Porosity
Field Cap,
Wilting Pt.
.523
.376
,217
.521
.377
.221
68
-------
TABLE 4. Initial HELP Model Test (October 84 - June 85.)
Results
Cover 2
HELP Model
Cover 3
HELP Model
Runoff
(in)
1.66
0.58
0.89
0.58
Drainage
(in)
16.64
7.53
9.54
7.53
% Error Leachate
(in)
1.80
-55. 0.83
1.31
-21. 0.83
% Error Total
(in)
20.10
-54. 8.94
11.74
-36. 8.94
% Error
-56.
-24.
design soil parameters and on-site daily
precipitation data for October 1984 to
June 1985. Comparison was based on the
overall model performance, i.e., totals
for the entire tested time frame and not
on a daily output basis. The overall
model predictions are compared to mea-
sured cover performance, for covers 2 and
3. As can be seen in Table 4, the HELP
model, using these initial parameter
estimates, underpredicted surface runoff,
drainage and leachate production. This
underprediction of flows for both covers
indicates that the HELP model, using
estimated parameters and Lexington,
Kentucky climatic data, overpredicted
evapotranspiration. The total predicted
HELP outflow quantities were underesti-
mated by 56%, and 24%, for cover 2 and 3,
re'spectively. '
HELP calculated leachate quantities
were also underpredicted by 54% and 36%
for covers 2 and 3, respectively. It
should be noted that these HELP model
results were based on "best estimates"
of what constructed cover parameters
might be expected to be and not on
in-situ measurements. In this regard,
model predictions within 50% are
considered quite good.
Calibrated HELP Parameters
Since differences in construction of
the topsoil portions of covers 2 and 3
TABLE 5. HELP Model Inputs
Parameter
Initial
Optimized
Cover 2 Cover 3
Topsoil
Porosity
Field Capacity
Wilting Point
K (in/hr)
Evaporation Coeff.
Drain Layer
Porosity
Field Capacity
K( in/hr)
Barrier Layer
Field Capacity
K(in/hr)
(cm/s)
General
Vegetation
Evaporation Zone
Curve Number
0.523
0.376
0.217
0.21
4.5
0.34
0.174
6.62
0.352
0.000142
1 x 10~7
Good Grass
10 inches
80.95(default)
0.45
0.288
0.217
0.56
4.5
0.34
0.174
6.62
0.352
0.00024094
1.7 x 10-7
Good Grass
1 inch
85
0.45
0.324
0.217
0.40
4.5
0.34
0.174
6.62
0.352
0.0001772
1.25 x 10~7
Good Grass
7 inches
85
69
-------
caused different behavior, the covers
were treated separately in an attempt to
calibrate the HELP model using the
October 1984 to June 1985 time frame.
Emphasis in calibration was first based
on matching leachate production and
secondly drainage. Runoff is really only
a concern with respect to erosion and
storm-water management.
The calibration procedure did not
include the entire possible range of all
parameters since it could have unjustly
biased the potential users' perspective
of HELP model predictive capabilities.
The only parameters used in the
restricted optimization were porosity, FC
and K for the topsoil layer and K for the
barrier layer. Also, the evaporation
zone was adjusted for both covers.
Initial and optimized HELP model inputs
are listed in Table 5 for covers 2 and 3.
Model results of the restricted
optimization are shown in Table 6. It
was possible to calibrate the HELP model
to do an outstanding job of matching
leachate, drainage, and surface runoff
for cover 3. The calibration for cover 2
gave an excellent match for the important
parameters of leachate and drainage.
Preliminary results indicate that a great
deal can be learned about the performance
'of a cover system through a thorough
modeling appraisal, based on restricted
optimized parameters which reflect actual
construction conditions or " test fill"
conditions (Ref. 3). See Ref. 10 for
details on construction conditions.
HELP Model Performance - Test Period
From the previous section we found
that wo could optimize HELP model para-
meters based on self-imposed restric-
tions, associated with understanding both
construction and cover mechanics. How
well will the HELP model predict the
cover performance for the non-calibrated
data period of July to December 1985
based on those previously calibrated para-
meters? The results are shown in Table
7.
Table 7 shows that the calibrated
HELP model did an excellent job in pre-
dicting leachate production during the
test period for both of the multilayer
covers. Model predictions were within 12
and 23 percent for covers 2 and 3, respec-
tively. For cover 2 the HELP model was
within 25 percent for predictions of run-
off, drainage, and leachate. It should
be noted that drainage and leachate were
predicted within 12%, which is considered
excellent for any hydrologic predictions.
Although leachate prediction for cover 3
was good, drainage and especially runoff
were vastly underpredicted and overpre-
dicted, respectively, by the HELP model.
But the numbers deserve closer inspec-
tion. Notice that if the predicted run-
off of cover 3 had matched the actual
runoff, then 2.55 additional inches of
water would have been available for
drainage and leachate. Most of that
would have gone to drainage, bringing
that predition to well within 10%, and
the additional head would have caused an
increase in the predicted leachate.
Consequently, the difference in the two
sets of predictions should be attributed
more to problems with the limited cali-
bration period than with any particular
failure of the E'ELP model. Due to a rela-
tively limited monitoring period the cali-
brations were biased against making good
predictions for the test period (summer-
time) because they were based on fall,
winter, and spring data. The fact that
TABLE 6. Results of Calibrated HELP Model With Optimized Inputs
October 84 - June 85.
Source
Runoff
(in)
Drainage
(in)
Error
Leachate
(in)
% Error
Cover 2
HELP 2
Cover 3
HELP 3
1.66
0.333
0.89
0.745
16.64
15.132
9.54
9.673
-9.1
1.4
1.80
1.820
1.31
1.30
1.1
-0.8
70
-------
TABLE 7. Results of Calibrated HELP Model Tests
July 1985 - September 1985
Source
Cover 2
HELP 2
Cover 3
HELP 3
Runoff
(in)
1.16
1.437
0.09
2.636
Drainage
(in)
5.44
5.874
5.33
2.449
% Error
_
8.0
-
-54.1
Leachate
(in)
0.70
0.616
0.54
0.415
% Error
_
-12.0
-
-23.1
cover 3 was the "better calibrated", in
terms of matching data for the calibra-
tion period, really means that it was
most biased against summertime predic-
tions. Summertime conditions are
significantly different, because the soil
moisture is more rapidly depleted and
often there is a longer time between
rainfall events. Data are currently
being compiled to extend the test period
through December 1985 in order to lessen
the bias against the test period.
The exercise of calibrating the HELP
model demonstrates that very realistic
results are possible through reasonable
adjustments to the input parameters.
Testing indicates that, if the calibra-
tion is of longer duration, the HELP
model may be very capable of making
excellent predictions.
ADVANTAGES AND LIMITATIONS OF THE HELP
MODEL
Advantages
The HELP model has several advantages
which make it a potentially valuable
design tool. The model uses published
methods to model the effect of all the
major hydrologic processes of moisture
movement and balance in a landfill. The
model ties all of the processes together
into a computer program which makes it
feasible to model long time periods.
Especially important is the inclusion of
a daily evapotranspiration methodology in
facilitating long term simulations.
The HELP model is designed to allow
.for various layer configurations and
drainage lengths that are typical in a
landfill design. This flexibility allows
for very quick comparisons of drainage
and leachate predictions for alternative
landfill designs. The computer program
includes default climatologic data for
102 cities, and it has very detailed
default data for soil characteristics.
The default data also facilitate relative
comparisons based upon different climates
or different soil types.
Limitations
Some of the algorithms used by the
HELP computer program to model the
various components of water movement have
specific limitations. One potential
limitation of the HELP model is that
water is not allowed to move laterally in
a designated vertical percolation layer.
That is not a serious problem for this
particular study. However, in many
typical landfills a drainage layer is not
used between the vegetated layer and the
barrier layer. If there is a significant
difference in hydraulic conductivity
between the two layers, the upper layer
would certainly act as a lateral drainage
layer (assuming the barrier layer is
sloped). This particular limitation can
be partially overcome by separating the
vegetated layer into two segments for use
in the model. The upper segment would be
designated a vertical percolation layer
and the lower segment would be a lateral
drainage layer with the same hydraulic
conductivity.
Another limitation indicated by
results of this experiment (Ref. 10) is
that flow in the topsoil layer may not be
strictly idealized porous media flow.
However, that is a limitation of any
model for saturated/ unsaturated soil
water movement which is based upon the
mathematical equations for porous media
flow. Potential users of the HELP model
71
-------
should be aware that the predictions have
much greater chance of validity if all
soil materials (including the topsoil
layer) are placed as uniformly as
possible.
Finally, the inclusion of default
soil data is a very attractive feature of
the HELP model. However, the parameters
are presented in the HELP model manual
(Ref. 7), for the various USDA textural
classes, as if they are established
values. An inexperienced model user may
incorrectly assume that, if the soil can
bo classified according to teztural
class, its characteristics are " known".
Yet practically any natural soil type can
be worked or compacted such that the soil
characteristics will not be consistent
with those listed in the manual.
CONCLUSIONS
The HELP model proved to be capable
of estimating drainage and leachate
production based on initial parameter
estimates of tested cover materials.
Optimized parameters, based on the actual
constructed cover, could better match
HELP model prediction with observed
results. The HELP model performs well,
for this application, and could be used
to document the expected performance of
alternative landfill cover configura-
tions. It is recommended that a range of
feasible input parameters be tested to
produce estimates of leachate production
in which future recycling, treatment, or
and/or hauling cost can be projected.
ACKNOWLEDGEMENT
The authors recognize the invaluable
assistance of Jarnos Wilson in construct-
ing monitoring, and maintaining the
experimental covers.
The investigation reported in this paper
(86-2-86) is in connection with a project
of the Kentucky Agric. Erp. Station and
is published with the approval of the
Director of the Experiment Station.
REFERENCES
1. Brady, N. C., 1974. The Nature and
Properties of Soils. Eighth Edition.
MacMillan. Publishing Co., Inc. New
York.
2. England, C. B. 1970. Land Capa-
bility: A Hydrologic Response Unit
in Agricultural Watersheds. ARS
41-172, Agricultural Research
Service, USDA.
3. EPA. 1985. Construction Quality
Assurance For Hazardous Waste Land
Disposal Facilities - Public Comment
Draft. J. G. Herrmann, Proj ect
Officer. EPA/530-SW-85-021. U.S.
EPA.
4. Lambe, W. T. and R. V. Whitman.
1979. Soil Mechanics. SI Version.
John Wiley and Sons, New York.
5. Lutton, R. J. , G. L. Regan, and L. W.
Jones. 1979. Design and Construction
of Covers for Solid Waste Landfills.
EPA-600/2-79-165. U.S. EPA.
6. Richards, L. A. 1965. Physical
Condition of Water in Soil. In:
Methods of Soil Analysis - Part I.
C. A. Black, editor. American
Society of Agronomy, Inc.
7. Schroeder, P. R. , J. M. Morgan, T. M.
Wai ski, and A. C. Gibson. 1984. The
Hydrologic Evaluation of Landfill
Performance (HELP) Model. Vol.1.
EPA/530-SW-84-009. U.S. EPA.
8. Tchobanoglous, G. , H. Theisen, R.
Eliassen. 1977. Solid Wastes:
Engineering Principles and Management
Issues. McGraw-Hill Book Company.
New York.
9. Warner, R. C. , J. E. Wilson, N.
Peters, H. J. Sterling, W. E. Grube,
1984. Multiple Soil Layer Hazardous
Waste Landfill Cover: Design,
Construction, Instrumentation and
Monitoring. In: Land Disposal of
Hazardous Waste-Proceedings of the
Tenth Annual Research Symposium.
EPA-600/9-84-007. U.S. EPA.
10. Warner, R. C., et al. (in progress).
"Demonstration and Evaluation of the
Hydrologic Effectiveness of a Three
Layer Landfill Surface Cover Under
Stable and Subsidence Conditions -
Phase I, Final Project Report." U.S.
EPA.
72
-------
REMEDIATION OF AN INDUSTRIAL DUMP SITE - A CASE HISTORY
Robert C. Ahlert and David S. Kosson
Department of Chemical and Biochemical Engineering
Rutgers, The State University of New Jersey
College of Engineering
P.O. Box 909
Piscataway, New Jersey 08854
ABSTRACT
The case history of design and implementation of a remediation strategy for a
hazardous waste disposal site is described. A number of industrial wastewater sludges had
been accumulated in an untined impoundment over several decades. Infiltrating rainfall
produced leachate with elevated TOC and created a source of slowly diffusing groundwater
contamination. Containment and treatment of leachate were straightforward means of
groundwater protection. The mass of organic matter present in the lagoon translated into
a prolonged period of leachate production and management. Long-term treatment and
monitoring costs were considered excessive, while changes in local geohydrology might
necessitate additional capital investment.
Leachate and sludges were sampled. A laboratory program was carried out, with two
goals. It was desired to evaluate increased rates of TOC removal by optimal, forced
extraction. Also, it was necessary to evaluate destruction of organic carbon species in
leachate produced by normal infiltration and by artificial extraction. An aqueous
extraction scheme was developed, an optimum alkaline extraction solution was found to
accelerate TOC removal by ten to fifty-fold. A parallel investigation of microbial
acclimation and degradation, demonstrated that traditional aerobic treatment of low-
strength and high-strength leachates was about 50% effective. A bench-scale study with
packed-bed bioreactors, utilizing aerobic and anaerobic microbial communities in series,
exhibited greater than 95% TOC removal.
A pilot plant was designed, built and operated on-site. The pilot plant includes
caustic solution preparation and application, extraction of sludge deposits in place,
forced extract neutralization and flocculation, nutrient addition, sequential microbial
treatment and effluent recycle or discharge. The pilot plant Is heavily automated and
semi-continuous in operation. Performance has matched design goals. Typical data are
presented. Pilot-scale operations have demonstrated forced leaching and complete sludge
carbon destruction at high rates. It 1s estimated that site renovation can be
accomplished in a period of 2-4 years, with an effluent suitable for discharge.
Process flow diagrams, mass and energy balances and control requirements for a
full-scale plant have been prepared. Plant design is largely dictated by scale-up of the
existing pilot plant. Operation and data acquisition for the new technology, at the pilot
level, were the critical steps in the remediation stratgey.
73
-------
1. SLUDGE LAGOON DESCRIPTION
1.1 Background
Over a period of several decades,
Industrial waste sludges were landfilled in
an unllned surface Impoundment. During the
period of operation, the compositions and
rates of deposition of sludges varied
greatly. Primary and secondary sludges
were deposited. The primary sludges were
lime neutralized inorganic matter, includ-^
ing neutralizatiion wastes, spent catalysts
and solid residues from diverse chemical
manufacturing operations. The secondary
slduges were blomass from aerobic treatment
of aqueous effluent from the same manufactur-
ing activities.
The resulting 4.1 acre site contains
approximately 30,000 yard^ (yd^) of sludge.
A representative lagoon cross-section is
presented in Figure 1. Two principal
layers are found within the fill material.
The first layer cansits of approximately
15,000 yd3 of secondary sludges deposited
over a ten year period, prior to 1967. In
response to complaints 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 yd^ 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 yd^
of contaminated fill material. At present,
the physical state of the sludges range
from solid to gelatinous. Leachate from
the sludges can impact local groundwater
resources.
.
16 OH
150--80'
140'
120
HO'-
100'
90'
90
EL. 152,5
EL 145.7
RGANIC SILT
AND SAND
FILL
Hi
CLAY AND
•30'
-20'
LEGEND
/ EH SILT AND BLACK SLUDGE
BLACK MOIST SLUDGE
3 WHITE/GREY SLUDGE
375' 300' 225' 150' 75' 0 75' 150' 225' 300' 375'
Figure 1. Lagoon Cross-Section
74
-------
1.2 leachate Characteristics
TABLE 1. LEACHATE CHARACTERISTICS
Several leachate samples were removed
from the lagoon for characterization. Two
samples were taken on June 26, 1984. They
were pumped in sequence from a monitoring
well after one well volume had been
rejected. Two samples were pumped on June
26, 1984, sequentially, from a 6-inch test
well in the lagoon. These samples were
representative of the water in the well
before existing pumping occurred. Two
additional samples were pumped from the
same 6-inch test well after one well volume
had been rejected. All four samples were
approximately 2.5 liters in volume.
A sample of 30 gallons was pumped from
the 6 inch test well on July 12, 1984 after
approximately one well volume had been
removed. A subsample was withdrawn from '
the sample drum for analysis. A second
subsample was withdrawn from the drum for
analysis on July 24, 1984. A 30-gallon
sample was pumped from the 6-inch test well
on August 28, 1984 after approximately one
well volume had been rejected. A subsample
was withdrawn from the drum for analysis.
All samples were assayed for total
organic carbon (TOC), total Kjeldahl
nitrogen {TKN}, ammonia and pH. Representa-
tive samples were assayed for nitrate,
electrical conductivity (EC), chloride
total phosphorus and volatile fatty acids
(VFA). Representative results of these
assays are presented in Table 1. Semi-
quantitative spectrographic analyses were
performed on several samples; typical
results are presented in decade format in
the table also. Acid and base titration
curves were developed for representative
samples.
1.3 Sludge Characteristics
Sludge samples were excavated from the
lagoon on July 3, 1984. A slit trench was
dug with a backhoe. Black sludge, presumed
to be primary sludge, was encountered at a
depth from just below the surface to
approximately 8 feet. Samples 1 and 2,
approximately 12 gallons each, were taken
from black sludge located at depths, between
6 and 8 feet. Water was encountered at
approximately 2.5 feet. The trench filled
rapidly with water. Plastic sheet was
encountered at a depth of approximately 4
feet.
pH
TOC (mg/i)
NH3 (mg N/Jl)
TKN {mg N/A)
N0§ .(mg/A)
Total P (mg/l)
a- .(rng/t)
6.5 - 7.7
170 - 5000
11 - 620
25 - 820
< 0.01
< 1.5
< 20
Residue (24 hr @ 103C) (mg/Jt) 2750 - 4280
TOS (mg/*) 2610 - 4080
Metallic Species (mg/A)
Ag
AA
B
Ba
Ca
Cr
Cu
Fe
Mg
Mn
Na
Ni
Pb
Si
Sn
Ti
Zn
.11 -
1.1 -
.011 -
.11 -
1140 -
.011 -
.11 -
11.4 -
114 -
11.4 -
114 -
.11 -
.11 -
11.4 -
.011 -
.11 -
1.1 -
.011
.11
.0011
.011
114
.0011
.011
1.1
11.4
1.1
11.4
.011
.011
1.1
.0011
.011
.11
Brown sludge, presumed to be secondary
sludge, was encountered at depths between 9
and 13 feet. This sludge had grey streaks,
apparently the result of the prior lime
addition. Samples 3 and 4, approximately
12 gallons each, were taken from sludge
located at depths between 10 and 13 feet.
Additional plastic sheet was encountered at
a depth of approximately 12 feet. The
maximum depth of the trench was less than
14 feet. The trench-was backfilled after
samples, were taken.
Samples 1 and 3 were assayed for
solids content and apparent density, see
Table 2. Semi-quantitative spectrographic
analyses and total carbon assays were
performed on samples 1 (1-8LK) and 3
(3-BRN). Typical results are presented in
75
-------
the table also, metals are in decade
format.
TABLE 2. PRIMARY AND SECONDARY
SLUDGE CHARACTERISTICS
Characteristic
Primary
Sludge
Apparent density
% Solids ' '
{24 hour (3 103 C)
% Solids
(12 hour 0 550 C)
Total C
(% on a dry
weight basts)
Inorganic SpTecies
(% on a dry
weight'basis)
1.16
,
28.0
22.4
\
7.2
• Secondary
Sludqe
1.12
23,0
' 15.6
16. 3^.
At
8
Ba
, Ca .
Cr,_
*Cu '*"*
Fe
. -Mg
Mn
Na
Ni
P
Pb
Si
Sn
Ti
V
10.
' .001'
.1
„ _ 10
1-
~ ' " M
. -I
->• * •• -.I'
.01
.1
.001
.1
.01
10
.1
.001
- 100
-'. .01
- 1
.-. 100.
- 1
"- .1 '
- 10
. — °> 1
-•...1
- 1
- .01
— 1
- .1
- 100
-
- 1
- .01
- • • I -
-
.001 -
. , 10 -
.01 -
""'voi -
•• '• 1 -
.- ...... .1 _
.01 -
„ .1 -
.001 -
.1 -
.01 -
1 -
•.'001""-"
' ' '.01 '-
' ;oor-'
10
.01
100
.1
.1
10
1
..1
1.
.01
1
.1
10 "
.01
.1
"01
1.4 Statement of Problem and Approach
Lagoon clean-up Is viewed as two
Interrelated problems. The first problem,
1s the removal of contaminants from the
lagoon, without major excavation. The
second problem is treatment of the stream
containing the stripped contaminants,
Including organic and inorganic species.
A research program was designed to
evaluate several treatment options for the
lagoon, refer to Figure 2. Elements of
this program included forced extraction of
representative sludge samples and laboratory
evaluation of aerobic (secondary) mixed
mlcrobial treatment and sequential aerobic
/anaerobic soil-based microbial treatment
of naturally occurring leachate and forced
extracts. Acidic, neutral and alkaline
aqueous extractants were considered for
removal of organic contaminants.
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. Recovered extract was
to be treated on-site using a soil-based,
mixed microbial treatment system. A
pilot-plant was constructed early in the
summer of 1985, it was operated success-
fully for a, period of 140 days, from July
through November 1985.
2. LABORATORY SLUDGE' EXTRACTION STUDIES
2.1 Approach -'Fea'sbi 11 ty
Initial extraction experiments were
performed to evaluate forced extraction for
controlled removal of organic species from
the primary and secondary sludges. An
extractant capable,of "forced leaching" the
sludges at an accelerated rate would allow
rapid renovation of the sludges.; without
excavation. The resultant leachate would
be contained, and treated on-site.
Extractants considered for in-situ
treatment, of- the sludges had to be inexpensive
and have minimum environmental impact.
"Extractants employed were acidic, neutral
and alkaline aqueous solutions. Parameters
evaluated included time-to-equilibrium,
extractant pH, and extractant-to-sludge
mass ratio.
2.2 Materials and Methods
2.2.1 Sludges. Primary,and secondary
sludge samples used in these experiments
were excavated on July 3, 1984. Primary
•sludge and secondary sludge samples were
identified as 1 and 3; refer to Section 1.3
for sludge characteristics.
2.2.2 Extractants. Alkaline extractants
consisted of sodium hydroxide in distilled
water at the specified pH. Neutral
extractants were distilled water only.
76
-------
, INDUSTRIAL SLUDGE LAGOON;
PROBLEM DEFINITION
CONTAMINANT REMOVAL
IN-SITU EXTRACTION
CONTAMINANT DESTRUCTION
SOIL-BASED MICROBIAL TREATMENT
LABORATORY EXTRACTION
STUDIES (FEASIBILITY)
LABORATORY SOIL COLUMN
STUDIES (FEASIBILITY.)
PILOT-PLANT DESIGN AND FABRICATION
.-,»- •, '
r i
t
MODEL DEVELOPMENT
t
PILOT- PLANT OPERATION
MODEL DEVE
LABORATORY EXTRACTION
STUDIES (PILOT-PLANT
EVALUATION)
LABORATORY SOIL COLUMN
STUDIES (PROCESS CONTROL AND
OPTIMIZATION)
MODEL CALIBRATION
AND'VERIFICATION
MODEL CALIBRATION
AND VERIFICATION
PROCESS MODEL
FULL-SCALE DESIGN
COST ESTIMATE +/- 25%
APPROVALS (SPONSOR AND NJDEP)
FINAL DESIGN
PERMITS
IMPLEMENTATION
Figure 2. Project Development Schematic
77
-------
Acidic extractants consisted of sulfuric
add in distilled water at the specified
pH. All chemicals were reagent grade.
2.2.3 Procedures. Studies were conducted
by mixing a measured mass of sludge with a
quantity of extractant. The sludge-extrac-
tant mixture was shaken for the period
specified for each case. The mixture was
centrifuged and supernatant solution
decanted. Supernatant was assayed for pH,
EC, TOG and TDS.
2.3 Results and Pi scussion
In preliminary experiments, 0.1N 1^804,
distilled water and 0.1N NaOH were considered
as extractants for both primary and
secondary sludges. The sludge-to-extractant
ratio was 10 mi extractant per gram of
moist sludge. One experiment employed
sequential extractions. Each extraction was
shaken for 24 hours. Three replicates were
performed for each data point. At the
conclusion of each extraction step, the
extraction mixture was centrifuged and the
supernatant decanted. Fresh extractant was
added and the mixture shaken again. This
cycle was repeated until four sequential
extracts were obtained for each sludge/
extractant combination.
Experiments were performed to establish
the time required for each sludge/extractant
combination to reach equilibrium. Single
extractions were allowed to shake for 6 to
48 hours. Two replicates were performed
for each data point. Based on the results
obtained from preliminary experiments, the
effect of alkaline extraction pH on
extraction effectiveness was studied
further. Aqueous sodium hydroxide solutions
at pH 9, 10, 12 and 13 was used. Sequential
extractions were performed; however, the
shaking period was extended to 48 hours.
Three replicates were performed for each
data point.
The effect of the extraetarit-to-sludge
ratio was examined also. Primary and
secondary sludges were extracted. Aqueous
sodium hydroxide, at pH 11 and 12-, was used
as the extractant. Sequential extractions
were performed; the ratio of extractant per
gram of moist sludge was varied. Ratios of
10 ml, 5 ml and 2 mjt per gram of moist
sludge were employed. Three replicates
were performed for each data point.
Extraction of secondary sludge proved
to be Independent of extractant pH;
however, extraction of the primary sludge
was dependent on pH. Alkaline extractant
proved to be much more effective than
neutral or acidic extractant. A 0.1N
solution of sodium hydroxide, at pH 13,
achieved the best removal of organic
matter. Satisfactory results were achieved
with NaOH solutions at pH between 11 and
12, Order-of-magnitude reductions in
caustic requirement and extract TOS made
extraction at pH at 11 and 12 desirable.
The ratio of extractant volume to
sludge mass was varied. Ratios as small as
2 ml extractant per gram of moist sludge
were found to achieve high organic species
removal. Extract TOCs as high as 3200 mg/jl
were observed.
Sequential extractions of fresh
quantities of sludge with the same extractant
were performed also. At the conclusion of
each extraction step, the mixture was
centrifuged and decanted. The supernatant
was added to fresh sludge and shaken. The
cycle was repeated to determine the maximum
extract TOC attainable. Equilibrium was
reached within five extraction steps. TOCs
as high as 8900 mg/£ were observed.
3. LABORATORY COLUMN STUDIES
3.1 Motivation
A biological method for the treatment
of natural leachate and forced extracts was
considered for implementation. This method
is a soil-based microbial treatment system.
In this process, a mixed microblal population
is developed in a soil structure. The
leachate or extract to be treated is
applied to the surface in a packed bed
configuration and allowed to percolate
through the soil column. During this
process aerobic degradation occurs near the
surface, where oxygen is available due to
atmospheric diffusion. Anaerobic degradation
follows and dominates at greater depths.
Process control is maintained through
management of influent loading rates and
concentrations. This section examines
development and application of the process
to the sludge lagoon case (see Kosson and
Ahlert [1] for additional details).
3.2 Materials and Methods
Samples of natural leachate (see Table
1.1) were obtained from the lagoon at the
study site. Samples of both primary and
78
-------
secondary sludges were excavated from the
lagoon; see Table 1.2. The extract for use
In the experiments were obtained by
contacting the appropriate sludge with
0.005N aqueous sodium hydroxide (pH 11.5)
for 48 hours. Two liters of extractant
were used per kilogram of sludge. The
resulting mixture was allowed to settle;
supernatant was centrifuged and recarbonated
wih pure €63, to a pH of 7. Recarbonated
liquid was filtered to remove carbonate
floe. Extracts from the primary and
secondary sludges were mixed in proportions
similar to those expected from the actual
sludge disposal lagoon. The characteristics
of the resultinq extract are presented in
Table 3.1,
Laboratory soil column experiments
were designed to investigate the ability of
mixed nricrobial populations in a soil
structure to biodegrade natural leachate
and forced extracts. Previous experiments
have shown that data obtained from laboratory
scale investigations are readily extrapolated
to field situations [1,2]. Hydraulic
conductivity (permeability), hydraulic
loading, organic loading, buffering and pH,
and soil adsorption, all important inter-
related variables, were considered during
the study. To obtain as much data as
possible from a. limited number of columns,
fractional factorial designs were employed.
The following three parameters were the
primary factors considered: column packing,
organic loading and retardation of bioslime
development.
Column packing was considered because
of an influence, via adsorption, on solute
retention, reduction and removal; packing
also controls hydraulic conductivity and
porosity. Two column packings were chosen.
One was a sandy loam, selected based on
previous experience with a similar soil and
availability on site. The second column
packing was sandy loam mixed with activated
carbon (Calgon PCB 30 x 140). The particle
size of this activated carbon is in the
sand range. Packing permeability and
adsorption capacity were increased.
The second variable was organic carbon
loading, an influential factor in the rate
of bioslime growth. The rate of bioslime
growth may be directly proportional to
influent TOC, depending on the composition
of the influent [3]. Two influent concentra-
tions, nominally full-strength or half-
strength leachate or extract, were considered
in each experiment. The third factor
79
examined was the effect of high concentra-
tions of chloride ion and/or sodium azide
on the retardation of'bioslime development.
Each laboratory column (1C) consisted
of glass process pipe,_ 3 inch in diameter
with an effective packing depth of 18
inches. The foundation of the packing
consisted of 1 inch of sand, a thin (1/8
inch) layer of glass wool and 2 inches of
glass beads. The columns were operated
with 150 mm Hg of vacuum applied continuously
at the base, through collection flasks.
The vacuum balanced capillary forces in the
soil structure, in an effort to replicate
field conditions. Each. LC received 30 mil
inoculum of'an acclimated mixed microbial
population from the secondary sludge of a
municipal sewage treatment plant. Acclima-
tion procedures are described by Bpyer et
al. [4].
3.3 Treatment of Natural Leachate
Sixteen laboratory soil columns were
operated to examine the treatment of
naturally occurring leachate. Two replica-
tions were performed for each trial case;
thus, seven trial cases were investigated.
Full-strength and half-strength leachates
were employed as influent for both packing
types. In addition, full-strength leachate
was supplemented with either chloride or
sodium azide and chloride to study inhibition
of microbial growth. LCs receiving
influent without inhibitors were operated
continuously for 100 days; LCs with
influent containing inhibitors were
operated continuously for 50 days.
Typical hydraulic flux rates for the
sandy loam and the sandy loam mixed with
activated carbon were 8 and 25 liters per
square meter (m2)' per day, respectively.
Columns fed full-strength leachate generally
established steady-state effluent TOC'
concentrations twice those of columns fed
half-strength leachate. No clear relation-
ship between effluent TOC concentration and
packing type was observed. On an integrated
mass basis, LCs fed full-strength leachate
removed approximately twice as much TOC as
those fed half-strength leachate, with the
same packing. In addition, LCs packed with
soil supplemented with activated carbon
removed approximately three times as much
TOC as those packed with soil only. This
effect was primarily a reflection of the
differences in hydraulic flux. TOC
reductions varied from 92 to
-------
TABLE 3.1 EXTRACT CHARACTERISTICS
PH
TOC (mg/X)
NH3 (mg/H)
TKN (rag/*)
IDS (g/JL)
VFA as TOC (mg/jl)
Acetic
Propionic
Isofautyric
Butyric
Total
Na+ (mg/t)
Ca++ (mg/Jt.)
Primary Extract
7,0
400
91
235
1.3
< 25
< 25
< 25
< 25
< 25
241
89
Secondary Extract
7.0
2500
400
640
6.8
772
283
158
267
1480
261
906
Combined Extract
7.0
2080
338
559
5.7
618
226
126
214
118
157
743
Effluent pH was a general indicator of
the status of the column. Initially, the
effluent pH was near neutral, but the
development of a balanced microbial
population led to a gradual rise in the pH
that stabilized at a slightly alkaline
level. The addition of CJ,~, at concentra-
tions of 1000 mg/£, as either NaCjl and
CaCJtg, to LC influent had limited inhibitory
effects. The concentrations of TOC in the
effluent were twice that of uninhibited
columns for sandy loam packings and were
SOX greater for sandy loam with activated
carbon packings. For packings with sandy
loam and activated carbon, the hydraulic
flux responses were similar regardless of
the CjT content.
3.4 Treatment of Sludge Extract
To examine the treatment of sludge
extract, four trial cases, each replicated,
were investigated. Full-strength and
half-strength extract was employed as
influent for the LCs, all packed with sandy
loam plus 5% activated carbon. Full-strength
extract was supplemented with sodium azide
80
to study the inhibition of microbial
growth. In addition, to estimate the
hydraulic characteristics of the packing,
two LCs were operated with only a chloride
solution as influent.
The columns fed full-strength extract
displayed TOC reductions in excess of 97%.
The columns fed half-strength extract
showed TOC reductions In excess of 99%,
throughout the experiment. On a mass
basis, the columns fed full-strength
extract removed approximately twice as much
TOC as columns fed half-strength extract.
In the absence of inhibition, TOC reduction.
was 99%.
The effluent pH, for LCs that were fed
half-strength extract, decreased rapidly
from an initial neutral level to an acidic
pH of approximately 4. The pH levels
fluctuated between 4 and 5 and paralleled
variations in hydraulic flux. LCs fed
full-strength extract followed the same
pattern.
-------
Two LCs were operated to estimate the
hydraulic characteristics of the packing,
using a chloride tracer at an influent
concentration of 100 mg/l Cl~. Estimates
of the packed bed porosity and dispersion
coefficient were 0.30 and 6.0, respectively.
A soil-based microbial treatment
process was found suitable for treatment of
natural leachate and forced extracts from
an industrial sludge lagoon. Reductions in
dissolved organic carbon ranged from 90 to
99£, depending on packing type, additives
in the leachate feed, and the rate of
influent. These results were extrapolated
to field conditions and formed the basis
for a pilot plant design and operation.
4. ON-SITE PILOT-PLANT
4.1 Pilot-Plant Design
4.1.1 Process description. The pilot
system was designed to consist of several
sequential process steps. The complete
process flow diagram is presented in Figure
3. The first process step is extraction of
sludges present 1n representative sections
of the lagoon. Sodium hydroxide solution
is mixed batch-wise in a 200-gallon process
tank (tank 1) and is applied to sludges
present in an extraction bed. The solution
can be applied to the surface of the
extraction bed, through perforated pipe, or
by injection into the sludges through six
well points. Extract is recovered from the
extraction bed through two wells. Recovered
extract is pumped through a basket strainer
and cartridge filter into a 1000-gallon
storage tank (tank 2),
The second process step is adjustment
of pH, dilution as necessary, and addition
of nutrients to extract stored in tank 2.
This occurs continuously in tank 3. Tank 3
is a 200 gallon process tank, baffled into
four sections by two overflow baffles and
one underflow baffle. Extract is pumped
from tank 2 to the first chamber of tank 3.
Carbon dioxide is bubbled through the
extract to adjust pH between 7.0 and 7.5,
Extract passes an underflow baffle to the
second chamber of tank 3, where floe
created by recarbonatlon is settled.
Clarified extract passes an overflow
baffled into the third chamber of tank 3;
dilution with recycle or potable water and
addition of nutrients occurs here. Extract
passes to the fourth chamber from which it
overflows and is applied to the surface of
the treatment, bed.
The third process step is treatment of
the modified extract (effluent from tank 3}
in an aerobic/anaerobic soil-based bioreaction
bed. Treatment is taking place in a soil
bed in which an aerobic microbial population
is maintained in the upper region and an
anaerobic microbial population is maintained
in the lower region. Extract applied to
the surface of the treatment bed percolates
through the soil column and is biodegraded.
Effluent from the treatment bed is recovered
through a screened well (well 3). Recovered
effluent is pumped to a 500-gallon storage
tank (tank 4} from which it can be recycled
onto the teatment bed, recycled to tank 1,
or discharged,
4.1.2Extraction bed design. The extraction
bed was designed to examine in-situ
extraction of the sludges in a representative
section of the lagoon. A 10 x 10 foot bed
was isolated from the remainder of the
lagoon by steel sheet piling. The sheet
piling extends from approximately two feet
above the lagoon surface to approximately
one foot into the lagoon bottom. The sheet
piling is interlocking and nominally
watertight.
Initially, extractant was applied to
the surface of the extraction bed through
parallel 1-inch perforated PVC pipes.
Subsequently, extractant has been injected
into the extraction bed through 2-inch
diameter, 24-inch long screened well
points. The well points are screened at a
depth of approximately 13 feet from the
lagoon surface. Extract is recovered from
the extraction bed through two wells. The
wells are constructed of a 6-inch diameter
PC casing with 2-foot long, 100 slot
stainless steel screens. Well 1 was
screened approximately 1 foot Into the
bottom of the lagoon. Well 2 is screened
near the bottom of the sludge within a
deposit of secondary sludge. During the
drilling of well 1 continuous split spoon
samples were taken. There is a stainless
steel deep well pump, a low water level
sensor and a high water level sensor in
each we!1.
4.1.3 Treatment bed design. A treatment
bed was constructed immediately adjacent to
the extraction bed. The perimeter of the
treatment bed consists of steel sheet
piling interlocked with the sheet piling of
the extraction bed. Sludge within the
lerimeter of the sheet piling was excavated
81
-------
oc
IN3
US
c
-?
n>
CO
i
•o
o
rt-
O
o
(B
O
s:
C3
DJ
PUMP
PROCESS FLOW DIAbKAM
TITLE: FLOW DIAGRAM
PROJECT : LAGOON 9*
OATt 227P5
RUTGERS UNIVERSITY
DEPT, OF CHEMCAL AMD BOCHaBAL INC.
PISCATAWAX NJ.
DRAWN BY
APOROVEB BY n^ Kt-'SON
DRAWING NO.
-------
to a depth of 8 feet and removed to an
alternate location within the lagoon. The
inside of the sheet piling and bottom of
the resulting pit were lined with 1/2-inch
plywood. Six inches of sand was placed on
the bottom and the entire pit was lined
with a continuous PVC liner.
Six inches of sand was placed inside
the liner and a 6-inch diameter well (well
3) was placed in the center. An additional
2 feet of sand was backfilled around the.
well. Two 1-inch methane vents were
installed on top of the sand and the
remainder of the bed was backfilled with 5
feet of soil mixture. The soil mixture
used to backfill the treatment bed consisted
of local top soil, sand and granular
activated carbon in 15:1:1 volume ratio.
The soil mixture was mixed on-site, with a
bucket loader, prior to backfilling.
Modified extract is applied to the
surface of the treatment bed through 1/2
inch perforated PVC pipes. Water level
sensors are located on the surface of the
bed to prevent flooding. Effluent from the
process is collected in well 3, Well 3"
contains a stainless steel deep well pump,
a low water level sensor and a high water
level sensor. Water collected from well 3
is pumped directly to tank 4.
4.1.4Process area. The process area
consists of a 30 x 35 foot compound located
on top of shale fill approximately 100 feet
east of the extraction and treatment, beds
at an elevation of approximately 10 feet
above the surface of the lagoon. The
compound is surrounded by an 8-foot
industrial chain link fence for security.
Tanks 1, 2, 3 and 4, a storage shed for C0£
cylinders and nutrient solution, electrical
power supply and controls, a wash basin and
safety shower are located within the
compound. All tanks are modified steel
fuel oil tanks. Potable water is provided
by pipeline from approximately 1/4 mile
from the compound.
4.1.5 Electrical power and controls.
Electrical power is provided from a 1 KVA
telephone pole line approximately 1/4 mile
from the process area.. .A pole transformer
provides 440 volts AC power to the process
area through two direct burial cables
buried in a shallow trench alongside the
pipe providing potable water. Within the
process area, a wood structure houses a
transformer, circuit breakers and process'
controls. All pumps, valves and controls
operate on 120 volts. All connections
between the power panel, sensors, valves
and pumps are within solvent cemented PVC
conduit.
The process controls consists of four
operationally independent controllers. All
controls are housed in a water tight
enclosure on the power panel'. Three of
the controls are identical units, i.e., one
pump controller each for wells 1, 2 and 3.
The basis of the well pump controls are a
high water level sensor and a low- water
level sensor (float switches) in each well.
The fourth controller controls operation of
pumps, valves and sensors associated with
tanks 1, 2 and 4 and sensors on the surface
of the treatment bed. Tank 1 is only
operatead manually. The primary function
of the fourth controller is to regulate
application of influent to the treatment
bed.
4.2 Results
Initially effluent volumes for the
extraction bed were low, and influent
volumes greatly exceeded influent volumes.
This was attributed to the initially
unsaturated state and low permeability of
the sludge.deposits. Well points were
installed to inject caustic extractant into
the sludges. Subsequently, extract
recovery increased to approximately 60%
with influent application rates of approxi-
mately .25 inches/day.
Effluent pH and TOC for the extraction
bed are presented in Figures 4 and 5.
Initially, both parameters were high. pH,
TOC and TOS of extract recovered from well
2 were approximately 10, 11,000 mg/i, and
28,000 mg/A,.respectively. However, all
three parameters decreased between days 40
and 60. On day 60,"pH TOC and TDS of
extract recovered from well 2 were 9.2,
5,200 and 11,000, respectively. Initial
high pH and TDS values are attributed to
lime previously applied to the sludges and
other soluble solids originally present in
the extraction bed. Increasing effluent pH
and TOC was observed on day 70. These
increases are considered to be in response
to .increased influent pH.. Effluent TOS
continued to decline during this period,
approaching influent TOS. •
Throughout the period of operation,
effluent TOC from the extraction bed
correlated closely with effluent pH. This
83
-------
I
EXTRACTION BED
WKLL 1 it
10.6-
1O-
9.5-
•-
8.5-
a -
7.6-
7 -
t
*™L. i_ -fr "^ , »!•
A . + . "fr *fr
Q *° B|?+ ***+ ° °
" D D _rftf?J»Ln
D ° ^a^, "** rf1 m
— r" r i i i i -i i i i i i i
? SO 40 80 80 100 1X0 1-
to
TlUt (DAYS)
• O WXLL t + I«X£ a
Figure 4. Extraction Bed Effluent pH
is
14-
13-
1S -
11 -
10-
8-
8 -
7~
0-
8 -
4-
a -
*
EXTRACTION BED
WKLL 1 * M
a -HI
a +
+ -M.
—^I I I | | i
40 00 so roo
TttfJf (DAYS)
Q WKLL 1 * WELL 2
Figure 5. Extraction Bed Effluent TOC
84
ISO
140
-------
phenomena was expected, based on the
previous laboratory sludge extraction
studies. In addition, all parameters were •
usually slightly lower for well 1 compared
to well 2. This was probably the result of
dilution from local groundwater.
Influent and effluent volumetric
fluxes for the treatment bed varied between
0.5 Inches/day and 1.0 inches/day, depending
on mode of operation and local rainfall.
Hydraulic response was rapid and recovery
of influent was almost quantitative
throughout the period of operation.
Influent and effluent TOC for the
treatment bed are presented in Figures 6
and 7. The peak observed in effluent TOC
between days 70 and 90 is the response
expected during 'development and acclimation
of the microbial population. This peak was
accompanied by a decrease and subsequent
increase in effluent pH. -This was an
expected response of'the system also. Both
of these phenomena were observed in
previous laboratory column experiments.
Integrated TOC reduction for the entire
period of operation was greater than 95% on1
a mass basis. Steady state TOC reduction,
after microbial population development was
greater than 98%.
Effluent from the treatment bed was
accumulated in tank 4 prior to recycle or-
discharge. Throughout operation, process
effluent pH was between 6 and 8, process
effluent TOC and TOS were less than 50 mg/i
and 900 mg/fc, respectively.
ACKNOWLEDGEMENT AND DISCLAIMER
The work described 1n this paper was
funded in part by tl.S.E.P.A. (Edison, N.J.)
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.
REFERENCES , . ... ,".
1, Kosson, O.S. and.R.C. Ahlert, "In-Sitcr
and On-Site Biodegradation of Industrial
Landfill Leachate,'" Environmental
Progress* 3, 176-183, August 1984.
2. Kosson, D.S., E.A. Dienemann and R.C.
Ahlert, "Characterization and Treata-
billty Studies on Industrial Landfill
Leachate (Kin-Buc I)," 39th Annual
Purdue Industrial Waste Conference,
W. Lafayette, IN, May 1984,
3. Kosson, D.S., E.A. D.ienemann and R.C.
Ahlert, "Treatment o'f Hazardous Land-
fill Leachates Utilising In-S1tu
Microbial Degradation, Part II,"
Hazardous Uastes and Environmental
Emergencies* HMCRI, Houston, TX,
289-292, March 1984.;
4.. Boyer, J.D., M.B. King, O.S. Kosson and
•R.C. Ahlert, "Aerobic Biodeqfadation of
Leachate and Forced Extracts from a
Sludge Disposal Lagoon," Toxic and
Hazardous Mastes* Te'chnomic Publishing
Co., Lancaster, PA, 497-508, 1985.
APPENDIX '
AEROBIC REMOVAL OF ORGANIC CARBON
Approach
Aerobic biodegradation of forced and
natural leachates can be used to eliminate
some of the organic compounds present in
solution. It has been observed that the
mixed microbial populations are capable of
.metabolizing organic compounds [1] found In
leachates from industrial landfills. The
microMal transformation of. supposedly
xenobiotic compounds can b^ incorporated
into more usual and naturally occurring
biochemical pathways. Metabolism may be of
two types [2]: '-....
J. A .chemical., that is-easily
metabol ized .can act as an
energy source and"support
. growth as we! 1; y "'
2. A chemical does riot serve
'** ' as a source of 'growth but
• ;_ can be. metabol i zed
•2(cometabo11sm).'" :
Cometabolism may play a significant
role in the degradation of anthropogenic
"compounds. This process is possibly a
result of the limited substrate specificity
for some microbial- transport mechanisms and
enzymes.^ -". -->
Kosson et al. [3] Investigated the
effects of initial glucose concentration on
the biodegradation of natural and forced
extracts. He reported that initial glucose
concentration did not effect percent
85
-------
NUTRIENT TANK
TANK 3 (to tr«o
ti"i
M
II
r*
£
3
-^
^
g
V*
g
1
1
**
*•.* -
M~
1.8 •
1.8 -
1.4-
1.2 -
1 -
o.a -
o.e •
0.4 -
o.x -
O_
100 -r
90 -
80 -
7O-
8O -
6O-
4O -
30-
SO-
1O-
O
o
t» D
• a
a o D
a
a
a
a °
*^ D
^i^
0
D
a
o so 40 eo ao too 120 1
TlUB (DAYS)
Figure 6. Treatment Bed Influent TOC
TREATMENT BED
WfLL 3
a
a
o
a a
a ° Q
o
a
aa
o ocPQ a
, ,_ __
• iiiitiiiiiik
20 40 80 8O 10O 120 14
TOfS (DAYS)
40
0
Figure 7. Treatment Bed Effluent TOC
86
-------
reduction of organic species. The effects
glucose has on the m1erob1al population and
Its ability to blodegrade natural and
forced extracts were examined more closely.
Procedure
A stock culture of micro-organisms,
acclimated to a feed solution containing
both glucose and leachate organic carbon
(GOC and IOC, respectively) was used. This
culture was derived from the secondary
sludge of a municipal sewage treatment
plant. Feed for the stock culture had a
1:1 ratio of GOC to LOC. The stock culture
was fed 12 hours prior to running an
experiment. This caused the microbes to be
1n the growth phase.
Leachate samples were aerated before
experimentation to stabilize pH. Stripping
experiments showed that no organic carbon
was lost. Then. pH was adjusted from
approximately 9.0 to 7.5 with 1M K2HP04 and
1H KH2P04. All experimental solutions were
prepared from the following constltutents:
500 mg C/i from LOC {as possible)
5% by volume buffer (1M K2HP04 and
1M KH£P04 mixed to obtain a pH of 7.5)
625 mg/Sl
-------
extract may be too complex for microorganisms
to metabolize. The secondary extract
originated from biological Industrial
sludge. The chemical constituents had
already undergone some blodegradatlon, and
may be more suitable for metabolic processing.
Conclusion
The composition of natural leachate
and forced extract from a sludge disposal
lagoon 1s complex, A supplementary carbon
source was required for microblal growth.
Glucose did not affect IOC reduction.
Aerobic degradation did not reduce the
organic species more than 50i. Therefore,
conventional aerobic blodegradatlon. I.e.,
secondary treatment, may not be appropriate.
Appendix References
1. Venkataraml et al., "Role of Cometabo-
Hsm 1n Biological Oxidation of
Synthetic Compounds," Biotechnology
and Bloengineering, Vol. 27, 1306-1311,
Sept. 1985.
2. Venkataraml et al., "Biological Treat-
ment of Landfill Leachates," CKC
Crftfca? Reviews in fnvlronm&ntal
Control, Vol. 14, Issue 4, CRC Press,
Inc., 333-337, 1984.
3. Kosson, D.S. et al., "Development and
Application of On-Site Treatment
Technologies for Sludge Lagoons,"
Proceedings from First International
Conference on the New Frontiers for
Hazardous Waste Management, Pittsburgh,
PA, Sept. 1985.
4. Standard Methods for the Examination
ofHater and Haste Mater, 15th Ed.,
APHA, AWWA, WPCF, Washington, DC, 1981.
TABLE A.I. INITIAL AND FINAL GOC VALUES
FOR CONTROL EXPERIMENT
GOC (rag/*)
Initial Final
16 0
84 0
188 0
288 0
400 0
500 0
'"•—- 88
-------
TABLE A.2. INITIAL AND FINAL 60C, TOC, LOC AND % REDUCTION FOR NATURAL LEACHATES
GOC (rng/i)
Initial
0
100
200
300
400
500
Final
0
0
0
0
0
0
TOC
Initial
157
243
335
440
540
640
(rag/Jt)
Final
94.
82.
77.
75.
74.
74.
7
2
3
0
2
3
LOC
Initial
157
143
135
140
140
140
(mg/l)
Final
94.
82.
77.
75.
74.
74.
7
2
3
0
2
3
% LOC
Reduction
40.0
42.5
42.7
46.4
47.0
47.0
TABLE A.3. INITIAL AND FINAL GOC, TOC, LOC AND % LOC REDUCTION FOR SECONDARY EXTRACT
GOC (mg/i!
Initial 1
0
100
200
300
400
500
1
:1nal
0
0
0
0
0
0
TOC
Initial
756
843
900
1000
1010
1100
(mg/l)
Final
280
321
339
337
326
343
LOC
Initial
756
743
700
700
160
598
(mg/i)
Final
280
321
339
337
326
343
* LOC
Reduction
67.0
56.0
51.6
51.9
46.5
42.5
TABLE A.4. INITIAL AND FINAL GOC, TOC, LOC AND % REDUCTION FOR PRIMARY EXTRACT
GOC (mg/M
Initial Final
0 0
100 0
200 0
300 0
400 0
500 0
TOC (mg/H)
Initial Final
67 66
167 62
267 64
367 64
367 72
568 65
LOC (rag/*.)
Initial Final
67 66
67 62
67 64
67 64
67 72
67 65
% LOC
Reduction
0
0
0
0
0
0
89
-------
REPORTABLE QUANTITIES GUIDELINES FOR CERCLA - DESIGNATED CHEMICALS
K. Jack Kooyoomjian
Richard Horner
Emergency Response Division
U.S. Environmental Protection Agency
Washington, DC 20460
Richard Field
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
Steve Gibson
Milton Kirsch
Gregory Ricci
Environmental Monitoring and Services, Inc.
Combustion Engineering Corp,
Newbury Park, CA 91320
ABSTRACT
This paper provides the technical methodology the U.S. Environmental Protection Agency
(EPA) has used to adjust reportable quantities (RQs) of hazardous substances, which when
released must be reported to the National Response Center. In accordance with CERCLA
Section 102, the EPA Administrator must promulgate regulations to establish the level
of release which triggers a response. The methodology consists of evaluating the
intrinsic physical, chemical and toxicological properties of the hazardous substance.
For each property a ranking scale was established, from which an RQ could be derived.
The final RQ for each substance is then considered for adjustment based on the
substance's degradation properties (Biochemical degradation, Hydrolysis, Photoylsis
(BHP)). The Administrator issued the Final Rule on April A, 1985 as 40 CFR Part 302.
90
-------
INTRODUCTION
CERCLA* (1980) requires immediate
notification (Section 103) from any person
in charge of a vessel or an offshore or
onshore facility who releases an amount of '
a hazardous substance equal to or greater ;
than its Reportable Quantity (RQ). The act
defines hazardous substances (Section
101<14)) and sets statutory RQs (Section
102(b)) for those hazardous substances at
one pound except those for which a
different RQ has been established pursuant
to Section 311(b)(4) of the Clean Water
Act. Section 103(a) of CERCLA requires the
persons in charge to notify the National
Response Center (NRC) as soon as they have
knowledge of the release. Once the NRC is
notified, they inform the predesignated
On-Scene Coordinator (OSC) pursuant to the
National Contingency Plan (40 CFR 300) of
the release. The OSC within the U.S. Envi-
ronmental Protection Agency (EPA) or the
Coast Suard then has; an opportunity to de-
cide whether the Federal government needs.
to respond to the release or threatened
release of a hazardous substance.
On April 4, 1985, EPA promulgated a
final rule (40 CFR Parts 117 and 302)
adjusting the statutory RQs and codifying
the hazardous substances defined by CERCLA.
The issuance of this rule coincidentally
provides a cost savings of about $17
million annually (EPA 1985a, Lowden 1986).
Thus the goal of setting RQs at a level
that would reflect the proper multimedia
toxicity and other criteria and would
result in few unnecessary reports was
realized.
While developing the technical basis
for RQ adjustments, an analysis was
performed of the threats that released
hazardous substances might have on human
health, welfare and the environment. It
became clear that two types of criteria
needed to be considered:
1. Those criteria characterized by
some acute or immediate threat,
such as fatal poisoning of mammals
or aquatic organisms, or disastrous
fires, explosions or otherwise
violent reactions.
* Comprehensive Environment Response,
Compensation and Liability Act of
1980, also known as CERCLA or
SUPERFUND.
2. Those criteria characterized by a
threat that is not immediately
identified but which have potential
for long term insults to humans
upon repeated and/or continuous
exposure.
The first criteria consider the intrinsic
chemical and physical properties of the
subtances, such as aquatic and mammalian
toxicity, ignitability and reactivity.
The second criteria consider the
chronic toxlcicity and potential
carcinogenic!ty of the hazardous
substances. A paper discussing the
methodology is given at this meeting
(Kooyoomjian 1986).
The present paper addresses the
technical methodology based upon the first
•type of criteria and includes the EPA
^technical methodology for (1) codifying the
CERCLA hazardous substances, and (2)
adjusting the statutory RQs.
BACKGROUND
In January 1981, less than one month
after the President had signed CERCLA into
law, EPA initiated the RQ rulemaking
process and assigned responsibility to the
Emergency Response Division, Office of
Emergency and Remedial Response. With
additional technical assistance from the
staff of the Releases Control Branch, in
EPA's Office of Research and Development,
Hazardous Waste Engineering Research
Laboratory in Edison, New Jersey, a program
was outlined to develop the CERCLA
hazardous substance list and to develop
strategies for EPA management to consider
for adjusting RQs of those listed hazardous
substances. The methodology used to
develop RQs must be scientifically sound
and sufficiently protective of the public
health, welfare, and the environment.
During 1981 and 1982, a number of
strategies and options were developed and
considered in keeping with the guidelines.
The scientific and programmatic aspects of
each strategy were tested. These tests
consisted of (1) degree of subjectivity
required for assigning firm numbers to the
RQs, (2) scientific validity, (3) ease of
implementation, and (4) degree of
complexity in developing the final RQs.
Concurrently, Regulatory Impact Analyses
were conducted for each of the candidate
strategies.
91
-------
On May 25, 1983, EPA published a
proposed rule (48 FR 23552) on the
Notification Requirements and strategies
for adjusting RQs for hazardous substances
which did not exhibit potential chronic
toxicity or potential carcinogenic
properties. Table 302.4 of the rule
contained the first codified list of the
CERCLA hazardous substances. Public
comments vere requested and 136 comment
letters totalling over 1000 pages were
received.
After analyzing each comment and using
the latest technical data for RQ
adjustment, a Final Rule was published on
April 4, 1985 (50 FR 13456). This rule
promulgated RQ adjustments for 340
hazardous substances. Concurrently, EPA
published a Notice of Proposal Rulemaking
to adjust RQs for 105 additional CERCLA
hazardous substances. Still in process are
hazardous substances that are being
considered for RQ adjustment mostly on the
basis of potential carcinogen!city and/or
chronic toxicity.
CODIFICATION OF CERCLA HAZARDOUS SUBSTANCES
CERCLA Hazardous Substance Definition
In CERCLA Section 101(14) a hazardous
substance is defined as:
1. Any substance designated pursuant
to Section 311(b) of the Clean
Water Act.
2. Any hazardous waste having the
characteristics identified under or
listed pursuant to Section 3001 of
the Solid Waste Disposal Act.
3. Any toxic pollutant listed under
Section 307(a) of the Clean Water
Act.
4. Any hazardous air pollutant
listed under Section 112 of the
Clean Air Act.
5. Any Imminent hazardous chemical
substance or mixture with respect
to which the Administrator has
taken action pursuant to Section 7
of the Toxic Substances Control
Act.
6. Any element, compound, mixture,
solution or substance the
Administrator determines to be
hazardous pursuant to Section 102
of CERCLA.
The Strategy for Developing the List
1. In 40 CFR 117 EPA has designated
and assigned RQs to 297 hazardous
substances pursuant to Section 311
of the Clean Water Act. These
hazardous substances were included
on the CE1CLA list.
2. In 40 CFR 261, 429 hazardous wastes
were designated pursuant to Section
3001 of RCRA. These consist of 107
acutely hazardous wastes, the "P"
list? 233 toxic wastes, the "U"
list; 13 hazardous wastes from
non-specific waste stream sources,
the "F" list; and 76 hazardous
wastes from specific waste stream
sources, the "K" list. In
addition, there is another RCRA
group of wastes called "unlisted"
or "ICRE" wastes (Ignilability,
Corrosivity, Reactivity, EP
Toxicity) that are hazardous
substances under CERCLA if they
meet specific test criteria
described in 40 CFR 261. The
number of substances that fall into
this unlisted hazardous waste
category cannot be quantified.
3. There are 65 compounds and classes
of compounds designated under
Section 307(a) of the Clean Water
Act. The BPA has identified 126
specific compounds as priority
pollutants that fall within these
65 compounds and classes of
compounds. The CERCLA list
includes these 126 specific
priority pollutants and the broad
generic classes named among the 65
compounds and classes of compounds.
4. EPA has designated eight substances
in 40 CFR 61 pursuant to Section
112 of the Clean Air Act. These
substances are arsenic, asbestos,
benzene, beryllium, coke oven
emissions, mercury, radionuclldes,
and vinyl chloride. Six of the
substances are also on the Clean
Water Act list.
5. No Imminently hazardous chemicals
have been designated pursuant to
Section 7 of TSCA.
92
-------
6. The Administrator has not, at this
time, designated any additional
hazardous substances pursuant to
the authority of CERCLA Section
102(3), except to reaffirm the
current CERCLA list on April 4,
1985 in 40 CFR 302.4(a).
After removing duplicates, there are
705 hazardous chemicals and wastes on the
CERCLA list. (As of the date of the Final
Rule (4 April 1985) there vere 698, but the
list was expanded on January 14, 1985 and
effective in July 1985 to include specific
dioxin waste streams.) Congress intended
that any modifications EPA made to the
various regulations pursuant to CWA, RCRA,
.CM, and TSCA would automatically modify
the CERCLA list. If EPA delisted any waste
from RCRA, for example, it would be
delisted from CERCLA provided that it was
not listed under CERCLA or by one of the
•other statues identified above. If EPA
added a waste or chemical to RCRA or the
other listing authorities, it would
automatically be a CERCLA hazardous
substance and its RQ would be one pound
until adjusted by EPA.
The Agency dealt with broad generic
classes of organic and metallic compounds
designated as toxic pollutants under
Section 307 of the Clean Water Act by not
requiring reporting of a release of any
member of such broad classes as
"chlorinated phenols", "phthalate esters",
"polynuclear aromatic compounds", "zinc and
compounds", and others. Many of the
generic classes of compounds encompass
hundreds or even thousands of specific
compounds making the task of adjusting an
RQ for each broad class virtually
impossible considering the varying
characteristics of all the specific
compounds in the class. The Agency,
therefore, assigned RQs to specific
compounds within the broad generic class.
Accordingly, only specific chemical
compounds and identified RCRA wastes need
:be reported when released in amounts equal
'to or greater than their RQs.
EPA further proposed that the 12 solid
metals originally listed in Section 307(a)
of the Clean Water Act would be reportable ,
under CERCLA if their particle size was 100
micrometers or less. The approach was to
establish a cutoff size 10 times larger '
than the maximum size considered by EPA to
be respirable to insure that releases into
the environment containing small particles
of metals tfould result in prompt
notification to the NRG. Thus the
inadvertent release of metal ingots and
similar metals need not be reported to the
NRC.
One of the issues in presenting the
list to the regulated community was the
selection of nomenclature for the hazardous
substances. There were three options
available! (1) use the names of the
chemicals as they appear in the
environmental statutes and implementing
regulations; (2) use the Chemical Abstracts
Collective Index System namei or (3) use
the major synonyms for each hazardous
substance.
TECHNICAL METHODOLOGY FOR RQ ADJUSTMENT
The CERCLA RQ list includes 705
hazardous substances. Of these, 297
chemicals designated as hazardous
substances pursuant to the Clean Water Act
Section 311 have previously been assigned
RQs of either 1, 10, 100, 1000, or 5000
pounds based on their level of aquatic
toxicity (EPA 1975). For the remaining 408
hazardous substances, CERCLA assigns a
statutory RQ of one pound. These statutory
RQs were intended by Congress to be of
temporary duration, pending EPA review and
adjustment of the RQs. The methodology
which was adopted by EPA and presented in
this paper explains the framework for
adjusting the RQs for the CERCLA hazardous
substances.
The Primary Criteria RQ Adjustment
Methodology(EPA 19155)
The RQ adjustment methodology uses six
environmental and human health and welfare
criteria! aquatic toxicity, mammalian
(oral, dermal, and Inhalation) toxicity,
ignitability, reactivity, chronic toxicity,
and potential carcinogeniclty. It examines
each CERCLA hazardous substance for the
likely hazards posed to human health or
welfare or the environment as defined by
these criteria, and associates an RQ with
each of them, based on scientifically
supported data. In general, the suggested
RQ is the smallest one that is associated
with any one of the hazards which a
material may pose.
The following discussion addresses the
primary criteria for adjusting RQs of
93
-------
hazardous substances based on their
intrinsic chemical, physical, and acute
toxic properties, i.e., aquatic and
mammalian toxieity, ignilability and
reactivity. Strategies for adjusting RQs
based on their carcinogenic and chronic
toxicity properties are given in a
companion paper (Kooyoomjian 1986).
Aquatic and Mammalian Toxicity. The
five-level RQ rating system derived for
aquatic toxicity (Table 1) under CWA
Section 311 consists of the following five
RQ categories, linked to the specified
aquatic toxicity ratings:
Table 1. Sating System - Aquatic Toxicity
Category Aquatic Toxicity
X LG50 < 0.1 mg/L
Founds
A 0,1 mg/L < LC50 < 1 mg/L 10
B 1 mg/L < LC50 < 10 mg/L 100
C 10 mg/L~< LC50 < 100 mg/L 1000
D 100 mg/L < LC50 < 500 mg/L 5000
This rating system is the one referred to
in the Clean ¥ater Act final rulemaking
which appeared in 44 FR 50766-50779, August
29, 1979. Similar five-level RQ rating
scales were developed for mammalian
toxicity. These rating scales are shown in
Table 2. The rating system for mammalian
toxicity includes separate scales for oral,
dermal, and inhalation toxicities.
The data for each of these criteria
were tabulated and an RQ was associated
with the range of data by a corresponding
dimensionless letter code (i.e., Category
X, A, B, C, or D).
The toxicity ranges for Categories X,
A, B, C and D were scaled for mammalian
toxicity in the same ratio as scaled for
aquatic toxicity for CWA Section 311, based
on the following rationale. Experimental
study shows that a normal swallow by a
small child amounts to about 4 to 4.5 cubic
centimeters. For a toddler weighing 15
kilograms (kg), this translates to a dose
of 300 milligrams (mg) per kg of body
weight, assuming an ingested material of
unit density. To provide for a safety
margin (allowing for lower body weights
and/or higher density materials), a
limiting LD^Q of 500 mg per kg of body
weight was cnosen. For reference purposes,
sodium chloride (common salt), has an
L05Q of 3000 rag/kg.
Applying this same series of
calculations to the "standard man" (70 kg
body weight, swallow volume of 21 cubic
centimeters) also yields a value of 500 mg
per kg of body weight for the limiting
LDcQ. However, adults are much less
liRely to ingest foreign materials than are
children. It should be noted that this
500mg/kg LDcQ value is the upper limit of
an LD50 range of 100-500, and corresponds
to a suggested RQ assignment of the maximum
value of 5000 pounds.
The upper-bound toxicity limits of 200
mg/kg for dermal toxicity and 2000 ppm for
inhalation toxicity were taken from the
National Academy of Sciences System for
Evaluation of Hazards of Bulk Water
Transportation of Industrial Chemicals
(see: A Summary of Hazardous Substance
Classification Systems, EPA/530/SH-171).
The many sources of aquatic and mammalian
toxicity data are given in EPA 1983 and EPA
1985b.
Ignitability and Reactivity Evaluations
The starting point for the scales used
to adjust RQs based on the primary criteria
of ignilability, corrosivity, and
reactivity have their origin in work done
by the U.S. National Academy of Sciences
(WAS) for the U.S. Coast Guard (USCG 1974).
The methodology developed for RQ adjustment
follows.
a. The Ignitability scale was adjusted
upward to accommodate an even more
hazardous category. Pyrophoric
substances which ignite spontaneously
on contact with air, are considered
the most hazardous and are assigned an
RQ of ten pounds. The other NAS
categories are moved up one level on
the ignitability scale.
b. The Reactivity scale was used with
slight modification,
c. Corrosivity was not used as a
criterion for adjusting RQs under
CERCLA because EPA did not develop or
identify an appropriate scale for
rating corrosivity. The RCRA
regulations define the characteristic
of corrosivity in terms of pH ranges
and rate of attack on steel (40 CFR
261.22). The pH ranges apply only to
aqueous wastes and the rate of attack
on steel test applies only to liquids.
94
-------
If EPA used a corrosivity scale based
on these tests, corrosivity could not
be used to adjust RQs when the
hazardous substances were neither in
aqueous solutions nor were liquids,
despite the fact that such substances
might pose hazards due to corrosivity.
The Agency attempted to apply other
tests and existing data in developing
a corrosivity scale, but was unable to
develop a five-tier scale that
addressed the corrosivity of all
CBRCLA hazardous substances.
Ignitability
The MAS scale does not take into
account self-ignitability. This
characteristic is considered even more
hazardous than simply having a very low
flash point, since it is capable of
starting a fire on its own. Three types of
compounds have this characteristic, and are
classed as follows!
Pyrophorics! Those compounds which
immediately ignite because of rapid
oxidation as soon as they contact oxygen.
Spontaneously Ignitable; Those
compounds that may not immediately ignite,
but that can do so if spread in a thin
layer, or brought into contact with iron
oxide or other catalysts.
Strong Oxidizer; Those compounds that
can cause fuel materials to ignite, and
support their combustion once they are
ignited.
Accordingly, the ability to start
fires was defined as the most hazardous;
the RQ of substances showing this ability
was assigned a value of ten pounds. The
other ignitability factors were moved up
one level on the table, discarding the
"noncombustible" category. This made the
upper category "Flash Point > 140°F",
which was equated to "Flash Point.
140 F-200°F", so that all the numerical
categories correspond to National Fire
Protection Association (NFPA) classes (NFPA
1980).
Reactivity
Two independent scales were used for
ranking reactivity characteristics. The
first is the degree of reactivity with
water, and the second is the degree of
self-reactivity. To scale the degree of
reactivity of a hazardous substance with •
iwater, known reactions of reference
^chemicals were used. The use of reference
compounds for ranking reactivity with water
is a measure of the violence of such
reactions, and is accordingly a measure of
heat release (heat of hydration, heat of
mixing) upon mixing the two. To scale the
degree of self-reaction, the
characteristics of the hazardous substance
was assessed to determine capabilities to
detonate or polymerize with heat release.
These rankings take into account all
the concerns expressed in the RCRA
description of reactivity (40 CFR Part
261.23, 40 FR 33122) except for the concern
.that certain wastes mixed with water or
acid can generate toxic gases. One case is
that of the cyanides, which can be ranked
appropriately on the basis of the CWA
Section 311 ranking for cyanides. Others,
such as those that could generate phosphine
(PH-) or hydrogen sulfide (H^S) have
been treated as special cases of
reactivity. The final ranking system based
on ignitability and reactivity is shown in
Table 3.
APPLICATION OF THE PRIMARY CRITERIA
Applying the RQ adjustment methodology
to a designated hazardous substance may
result in several different RQs for each of
the six primary criteria. For example, the
RQ based on aquatic toxicity for
crotonaldehyde is 100 pounds; but the RQ
based on mammalian toxicity is 5000 pounds.
Crotonaldehye's ignitability correlates
with a 1000 pound RQ. All other criteria
would put the RQ at 5000 pounds. Thus
there are four candidate RQs for
crotonaldehyde. Based upon the
methodology, the lowest RQ is the final
adjusted RQ, which in this case is 100
pounds.
The 'following summarizes the RQ
adjustment methodology based on the primary
criteria:
The aquatic toxicity, mammalian
toxicity, ignitability, reactivity,
chronic toxicity, and potential
carcinogenic!ty data for each
substance were evaluated. The data
for that criterion which resulted in
the lowest RQ according to the
developed scales were used as the
95
-------
basis for the proposed RQ. In
general, the RQ for each hazardous
substance was adjusted based on
specific data for the individual
substance, with the following
exceptions.
A. All cyanide compounds were assigned
a proposed RQ of 10 pounds based on
the aquatic toxieity of cyanide as
developed under Section 311 of the
Clean Water Act.
B. In instances where no data were
found for the chemicals, RQ
assignments were generally set at
the maximum level, Category D (5000
pounds) or at a level judged
appropriate on the basis of
chemical similarity to other CERCLA
hazardous substances.
OTHER ADJUSTMENTS
Certain substances may have their RQ
adjusted further after application of the
primary criteria. Natural degradative
processes may reduce the hazards posed by
their release. Three degradation processes
were considered for adjusting RQs:
biodegradabllity, hydrolysis, and
photolysis - (BHP). If a hazardous
substance is subject to degradation to a
less hazardous form then its primary
criteria RQ is raised one level, e.g. from
10 to 100 pounds, if one of the following
standards is met:
(1) The reported biochemical oxygen
demands of the substance over a
five-day period at 20 degrees Celsius
is at least 50% of the theoretical
oxygen demand or;
(2) The estimated half-life of the
substance is equal to or less than
five days when the substance is
subjected to hydrolysis or photolysis
in conjunction with biochemical
degradation (EPA 1979).
The RQ is not raised, however, when
the RQ has been adjusted to 5000 pounds
based upon the primary criteria. Also, if
the substance is a potential carcinogen,
then the RQ will not be adjusted above 100
pounds based on BBP.
FUTURE PLANS
The EPA is currently completing the RQ
adjustments. When these efforts are
complete, proposed rules will be issued and
comments on the rule will be sought.
REFERENCES
Kooyoomjian, K.J., M. Chu and C« DeRosa,
Potential Carcinogenic!ty and Chronic
Toxicity for RQ Adjustments - Proceedings
of the 1986 Hazardous Material Spill
Conference, St. Louis, MO, April, 1986.
Lowden, G. and E. Deutsch, Economic Effects
of Adjusting Reportable Quantities of
CERCLA Hazardous Substances - Proceedings
of the 1986 Hazardous Material Spill
Conference, St. Louis, MO, April, 1986.
National Fire Protection, "Flammable and
Combustible Liquid Code, NFPA No. 30, 1980.
U.S. Coast Guard, "System for Evaluation of
Hazards of Bulk Water Transportation of
Industrial Chemicals, USCG-D-113-74,
February 1974.
U.S. Environmental Protection Agency,
Technical Background Document to Support
Rulemaking, EPA-440/9-75a through d,
January, 1975.
U.S. Environmental Protection Agency, Water
Related Environmental Fate of 129 Priority
Pollutants, EPA-440/4-79-029a and b, 1979.
U.S. Environmental Protection Agency,
Background Document to Support the Notice
of Proposed Rulemaking Pursuant to CERCLA
Section 102(b), May, 1983.
U.S. Environmental Protection Agency,
Economic Analysis of Reportable Quantity
Adjustments Under Sections 102 and 103 of
the Comprehensive Environmental Response,
Compensation, and Liability Act, Volume III
- Background Report in Technical File to
Support Rulemaking 40 CFR 302, January,
1985a.
U.S. Environmental Protection Agency,
Technical Background Document to Support
Rulemaking Pursuant to CERCLA Section 102,
Volume I - Background Report in Technical
File to Support Rulemaking 40 CFR 302,
March, 1985b.
96
-------
Table 2, MAMMALIAN TOXICITY RANKING SCALE
Mammalian
Category Toxicity (Oral)
Mammalian
Toxicity (Dermal)
Mammalian loxicity
(Inhalation)
RQ (Pounds)
X
A
B
C
D
U>50 50 <.Q4 mg/kg
0.1 mg/kg
-------
Table 3. REACTIVITY-IGNITABILITY RANKING SCALE
UD
CO
Category
X
A
B
C
D
Ignitability
(Fire)
(Not used)
Pyrophoric or
Self-Ignitable
FP(cc) <100°F (37.8°C)
BP <100°F (37.8°C)
FP <100°F (37.8°C)
BP >100°F (37.8°C)
FP(cc) 100-1AO°F
(37.8-60°C)
Reactivity
With Water
(Not used)
Inflames
Extreme reaction,
e.g., S03
High reaction,
e.g., oleum
Moderate reaction,
e.g., NH3
Self-Reaction
(Not used)
Extreme self-reaction;
may cause explosion
or detonation
High; may polymerize;
requires stabilizer
Moderate; contamination
may cause polymerization;
no inhibitor required
Slight; may polymerize
with low heat release
RQ (Pounds^
1
100
100
1000
5000
Notes: FP(cc) = Flash Point, closed cup
BP " Boiling Point
-------
PREDICTING HAZARDOUS WASTE LEACHATE COMPOSITION!
Danny R. Jackson
Battelle Columbus Division
Columbus, Ohio 43201
OBJECTIVE
Information on the Teachability and
composition of leachate from wastes is
needed in the design of disposal facil-
ities and in the review of permit
applications. This project is studying
the use of laboratory batch extraction
procedures for predicting the composition
of aqueous leachate from wastes in the
field. Water was passed through large
diameter columns filled with solid wastes
to simulate waste leaching under field
conditions. The ability of selected
laboratory batch procedures to predict
total amounts of Teachable material and
maximum analyte concentrations in leach-
ates from the large columns was invest-
igated. This presentation describes the
wastes and test conditions and compares
the maximum analyte concentrations found
in the column leachates and batch ex-
tracts. A previous paper? reported
results of a comparison of waste leaching
in small columns (5 cm diameter) and in a
batch procedure.
WASTES
Five wastes, representiny different
industrial processes, were selected for
this project. These wastes included an
electroplating sludge (ES), an electric
arc foundry dust (EAFD), an automotive
paint incinerator ash (PIA), a municipal
refuse incinerator ash (MRIA), and a mine
tailing (MT). Physical and chemical prop-
erties determined for each waste included
hydraulic conductivity, bulk density,
particle size, solids content, elemental
content, and pH. Results of these anal-
yses indicate that the waste materials
represent a broad range of physical and
chemical properties.
1 The research described in this paper has been funded in part by the United States
Environmental Protection Agency through Cooperative Agreement CR-810272 with
Battelle Memorial Institute. The EPA project officer is Mike H. Roulier of the
Hazardous Waste Engineering Research Laboratory, Cincinnati, Ohio 45268.
2 Jackson, D. R., B. C. Garrett, and T. A. Bishop. 1984. Comparison of Batch and
Column Methods for Assessing the teachability of Hazardous Waste. Environmental
Science and Technology 18(9): 668-673.
99
-------
WASTE EXTRACTION PROCEDURES
Five laboratory batch procedures
were used to extract each of the waste
materials. These procedures included
the EPA Extraction Procedure (EP), the
proposed Toxicity Characteristic Leaching
Procedure (TCLP), Solid Waste Leaching
Procedure (SWLP), Ham Procedure-C
(Ham-C), and Saturated Paste (Paste).
These procedures represent two types of
leaching media, and four liquid to solid
ratios. The greatest differences between
the methods in terms of analytes extract-
ed from the wastes were attributed to the
leaching media. There were both qual-
itative and quantitative differences
in the efficiency of distilled water and
acetic acid as extractants for analytes
in the wastes. The batch procedures use
100 g of waste and varying amounts of
extracting liquids. The extractions,
except for the paste method, are conduct-
ed in a National Bureau of Standards
rotating extractor. The paste method is
conducted in a glass beaker.
COLUMN STUDIES
High density polyethylene tanks,
having dimensions of 38 cm dia x 120 cm
height, were used to contain the waste
materials. Acid-washed gravel and sand
were placed in the bottoms of the tanks
to act as a prefilter to prevent finely
divided particulates from passing into
the leachate collection bags. The
particle sizes of the wastes ranged
from greater than 2.0 cm to less than
0.005 mm. Particles larger than 2.5 cm
diameter were removed from the wastes
prior to loading the columns. Individual
columns were filled with air-dried
samples of waste by weighing portions of
waste sufficient to fill 12 cm of column
depth. Each weighed portion was compact-
ed if necessary before adding additional
material. The conical bottoms of the
columns were fitted with tygon tubes
leading to Tedlar bags for leachate
collection.
The wastes were initially saturated
by aplying deionized water from the base
of the columns. For subsequent leaching,
deionized water was applied to the top
surface of the waste columns for 8 hours
at 2-week intervals. Free drainage and
partial aeration of the upper portions of
the columns were allowed during the periods
between water applications. Leachate was
collected in Tedlar bags and withdrawn
using a syringe for taking sample aliquots.
Each leachate sample was filtered and acid-
ified for metal analysis. A second sample
was tested for pH and electrical con-
ductivity (EC) immediately after it was
removed from the Tedlar bag. Total organic
carbon (TOC) was determined on unpreserved
samples shortly after they were removed
from the bags.
RESULTS AND CONCLUSIONS
The large columns were of a manage-
able size, accomodated a waste particle
size up to 2.5 cm in diameter and offered
a high degree of flexibility for waste
leachate studies. Analyte concentrations
found in the leachate were reproducibile
among the replicate columns. Leaching
profiles were obtained for each waste to
evaluate maximum analyte concentrations
and characteristic leaching patterns.
These columns can produce quantities of
leachate to support ancillary projects
such as liner compatibility or bioassay
studies.
Analyte concentrations produced from
the batch extraction methods were compared
to maximum analyte concentrations produced
by the large columns. Batch extraction
methods using deionized water and tumbling
produced results most representative of
maximum analyte concentrations found in the
leachate from large columns. Extraction
methods using acetic acid as the extracting
medium were least representative of the
maximum analyte concentrations found in the
large column leachates.
100
-------
CH01ICAL RESISTANCE OP SYMIUKTIC
WASTE CCMBINATICNS
Gordon Bellen and Rebecca Carry
National Sanitation Foundation
3475 Plymouth Road
Ann Arbor, MI 48105
ABSERftCT
Laboratory testing was conducted under the sponsorship of the Environmental Protec-
tion Agency (EPA) to study the chemical resistance of flexible membrane liner materials
(EMLs). Immersion tests were performed with six FMLs and twenty chemical solutions to
develop liner selection guidance for EML users. The effect of time, temperature, and
chemical concentration were studied. The length of immersion time ranged from one day to
two years. To evaluate the effect of these challenges on the FMLs, weight, dimension,
tensile properties, and tear resistance were measured. The results of the study were
used to develop general principles for interpreting FML chemical immersion tests. This
paper presents results and discusses interpretation principles for one ML material, a
plasticized thermoplastic.
effect of a small amount of an incompat-
ible chemical, Figure 1 compares the
change in weight of a thermoplastic FML
immersed in two solutions of very differ-
ent concentrations. Percent weight change
is shown as a function of immersion time.
While a saturated brine solution (approx-
imately 35% by weight) causes very minor
changes in FML weight, a, 0.5% solution of
1,2-dichloroethane results in much greater
weight gain.
Por these reasons, the EPA requires
that FML selection be based on evaluation
of changes in physical properties result-
ing from immersion in the actual waste to
be contained. Several similar immersion
test methods have been developed (6) to
determine chemical resistance, with
guidance on duration, temperature, and
types of tests to be performed. Evaluation
of the test results has not been well
defined, however. The EPA considers "any
significant deterioration in any of the
neasured properties to be evidence of
incompatibility unless a convincing demon-
stration can be made that the deteriora-
Flexible membrane liners (FMLs) are
used increasingly as lining materials for
hazardous waste containment, in landfills
and surface impoundments. A double-liner
system, including at least one synthetic
liner, is required in all new installa-
tions. (7) The ML used in a waste con-
tainment application must show long-term
chemical resistance to the waste stream.
Waste streams are mixtures of chemi-
cal substances, with some chemical compon-
ents present in small or trace quantities.
The presence of some trace components may
be unknown at the time a liner is being
selected. Published chemical resistance
tables generally list only pure compon-
ents, or mixtures of one component in
water. A small amount of a substance
deleterious to the liner could be present
in a mixture whose major component has no
effect. Looking only at the effect of the
major component in tables, the liner would
appear to be resistant. The presence of
the incompatible chemical could mean the
difference between success and failure of
the installation. As an example of the
101
-------
tion exhibited will not impair the liner
integrity over the life of the facility"
(9).
Not enough guidance information is
available for EML users, who must select
the most appropriate lining material and
demonstrate its resistance to the wastes.
Theoretical methods of predicting chemical
resistance such as Hansen solubility
parameters and cohesive energy density
numbers (2) are not yet well developed for
IMLs, and expert systems for data in-
terpretation (43 are not generally avail-
able. Acceptable changes in a physical
property such as tensile strength at a
given temperature and immersion time may
be different for different EML materials.
This project was initiated by the EPA
(cooperative agreement £CR 810-727-01-0)
to help develop chemical resistance
selection guidance information for EML
users.
In this project, immersion testing of
six EMLs was conducted at two temperatures
with a broad range of chemical exposures.
EML property changes were measured for
exposure periods from one day to two
years. Results were studied to determine
the basic IML responses to combinations of
chemical challenge, concentration temper-
ature, and time. The focus in data inter-
pretation was on the types of degradation
encountered, stabilization of the material
response, the extent of property change,
and indicators of non-resistance (5), The
method of interpretation can then be gen-
eralized to provide guidance to EML users
testing FMLs with specific waste streams.
This paper presents a discussion of
chemical immersion results for one thermo-
plastic liner material as an example of
determining chemical resistance using
immersion testing. The EML is not named in
order to focus attention on the methodol-
ogy of data evaluation rather than on the
specific material.
ftPBRCBCH
Objectives
The purpose of the project was to
investigate the chemical resistance of
flexible membrane liner materials, in
order to help liner users select the most
appropriate liner for their waste
containment application. To accomplish
this, the following objectives were
established:
1. Expose liner materials to a range
of chemicals under controlled
laboratory conditions.
2. Measure changes in weight,
dimension, tensile properties, and
tear resistance of these exposed
liners.
3. Study the FML response (changes in
properties) to chemical challenge
and develop criteria for chemical
resistance.
4. For each FML, generalize the types
and degree of changes observed,
and the criteria for chemical
resistance so that they can be
applied to the evaluation of other
data.
Methods and Materials
Eaqaerimental Structure. The testing
lasted for twenty-eight months, and can be
divided into two main sections: measure-
ment of changes in mechanical properties,
and measurement of long-term weight and
dimension changes. The testing protocol
was similar to EPA Method 9090 in some
respects, but not identical to it (6).
Method 9090 had not been adopted at the
time this project was begun.
1. Mechanical Property Changes.
Weight and dimension, tensile
properties, and tear resistance
were measured before and after
immersion periods of 1, 7, 14,
28, and 56 days, and percent
changes were determined.
2. Long-term Weight and Dimension
Changes. Each FML was also
measured for changes in weight
and dimension every four months
for up to two years immersion
time. At the end of the test
period, tensile strength and
elongation were, measured.
FML Materials. Six FML materials,
donated by six manufacturers, were tested
in this project. All materials except one
were 30 mils thick (0.03"). The sixth was
not available in 30 mil thickness, and a
60 mil liner was used. Only unsupported
FML liners were tested (material with .no
reinforcing scrim). Two crosslinked
rubber, one unvulcanized rubber, two
thermoplastics, and one partially
102
-------
crystalline thermoplastic were tested.
Chemicals. Twenty chemical exposures
were used in ML chemical resistance test-
ing. The goal in selecting the chemical
types was to have a broad range of chemi-
cals found in hazardous waste streams.
Because a 'typical1 waste is impossible to
identify or produce, and because studying
the liner response to chemical challenge
was a project goal, a solution of one
chemical in water was chosen for each
test. The three concentrations of each of
the organic chemicals were selected based
on results of initial screening tests.
Table 1 shows the chemical exposures used
for testing ML chemical resistance.
Test Procedures. Samples of each EML
were cut for chemical immersion. The two
year immersion samples (three (3) one inch
by three inch coupons) were cut with the
long side parallel to the machine direc-
tion. The shorter—term samples, for
mechanical property testing, consisted of
three coupons; two were approximately
eight by eight and one-half inches in
size, and the third was the same as the
two-year samples. This small coupon was
used as the indicator for weight and
dimensional changes; and the larger
coupons were used to cut specimens for
mechanical property changes after
immersion.
All of the small die-cut specimens
were weighed and measured for thickness,
width, and length prior to immersion.
Stall holes were punched in the top
corners of the samples, and they were hung
in immersion jars using teflon cord. Quart
mason jars were used for the two-year
immersion samples and two-gallon glass
jars were used for the larger sets
(shorter term samples). All immersions
were conducted with a liner surface-to-
volume ratio of approximately 40 milli-
liters per square inch. Immersions were
incubated at two temperatures, 23° and 50°
Centigrade. •
After the specified length of
immersion time, sample jars were removed
from the incubation chambers and samples
removed. Low temperature samples were
removed from the solutions, blotted, and
cleaned or rinsed if necessary (oily
exposures, acid and basic solutions) and
placed in plastic bags. High temperature
samples were reimmersed in a 23° C
solution of the same composition for at
least one hour before bagging per ASTM
D543 (1). The small coupons were weighed
and dimensioned. Two year immersion
coupons were then returned to a fresh
solution of the same composition for more
incubation.
Tear resistance and tensile property
specimens were then cut from the larger of
the shorter-term samples. Five of each
type were cut in both the machine direc-
tion and the transverse direction. Tensile
property (S-100 modulus and breaking
factor for the material discussed here)
and tear resistance tests were performed
on the samples. ASTM methods appropriate
to the type of liner material were used,
specifying specimen shapes and test
conditions.
Changes in EML weight and dimension
are expressed in percent change (as
compared to initial values). Changes-in
tensile properties and tear resistance are
expressed as percent retention of original
property (as measured by testing unexposed
control samples).
RESULTS AND DISCUSSION
Treatment o£ Data
Mechanical property data is presented
for a 30 mil plasticized thermoplastic
liner cut in the machine direction. The
machine direction was chosen in order- to
allow presentation- of the two-year tensile
property results (for which there were
only machine direction samples). In
general, strength is higher in the machine
direction than in the transverse direction
for this material, but the response to
chemical challenge is similar. Mechanical
property changes are given as percent
retention of initial value (unexposed
control samples).
Weight is used as the indicator for
swelling responses. Weight and volume
changes are closely related, and because
the weight measurement is more precise,
weight change was chosen for presentation.
Weight changes are expressed as percent
change (compared to initial weight).
Table 1 gives a summary of the resis-
tance of this EML to the chemicals tested.
Also listed in Table 1 are the types of
response seen, whether there were tempera-
103
-------
ture or chemical concentration effects,
whether the FML response stabilized within
the test period, and the reasons for
non-resistance. The procedure followed in
determining the chemical resistance is
explained in the following section.
Criteria for Determining EML Chemical
Resistance
Parameters for evaluating ML chemi-
cal resistance include: stability of the
liner's properties in contact with the
waste, the magnitude of physical changes
observed, and the type of change. For this
ML, based on the overall response of the
indicators in this project, the following
criteria are proposed:
1. Stability of weight change and
mechanical properties with tine.
2. A stabilized weight change of not
more than five percent gain or ten
percent loss.
3. The breaking factor must be at
least 80 percent of the initial
value and equal to or greater than
the minimum as-received value in
the material properties table of
NSP Standard 54 (3).
4. Ihe percent elongation at break
must be at least 70 percent of the
initial value and equal to or
greater than the minimum as-
received value in the material
property table of NSF Standard 54.
5. The S-100 modulus must be between
60 and 140 percent of the initial
value (based on the relationship
of modulus to breaking factor and
elongation).
These criteria are used in evaluating
the FML's response to the chemicals in
Sable 1, Resistance or non-resistance is
listed in the second to last columns, and
the reasons for a non-resistance rating
are given in the last column.
Besponse Time and Stability
The stabilization of an FML's
physical property changes in response to a
chemical challenge is an important param-
eter in the evaluation of chemical
resistance. The response must stabilize in
order for the FML to be considered chemi-
cally resistant.
Response time, or the time required
for the liner to stabilize in a given
solution, varied from less than one week
to more than two years for this FML.
Response time was generally faster for
weight gain (absorption of chemical or
water) than for weight loss (extraction of
plasticizer).
Magnitude of Change
A two-part evaluation of the magni-
tude of physical property change was
followed in determining FML chemical
resistance. Both an absolute minimum value
for a mechanical property (for example,
elongation at break) and an allowable
percent change of that property (from the
initial value) were detemined.
absolute amount. There are no pre-
viously established or accepted benchmarks
of performance based on immersion tests.
The NSF Standard 54 physical property
values suggest a possible benchmark which
correlates well with existing chemical
resistance data. For liners tested in this
study for which there were material
properties tables in NSF Standard 54,
results of liner strength tests (e.g.,
breaking factor, etc.) generally fell
below Standard 54 values after immersion
in chemicals for which the liner was known
to be non-resistant. The EPA draft
guidance document suggests that the NSF
Standard 54 specifications be used as
absolute minimums for physical properties
in immersion tests (9).
Using Standard 54 values for virgin
material to evaluate the physical proper-
ties of immersed material is further
justified when one considers the increased
variability in physical properties
measurement after chemical immersion. For
the thermoplastic material discussed in
this paper, the coefficient of variation
(5/50 for breaking factor was 2.5 percent
for virgin material, and as high as 11.5
percent for the FML which had stabilized
in 4 percent furfural. To pass Standard 54
the mean breaking factor for five speci-
mens must be at or above the minimum value
in the physical properties table. Compar-
ing the average value of test results to a
Standard 54 specification as a pass/fail
criteria implies that 96 percent of the
liner coupons tested are within HH 2s of
the Standard 54 value. If the Standard 54
value was 50 lb/in., 2s would equal 2.5
Ib/in. for the unexposed EML and 11.5
104
-------
Ib/in. for the ML after immersion in 4
percent furfural. This means that, if the
mean results for unexposed and immersed
coupons are. equal, 2 percent of the unex-
posed liner could have a breaking factor
less than 47.5 Ib/in. and still be accept-
able, but 2 percent of the exposed liner
could have a breaking factor less than
38.5 Ib/in. Therefore when increasing
material variability after immersion is
taken into account, the Standard 54 value
is not too strict for evaluating immersed
samples.
Percent Change fron Initial Value. A
maximum allowable change in physical pro-
perties is also necessary for evaluating
chemical resistance for three reasons. If
an FML's initial properties are much
higher than the Standard 54 values, sub-
stantial material degradation could occur
before the Standard 54 limit would be
passed. Also, Standard 54 values are
minimums, and give no guidance in cases
where a property (such as strength)
increases as a result of chemical
exposure. Finally, acceptable changes in
weight and volume can not be determined
from Standard 54.
To determine reasonable percent
change limits for chemical resistance
evaluation, three steps were followed.
First, the variability 'of the property was
evaluated. For example, Figure 2 shows the
overall relationship of breaking strength
and weight change (the figure shows weight
changes between -30 and +30 percent weight
change) for this ML. The variability of
breaking strength measurements for this
ML was roughly +_ 15 percent, as measured
by the center locus of points that gener-
ally showed resistant behavior. Next, the
limits of variability defined in step one,
and the overall range of property change
observed were compared to known chemical
resistance information. Limits of accept-
able change were then set. Third, limits
for parameters such as percent weight
change were set by determining the amount
of change that correlated with the limits
set for those properties listed in the
Standard 54 tables.
Indicators of Non-Resistance
For this ML, the most important
indicators of non-resistance are weight
change, breaking strength, S—100 modulus,
and elongation at break. Tear resistance
generally followed the same trend as
breaking strength.
Changes in weight corresponded to
changes in tensile strength and stiffness
(S-100 modulus) for this liner. (The
exception seen for this FML was for immer-
sion in hot hydrochloric acid, which
caused weight gain but no loss of
strength.) In general, strength and
modulus decreased with weight gain, and
increased with weight loss.
When absorption of liquid (weight gain) is
the response to chemical challenge, weight
change, breaking strength, and modulus are
key indicators of chemical resistance.
Weight changes of about 5 percent corres-
pond to a 20 percent loss of breaking
strength, and a strength less than the
minimum acceptable as-received value for
this FML according to NSF Standard 54
(Figure 2).
Weight loss indicates loss of plasti-
cizer from the liner material, ftn ML for-
mulation of this material can contain up
to 35 percent plasticizer. For extraction
of plasticizer, changes in weight and in
modulus are particularly important. The
beginning of plasticizer extraction (which
can continue until the material becomes
brittle) can be detected earlier by these
measurements than by breaking strength or
elongation. As shown in Figure 3, the rate
of change in modulus with respect to
weight change is greater than for breaking
factor.
Elongation at break appears to be the
least sensitive, but the most decisive,
measurement. The variability in measure-
ment is large, roughly +_ 20 percent. Until
the sample is very stiff, and nearly
brittle, elongation does not always
correlate with other indicators. Figure 4
shows retention of ultimate elongation as
a function of weight change. Elongation is
not a good indicator of material change
for small weight losses or for weight
gains for this PML, Oily when the weight
loss is around ten percent does the
elongation show substantial decreases.
Types of Effects
The results in Table 1 for this
plasticized thermoplastic FML are grouped
according to the type of effects seen. A
temperature or concentration effect is
105
-------
noted in the third and fourth columns.
Five basic types of response to chemical
immersion were observed with this liner:
minor change, swelling, swelling and
softening with loss o£ strength, shrinking
and stiffening with loss of elongation,
and a combination of swelling and
shrinking depending on immersion
conditions.
TOmperature Effects. The response of
this liner to increased immersion temper-
ature shows that higher temperatures (at
least up to 50° C) can be used to
accelerate the material response.
When the response was minor or the
response time very fast (inorganic salts,
solvents methyl ethyl ketone and
1,2-dichloroethane), the difference in
results between the two temperatures was
small. In these cases, a higher exposure
temperature did not significantly
accelerate the response, but neither did
it change the response. This would
indicate that the higher temperature
itself did not affect the material.
•Temperature affected this FML's
response to two conditions: immersion in
both hydrochloric acid and in those
chemicals which caused a weight loss. The
liner response to hydrochloric acid at 50°
C shows a much larger weight gain compared
to the sample exposed at 23° C. But, at
both temperatures, the weight gain was
slow and steady, and had not stabilized
after two years.
Both the rate and the magnitude of
weight loss are affected by immersion
temperature for the stiffening response.
The initial rate of sample weight loss in
NaOH was faster at 50° C, and the final
weight loss was greater. The rate of loss,
though, became similar at both tempera-
tures after about 300 days of immersion.
Figure 5 illustrates the effect of
temperature on weight change. This figure
shows weight change as a function of time
for both hydrochloric acid and sodium
hydroxide imnersion samples, at both 23°
and 50° C. While the acid solution caused
a weight gain and the caustic solution
caused a weight loss, increasing the
temperature did not change the nature of
the response in either case. Both
responses were accelerated, however. This
indicates that an elevated temperature (at
least up to 50° C) can be useful in
accelerating chemical resistance testing
for this FML.
Concentration Effects. Four organic
chemicals were tested in three concentra-
tions. All responses seen with these
chemicals were affected by concentration.
Ml four saturated organic solutions
caused significant change in the liner
properties, in some cases almost total
loss of strength. But the response at
lower concentration is much less severe,
and appears to stabilize over time. Table
1 shows that the lowest concentration
tested of furfural (23° C only),
1,2-dichloroethane, and phenol (23° C
only) are considered resistant using the
above criteria.
Figure 6 shows the effect of furfural
concentration on weight change, breaking
strength, and S-100 modulus. The response
to furfural (swelling and softening)
stabilized with time at all concentra-
tions. Stabilized properties (log scale)
are shown as a function of concentration
for the 23° C exposure samples. The
response (absorption of furfural) appears
to be first order with respect to concen-
tration, and could be represented as:
bx
y = ae
• where y = liner response (percent
retention of modulus, for
example)
x = furfural concentration
(wt/wt percent)
b = liner sensitivity factor
The intercept of the line would be
log(a), and the slope b*log(e). This type
of graph could be useful in determining
acceptable chemical concentrations in a
waste stream, based on allowable stabi-
lized property changes. Note that it would
be of use only if the FML material stabi-
lizes in the waste solution, and if the
response is first order, and if the
chemical concentration is not expected to
increase over the life of the application.
Minor change. In water, salt, and
potassium dichromate solutions only minor
changes in weight and mechanical proper-
ties were observed . An example is shown
in Figure 7. This figure shows the
response of this liner to a saturated salt
106
-------
solution at 23°. Very little change is
seen in weight or mechanical properties,
even after two years.
Swelling. Hydrochloric acid immer-
sion was the only exposure condition for
this FML in which an increase in weight
does not correspond to a decrease in
strength. The liner had a long, slow
weight and volume gain at 50° C that was
an order of magnitude greater than the
gain at 23° C (34% compared to 4% after
two years). There was no corresponding
change in mechanical properties, however,
Figure 8. shows the response of the liner
to hydrochloric acid at 50° C.
Swelling and softening. In furfural
and 1,2-dichloroethane, this plasticized
thermoplastic liner swelled, softened, and
decreased in strength, indicating absorp-
tion of liquid. Figure 9 shows an example
of the liner response to 8% furfural.
In furfural, the liner's weight gain
was slower and continued over a much
longer time period than with 1,2-dichloro-
ethane, but in both exposures the breaking
strength response was fast and stablized.
Elongation shows a long-term decrease in
furfural that would not be predicted by
looking only at the short-term data.
Shrinking and stiffening. Shrinking
and stiffening, with gain in modulus,
breaking strength, and loss of elongation,
indicate loss of plasticizer. This
response occured with exposure to ASTM £2
oil and 10 percent sodium hydroxide. In
both chemicals, the liner became very
stiff as immersion time increased, and
sometimes was brittle, breaking as scon as
a test load was applied.
Figure 10 presents the liner response
to a 10 percent solution of sodium hydrox-
ide at 23° C. The response was similar but
more severe at 50° C. At 50° C, the liner
sample became brittle after a two-year
exposure and broke as soon as a load was
applied. The S-100 modulus, although not
shown on these figures, also increased
with exposure to oil and sodium hydroxide.
Combination Response. The fifth type
of response seen was a combination of
swelling and shrinking responses. Immer-
sion in methyl ethyl ketone and phenol
resulted in both swelling/softening and
shrinking/stiffening depending on time,
temperaure, and concentration.
Exposure to these chemicals caused
both absorption of the chemical or water
into the liner and extraction of the
plasticizer out of the liner. In general,
the absorption process is faster, result-
ing in an initial weight gain that is
dependent on concentration. As exposure
time and/or chemical concentration
increases, the slower extraction process
becomes significant.
Figure 11 shows percent weight gain
as a function of immersion time for
samples exposed at 23° C. Weight gain is
shown for samples exposed to 1, 4, and 8%
phenol solutions. After a one day
exposure, all samples have gained weight,
and the weight gain increased with
increasing phenol strength. This initial
relationship of weight to concentration
becomes reversed as the immersion time is
lengthened. The lowest phenol concentra-
tion (1%) has the largest weight gain, and
the highest concentration (8%) has the
most weight loss, after two months
immersion.
The liner's mechanical properties
also showed this combination response.
Figures 12 a, b, and c show percent weight
change, percent retention of breaking
factor, and percent retention of ultimate
elongation with immersion time at 50° C.
In Figure 12a, response to a 1% phenol
solution, a stabilized decrease in break-
ing strength corresponds to the moderate
weight gain. Elongation is not signifi-
cantly affected. In a 4% solution (Figure
12b), the initial weight gain recedes with
time. The corresponding initial loss of
breaking strength (from softening due to
absorption) is recovered and breaking
strength increases with time. Elongation
still shows no significant change.
The liner response to an 8% solution
(Figure 12c) shows the continuation of
these trends. The initial weight gain (and
initial loss of strength) becomes a weight
loss (and increase in strength). In 8%
phenol, however, an additional effect is
apparent: the elongation at break has
decreased after a two-year exposure,
another sign of stiffening and possible
eventual embrittlement.
107
-------
OCNCDOSICNS
REFERENCES
1. Immersion testing of a liner in
the waste it is intended to con-
tain is essential for determining
chemical resistance. Low con-
centrations of an incompatible
chemical can cause significant
change in FML physical properties.
2. 'She stabilization of a material's
response to a chemical challenge,
when considered in conjunction
with the magnitude of that
response, is an important param-
eter in the evaluation of chemical
resistance.
3. For the FML discussed in this
paper, increasing the immersion
temperature does appear to
accelerate the response without
changing it.
4. For the FML discussed in this
paper, water was not an aggressive
medium.
5. Increasing the concentration of
organic solvents in water solution
increased the magnitude of the ML
response (physical changes).
6. The indicators of FML
non-resistance to a chemical
depend on the type of FML tested.
For the thermoplastic discussed in
this paper, the S-100 modulus,
weight change, and breaking
strength were the most important
indicators of non-resistance.
7. Minimum as-received property
values listed in NSF Standard 54
for Flexible Membrane Liners can
be useful as benchmarks in
evaluating chemical resistance
test results.
American Society for Testing and
Materials, Standard Test Method
for Resistance of Plastics to
Chemical Reagents, AS1M D543-67,
Reapproved 1978.
Barton, Allen P., 1983. CRC
Handbook of Solubility Parameters
and Other Cohesion Parameters. CRC
Press,Inc., Boca Raton, Florida.
National Sanitation Foundation,
1983. Standard 54 for Flexible
Membrane Liners. NSF Ann Arbor,
Michigan.
Personal communication, Mary Ann
Curran, October, 1985.
Rossiter, Walter J., 1983. A
Methodology for Developing Tests
to Aid Service Life Prediction
of Single-Ply Roofing Membranes,
presented at the NBS/ftBCA 7th
Conference on RoofingTechnology.
Tratnyek, Joseph, Peter Costas,
and Warren Lyman, 1984. Test
Methods for Determining the
Chemical Waste Compatability of
Synthetic Liners. Final Report to
U.S. Environmental Protection
Agency, Contract No. 68-01-6160.
U.S. Code of Federal Regulations
Part 40, Sections 264.221 and
264.301.
U.S. Environmental Protection
Agency, Method 9090. Cotnpatability
Test for Wastes and Membrane
Liners, October, 1984.
U.S. Environmental Protection
Agency, Draft Minimum Technology
Guidance on Double Liner Systems
for Landfills and Surface
Impoundments — Design,
Construction, and Operation, May,
1985.
108
-------
Table 1.
PLASTICIZEO THERMOPLASTIC: RESPONSE SUMMARY FOR CHEMICAL RESISTANCE
CHEMICAL TEMPER- CONCEN- TYPE OF
.NAME TYPE ATURE TRATION RESPONSE
EFFECT EFFECT
Water - no - Minor Change
Sodium Chloride SALT no no
Potassium Dichromate OXIDIZER no
Hydrochloric Acid ACID yes
Furfural ALDEHYDE no yes
1 ,2-Dichloroethane CHLORINATED no yes
HYDROCARBON
Minor Change
Minor Change
Swel ling at 50 C
Swe 1 1 ing and
Softening
Swel 1 i ng and
Soft ening
Sodium Hydroxide BASE yes - Shrinking .and
1 St i f f eni ng
ASTM »2 Oil OIL yes - | Shrinking and
St i f fen i ng
Methyl Ethyl Ketone KETONE no yes Swelling and
Sof t eni ng with
some later
Shrinking and
St if f ening
Phenol PHENOL no yes Swelling and
Sof ten i ng at low
concentration and
shrinking and
stiffening at
nigh concentration
STABILITY RESISTANT- REASONS FOR
ACHIEVED- NON-RESISTANCE *
CONDITIONS
23
50
23
50
23
50
23
50
23
50
23
50
23
50
23
50
23
50
23
50
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
-
10% &
sat 'd
10%
10%
1%
4%
8%
1%
4%
8%
, 1%
.5%
.8%
. 1%
.5%
.8%
10%
100%
3%
13*
26%
3%
13%
26%
1%
4%
8%
1%
4%
8%
Y
Y
Y
Y
Y
Y
N
N
Y
Y
Y
Y
Y
Y
Y(v)
V(v)
V(v)
Y(v)
Y(v)
V(v)
Y
N
Y
Y
Y(v)
Y
V(v)
V{v)
Y
Y(v)
Y
Y
N
Y
N
N
Y
Y
Y
Y
Y
Y
N
N
Y
N
N.
N
N
N
Y
N
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
S
S, WG
WG, BL, SL
WG, BL, SL
WG (margina
wG, BL, SL
WG, BL, SL
WG, BL,
WG, BL,
WG, BL,
WG, BL,
WL,
WL,
WL,
WL,
BL,
WG, BL.
WG, BL,
WG, BL,
WG, BL,
WG, BL,
WG(rnarg
S
S, WL,
WG
s
S, WL,
SL
SL
SL
SL
SG,
SG,
SG
SG,
SL
SL
SL,
SL
SL
SL
ina!
SG,
SG,
)
EL
EL
EL
EL
EL
EL
* S = No Stability, WG = weight gain, WL = weight loss, BL = Breaking Factor Loss, EL = Elongation at Break Loss,
SG = S-100 Gain, SL = S-100 Loss
Y = Yes, N = No, (v) = variable results due to volatility of chemical
-------
p
E
R
C
E
N
T
ou-
50-
40-
30-
20"
10-
J
• • Wl4fN ttxno (H) .IH DCC /
«. wrifN t)>ii<«> (») fer* MCI y
/*
/*
s
s-m •
• — -""* *
ji — « — "" *•
i 0
2 5 10 20 50 100 200
LOG DAYS IMMERSION
500 1000
Figure 1. Comparison of Wsight Changes of a Thermoplastic FML with
Time In Two Chemical Solutions at 23 Centigrade..
-30 -20 -16 8 16
PERCENT WEIGHT CHANGE ALL EXPOSURES
Figure 2, Relationship of Breaking Factor to Percent Weight Change
for a Plasticizad Thermoplastic FML.
iet
-28 -15 -18 -5 8 6
PERCENT WEIGHT CHANGE IN ASTM #2 OIL AND 10% SODIUM HYDROXIDE
R
E
T
E
N
T
I
0
N
* tf***'.
+
i
Elongation at Break
-36 -26 -te 0 te ae
PERCENT WEIGHT CHANGE ALL EXPOSURES
Figure 3. Relationship of Breaking Factor and S-100 Modulus to Weight Loss
for a Plasticized Thermoplastic FML.
Figure 4. Relationship of Elongation, at Break to Weight Changes ior a
Plasticized Thermoplastic FML.
-------
30
P
E »
R *>
C
E °
N ..
T
•Vf
• O—'
S0 C
\
10-2x
10 % Hydrochloric Acid
23 C
fl--o a- _ -_a i
10 % Sodium Hydroxide
1 i i i I
50 C
100200
300 400 600
DAYS IMMERSION
600
700800
Figure 5. Comparison of the Effect of Temperature on Weight
Changes of a Plasticized Thermoplastic FML.
CONCENTRATION (wt/wt percent)
Figure 6. Relationship of Change in Physical Property to Furfural
Concentration at 23 Centigrade for a Plasticized Thermoplastic FML.
P 150-
E ,
R 1001
C
50 -
N
T 0 -
-50-
V
. , , fl 0
~ -x "sr
i tf D
•
a
r 9 _g)
1 . , ,1
_|_.JL
U Q
— — ft
= Weight Change (%)
= Breaking Factor (% retention)
= Elongation at Break (% retention)
• m
W-J ! UO-U4.
P
E 100^
R '
C
E
N 50-
T
A 4
V o i-i n n
ij n
n ^
"
. •= Weight Change (%)
a = Breaking Factor (% retention)
n- Elongation at Break (% retention) •-•'*
-.•*"'*"
__ _,-*-— •—-•-'".'* ,
10 20 50 100
LOG DAYS IMMERSION
200
500 1000
10 20 50 100
LOG DAYS IMMERSION
200
500 1000
Figure 7. Physical Property Change with Immersion in Saturated Salt
Solution at 23 Centigrade for a Plasticized Thermoplastic FML.
Figure 8. Physical Property Change with Immersion in 10% Hydrochloric Acid
Solution at 50 Centigrade for a Plasticized Thermoplastic FML.
-------
ISO-]
p \
E 100-
R C
C
5 S
\7---^^_
- •* Weight Change (%) "~^*V«
1 Q= Breaking Factor (% retention) _— J?
y= Elongation at Break (% rotentlon) ^,—-~"~""~
E «-.--"""~"
N ^T r'"'* * " * •
T ,
0J
•"' i 1 - i • • i| f i • i i i i t| • i 1 t • •
2 5 10 20 50 100 200 500 1000
LOG DAYS IMMERSION
Figure
200-
P 150-
E
R 100-j
C
E 50 -
N
T 0 i
-50-
9, Physical Property Change with Immersion In 8% Furfural
Solution at 23 Centigrade for a Plasticlzad Thermoplastic FML.
' !_ JL TO — •^^^^'^
» W TT 'lP "^'
»= Weight Change (%) '-^
o= Breaking Factor (% retention) '"Xn
\J= Elongation at Break (% retention) «
L__._,-l...._._..t_L..i. t t t | 1 .-.___ .._l- . .J-l^l 1 1 It! II
2 5 10 20 50 100 200 500 1000
LOG DAYS IMMERSION
Figure 1
15
!
P 10
5
R
C 0'
E
N -5
T
-10-
-is
0. Physical Property Change with Immersion in 10% Sodium Hydroxide
Solution at 23 Centigrade for a Plasticizod Thermoplastic FML.
L e= Weight change (%) 1% Phenol
!\ n= Weight change (%) 4% Phenol
j— ^______^ V« Weight change (%) 8% Phenol
;.,,_^----^|^.^.
"•^-^
P \ t 1 r * ! ' 1 I r iritii
5 10 20
LOG DAYS IMMERSION
50
100
Figure 11. Weight Change with Immersion in Phenol Solution at 23 Centigrade
for a Plasticlzed Thermoplastic FML.
112
-------
p
E
R
C
E
N
T
150-
100-
1
(
50-
0 J
«_ „ t? _J» „
•^.-- 9""
a r Q Q
* = Weight Change (%)
D = Breaking Factor (% retention)
y = Elongation at Break (% retention)
n
^ , -•
5 10 20 50 100 200 500 1000
LOG DAYS IMMERSION
Figure 12». Physical Property Change with immersion in 1% Phenol Solution
at 50 Centigrade (or a Plasticized Thermoplastic FML.
150
P
E
R
C 50 +
E
N
T
o --
-50
• a Weight Change (%)
D = Breaking Factor (% retention)
\J a Elongation at Braid (% retention)
2 5 10 20 50 100 200
LOG DAYS IMMERSION
500 1000
Figure 12b. Physical Property Change with Immersion in 4% Phenol Solution
at SO Centigrade for a Plaaticizad Thermoplastic FML.
150
1001 :
E
R
C 50 -f-
E
N
T
-50
• - Weight Change (%)
a,a Breaking Factor (% retention)
^ = Elongation at Break (% retention)
•4-
2 5 10 20 50 100 200
LOG DAYS IMMERSION
500 1000
Figure !2c. Physical Property Change with Immersion in 8% Phenol Solution
at 50 Centigrade for a Plasticlzed Thermoplastic FML.
113
-------
VERIFICATION OF THE HYDROLOGIC EVALUATION OF
LANDFILL PERFORMANCE (HELP) MODEL USING FIELD DATA
Lee Peyton
Department of Civil Engineering
University of Missouri - Columbia
Columbia, MO 65211
Paul R. Schroeder
Environmental Laboratory
USAE Waterways Experiment Station
Vicksburg, MS 39180
ABSTRACT
Simulations of twenty landfill cells from seven sites were performed using the HELP
model. The results of the simulations were compared with the field data obtained from the
cells to verify the HELP model and to Identify inaccuracies and shortcomings of the model.
Field data were complied from eight large field lysimeters at the University of
Wisconsin in Madison, from five landfill cells monitored in a gas generation demonstration
project at Sonoma County, California, from a demonstration landfill at Boone County,
Kentucky, from three county landfills in Wisconsin and from three cells of a hazardous
waste disposal facility in Niagara Falls, New York. Results for the Madison, Wisconsin,
the Sonoma County, California and the Boone County, Kentucky sites are presented in this
paper. Measured components of the water balance ranged from just leachate collection to
percolation, lateral drainage and runoff. The descriptions and soil properties of the
sites were loosely defined, requiring much judgment In the selection of HELP model Input
specifications and allowing significant variance in the simulation results.
Using best judgment prior to calibration, runoff was overpredlcted for five cells by
an average of SOX and underpredicted for six cells by an average of 20%. Evapotranspira-
tion appeared to be underpredlcted by an average of 8%, but low values of evaporative
depth were used in the simulations which may account for this. Lateral drainage was
overpredicted by 11% In two cells where very high leachate collection rates were observed.
In three cells where very small quantities of leachate were collected, lateral drainage
was underestimated by 97%, although this difference only amounted to 1.4 Inches per year.
Of the remaining nine cells, lateral drainage was overpredlcted by an average of 7% in
covered cells and overpredlcted by an average of 52% In permanently uncovered cells with
a weathered waste surface that supported dense vegetation.
INTRODUCTION
Among the potential problems posed by
landfills Is the possible contamination of
ground and surface waters by the migration
of leachate offslte. Therefore, It is
essential that adequate design tools are
available to accurately predict and con-
trol this liquid generation and movement*
The Hydrologlc Evaluation of Landfill
Performance (HELP) model (1,2) was devel-
oped to help hazardous waste landfill de-
signers and evaluators estimate the magni-
tude of water budget components and the
height of water-saturated soil above bar-
rier soil layers. This model performs a
sequential dally analysis to determine
runoff, evapotranspiration, percolation
and lateral drainage Cor the landfill and
obtain dally, monthly and annual water
budgets.
114
-------
Determining the water budget is not a
simple task. The interrelationships be-
tween climate, vegetation and soil charac-
teristics, and their effects on runoff,
evapotranspiration and vertical drainage
are complex. It is necessary to verify
that such a model accurately represents
reality. This report presents the results
of an attempt to verify the HELP model
using existing data collected at land-
fills, test cells and lyslmeters.
PURPOSE AND APPROACH
The purpose of the study was to as-
sess the adequacy of the HELP model in
simulating liquid movement through land-
fills and to validate the use of the model
In evaluating landfill designs. These
assessments were based on field data ob-
tained at various sites in California,
Wisconsin, Kentucky and New York. The
landfills varied In their designs, oper-
ating conditions and soil characteristics.
In general, the landfill characteristics
were loosely defined and introduced uncer-
tainty in the selection of values to de-
scribe the landfill. The water budget
field data varied in the number of compo-
nents measured—only leachate collection
was measured in some cases, while precipi-
tation, runoff, percolation and leachate
collection were measured In other cases.
Consequently, the level of verification
differed from site to site, but efforts
were made in all cases to determine the
accuracy of the model in simulating Iso-
lated components of the water budget.
RESULTS
University of Wisconsin Lysimeters
From 1970 to 1977, eight large lyslm-
eters filled with either shredded or un-
processed refuse were monitored for sur-
face runoff and leachate production at the
University of Wisconsin-Madison (3). Each
cell was 60 feet long by 30 feet wide.
The depth of refuse was either 4 or 8
feet. Refuse was underlain with a 4-lnch
layer of crushed granite over a 6-tnll
polyethylene barrier. Bottom slopes of
approximately 3% directed leachate to a
collection box at the center of the cell
where it was periodically pumped and the
volume of leachate was measured. Four
cells were covered with a 6-inch thickness
of sandy silt soil; the remaining cells
were left uncovered. The top surfaces of
all cells were sloped at 3% toward one of
the 60-foot walls where surface runoff was
collected and measured.
Cllmatological conditions at this lo-
cation are as follows! Average tempera-
ture Is 45 F; average annual precipitation
is 31 Inches; minimum daily temperature
falls below freezing on 163 days per year;
and average daily solar radiation is 330
langleys.
The simulations used actual daily
precipitation and mean monthly tempera-
tures from National Oceanic and Atmospher-
ic Administration (NOAA) records for
Madison, Wisconsin. Default values for
solar radiation at Madison were used.
Since hydraulic parameters for the soil
and waste were not available, default
characteristics were used. Default soil
texture 9 was chosen for the sandy silt
soil cover, and default soil texture 19
was chosen for the waste layer.
The vegetative cover was mixed volun-
teer vegetation, comparable to meadow
grass, which became established on both
covered and uncovered cells over a several
year period. This vegetation grew more
quickly and more densely on the uncovered
cellsj therefore, the default option for
"fair grass" was chosen to describe vege-
tation on the uncovered cells and the
default option for "poor grass" was chosen
for the covered cells. Evaporative depths
were set at 12 and 8 Inches, respectively,
to be consistent with suggested values in
the HELP model documentation. SCS runoff
curve numbers were selected by analyzing
observed precipitation and runoff data.
The first two years were treated as
periods of equilibration for both the HELP
simulations and the test cells and there-
fore are not included in the following
comparisons. It was assumed that this
period was necessary to bring Internal
moisture to normal levels and to establish
a weathered surface on the uncovered cells
and volunteer vegetation on all of the
cells.
- The simulation results are presented
In the form of cumulative comparisons
between model predictions and field mea-
surements for the following components of
the water balances runoff; evapotrans-
piration plus change In moisture storage
(ET+AS); and lateral leachate drainage.
115
-------
Field "measurements" of ET+AS were actu-
ally computed from
1T+AS - PROP - ROT - DRG
where PROP * measured precipitation, ROT =
measured runoff, and DRG " measured later-
al leaehate drainage. It was assumed that
all leachate was recovered by lateral
drainage due to the presence of an Imper-
vious synthetic liner below the lateral
drainage layer.
Comparisons of runoff, El+AS, and
leachate drainage are plotted in Figure 1
for one of the covered cells. The simula-
tions overpredlcted cumulative runoff for
two covered cells and underpredicted cumu-
lative runoff for the remaining two cov-
ered cells. However, the runoff simula-
tions generally compare well prior to a
large storm In July 1975» The comparisons
for cumulative ET+AS and cumulative leach-
ate drainage were generally better than
for runoff. On the average, runoff
accounted for 7.6i2.2 percent of the pre-
cipitation from the covered cells. The
HELP model predicted 8.1 percent, yielding
an error of 0.5±2.6 percent of the preci-
pitation or a relative error of 6.8 per-
cent of the measured runoff. Leachate
drainage accounted for 23.5±4.5 percent of
the precipitation while the model esti-
mated 25.1 percent. The average error was
1.6*3.7 percent of the precipitation and
the relative error was 7.0 percent of the
measured leachate. The "measured" ET+AS
was 68.9±3.1 percent of the precipitation
while the predicted value was 67.1 per-
cent. The model underpredicted by 1.8±3.3
percent of the precipitation or 2.7 per-
cent of the "measured" ET+AS.
Comparisons for an uncovered cell are
plotted in Figure 2. Simulation of these
cells was not as successful, particularly
for leachate drainage. On the average,
runoff accounted for 3.4±0.5 percent of
the precipitation for uncovered cells
while the HELP model estimated 2.6 per-
cent. The model underestimated by 0.8±0.5
percent of the precipitation, yielding a
relative error of -22 percent of the mea-
sured runoff. Leachate drainage accounted
for 12.749.4 percent of the precipitation
which shows the highly variable nature of
uncovered waste. The model predicted: 19.3
percent, producing an error of 6.6±10.2
percent of the precipitation and a
relative error of 51.9 percent of the
measured leachate. While the error In
drainage is large, it Is still smaller
than the variability In the measured
values. The "measured" ET+AS was 83.9±9.1
percent o£ the precipitation while the
predicted value was 78.0 percent. The
model underpredicted by 5.9±10.1 percent
of the precipitation or 7.0 percent of the
"measured" ET+AS. The differences between
the predicted values and the actual mea-
surements for both covered and uncovered
cells were generally equal to about one
half of the standard deviation of the
measured values which indicates fairly
good agreement between the predictions and
measurements.
The difference between the actual
measurements and the model predictions
could be Influenced by a number of fac-
tors. Errors were present in the field
measurements of runoff and leachate drain-
age due to short periods of pump and run-
off system malfunction in the Springs of
1973, 1974, and 1975. Errors may also
have been introduced by using precipita-
tion measurements collected several miles
from the test site. This could explain
the large discrepancy during July 1975 as
seen In the runoff comparisons. Based on
a comparison with nearby weather stations,
the major storm responsible for this run-
off was highly localized so that the mea-
sured rainfall may not accurately repre-
sent conditions at the test site. Other
uncertainties were introduced In selecting
parameter values to describe extent of
vegetative growth, leaf area Index, winter
cover factor, evaporative depth, soil and
waste characteristics, and degree of com-
paction.
Model predictions showed no runoff
during the winter months. This Is consis-
tent with the HELP model methodology which
stores all precipitation on the surface as
snow when average temperatures are below
freezing. Thus, when Wisconsin tempera-
tures warmed In April, all precipitation
stored at the surface by the model was
allowed to either run off or Infiltrate.
The measured data suggest that this metho-
dology Is not appropriate. Significant
runoff did occur in the field throughout
the extended periods of daily average
below-freezing temperatures and therefore
the field runoff in April was signifi-
cantly less than predicted.
116
-------
10.0-1
5.0 -1
0
100
z
co- 80
5 40
1 20
o
0
g 40
UJ
I 3°
1
Q 20-
LEGEND
' HELP SIMULATION
• FIELD
JAN
1973
JAN
1974
JAN
1975
JAN
1976
DATE
Figure 1. Cumulative Comparison of HELP Simulation
and Field Measurements for University of
Wisconsin-Madison Covered Cell.
5
4
3
1-
0
_125
g-100
*
C 75-j
LU
|n|
I 25J
LEGEND
' HELP SIMULATION
• FIELD MEASUREMENT
0
z 40
uJ
1 30-
r-^
20-
UJ
1
10-i
JAN
1973
JAN JAN
1974 1975
DATE
JAN
1976
JAN
1977
Ftgure 2. Cumulative Comparison of HELP Simulation
and Field Measurements for University of
Wisconsin-Madison Uncovered Cell.
-------
Sonoma County lest Cells
A solid waste stabilization project
was conducted la Sonoma County, California
from 1971 to 1974 to determine the effect
of applying excess water, septic tank
pumpings, and recycled leachate to land-
fills (4). Each of the five cells studied
was 50 feet long by 50 feet wide and 8
feet deep. Three cells were constructed
and operated as typical sanitary landfills
except that the moisture content of the
refuse was brought to field capacity with
water prior to capping in one cell and
with septic tank pumpings in another cell.
The fourth and fifth cells were con-
structed with an inflow pipe network and a
sand/gravel distribution medium installed
between the waste layer and the soil
cover. Water was continuously added to
the waste through this network In the
fourth cell, whereas landfill leachate was
continuously recycled through the waste
layer in the fifth cell.
This test site was located 45 miles
north of San Francisco. The mean tempera-
ture is 58 V, with the dally minimum tem-
perature falling below freezing on 39 days
per year. Mean annual precipitation is 31
inches. Ninety—five percent of this prer
cipitation occurs from October to April.
The mean dally solar radiation is approxi-
mately 410 langleys.
Dally precipitation measurements made
at the test site were used in the simula-
tions. Mean monthly temperatures were
taken from NOAA records for Santa Rosa,
California, 10 miles from the test site.
Default values for solar radiation at
Sacramento, California were used. Default
soil texture 14 was chosen for both the
sandy clay soil cover and the sandy clay
liner. Both were treated as compacted
soils In the model. The hydraulic conduc-
tivity of the liner was set to 0.000277
In/hr (2 x 10"7cm/sec) based on field
tests. The thickness of the liner was set
to 24 Inches since this was the depth of
clay liner replacement where pervious
lenses were encountered during excavation.
The cover thickness was also 24 inches.
Surface vegetation was assumed to be
absent since its presence was not ad-
dressed in the project report (4) and
since summer precipitation Is inadequate
to support significant vegetation. The
evaporative depth was chosen to be four
inches as suggested for bare ground by the
HELP model documentation report (1). SCS
runoff curve numbers were estimated based
on rainfall and runoff data measured at
the test cells.
Results for one of the cells con-
structed and operated as a normal sanitary
landfill (except for initial moisture con-
ditions) are shown In Figure 3. The first
year of cell operation and model simula-
tion was considered to be a period of
equilibration and is not plotted. Com-
parisons are made between model predic-
tions and field measurements for the fol-
lowing components of the water balance:
runoff; evapotranspiratlon plus change in
moisture storage plus percolation through
the clay liner (ET+&S+PERC)5 and lateral
leachate drainage. Field "measurements"
of ET+&S+PERC were actually computed from
ET+AS+PERC - PROP - RNF - DRG
where PERC, RNF, and DRG represent mea-
sured values for precipitation, runoff,
and lateral drainage. The model predic-
tion for percolation through the clay
liner Is also Included as diamonds on the
drainage plot.
The comparison in Figure 3 is gener-
ally typical of all three cells which did
not contain a pipe network inflow system
within the cover. On the average for the
three cells, runoff accounted for 61.0+2.7
percent of the precipitation. The HELP
model predicted 71.9 percent, yielding an
error of 10.9±2.7 percent of the
precipitation or a relative error of 17.9
percent of the measured runoff. Leachate
drainage accounted for 2.94±1.56 percent
of the precipitation while the model
estimated 0.05 percent. The average error
was -2.89±1.56 percent of the precipi-
tation and the relative error was -98.3
percent of the measured leachate. The
"measured" ET+AS was 36.0+3.4 percent
of the precipitation while the predicted
value was 27.9 percent. The model under-
predicted by 8.1+3.3 percent of the pre-
cipitation or 22.4 percent of the "mea-
sured" ET+AS.
The obvious major discrepancy is re-
lated to lateral leachate drainage. The
model predicted that practically all
leachate would leave through barrier soil
percolation rather than lateral drainage.
This could be due to the form of the
118
-------
percolation and lateral drainage equations
used in the model. As the depth of leach-
ate ponding above the barrier soil ap-
proaches zero, the lateral drainage pre-
diction approaches zero while predicted
percolation continues at a rate equal to
the saturated hydraulic conductivity times
a unit hydraulic gradient. The very snail
rates of infiltration through the clay
cover produced such small ponding depths
that this percolation/drainage partition-
ing scheme predicted that the infiltration
would leave the landfill as percolation
rather than lateral drainage. This scheme
is not necessarily in error but the dif-
ference in leachate drainage between pre-
dicted and measured probably is the result
of the overprediction of runoff. The
underestimated infiltration would have
significantly increased the leachate
drainage and ET+AS+PERC which were both
underpredlcted. Runoff may have been
overpredlcted due to the presence of vege-
tation that was not mentioned in the
project report (4). With vegetation the
runoff curve number would have been much
smaller and evaporative depth would have
been greater, increasing infiltration and
evapo transpira tIon.
Results for one of the two cells with
the pipe network inflow system are shown
in Figure 4. The entire period of record
is shown on the plot. Since surface run-
off was not measured from these cells, the
only comparison shown in Figure 4 is for
lateral leachate drainage, although the
model predictions for runoff, ET+AS+PERC,
and percolation, as diamonds on the drain-
age plot, are also plotted. The curves
for the HELP simulation leachate drainage
results and the field drainage data lie on
top of each other with a relative error of
less than 0.7 percent.
Based on measurements from the cells
without the inflow system, rainfall ac-
counted for about one percent of the
leachate drainage in the two cells with
the inflow. This provides a unique test
of the HELP model methodology for dividing
flow between lateral drainage and vertical
percolation since inflows and lateral
drainage outflows are known. The excel-
lent reproduction shown in Figure 4 ap-
pears to confirm the appropriateness of
the HELP model methodology for large in-
flow rates.
Boone County Test Cell
A test cell was studied from 1971 to
1980 in Boone County, Kentucky to evaluate
volume and characteristics of the leach-
ate, composition of gases, internal tem-
perature, settlement, and clay liner effi-
ciency (5,6). fhe cell consisted of a
30-foot wide by 150-foot long trench with
vertical walls and ramps on both ends
sloping toward the center at 14 percent.
The middle 50 feet were sloped at 7 per-
cent to the transverse centerline. A 30-
mil synthetic liner, 30 feet wide by 50
feet long, was centered over the base of
the cell. A leachate collection pipe
embedded in gravel was placed on the syn-
thetic liner at the bottom of the cell.
An 18-inch thick compacted clay liner was
placed over the synthetic liner and col-
lection pipe. This liner was found to
have an average in-place hydraulic conduc-
tivity of 4.0 x 10"?cm/sec at the conclu-
sion of the cell study. A second pipe was
embedded in a gravel-filled section of the
clay liner directly above the lower pipe
to collect lateral leachate drainage above
the clay liner. A 6-mil polyethylene
strip was placed beneath this pipe to
prevent leachate from shortcircuitlng to
the lower pipe. Residential refuse was
placed and compacted above this liner
system, and a 2-foot layer of CL (Unified
Soil Classification System) cover soil was
deposited onto the completed waste layer.
The cover soil was found to have an aver-
age J.n-place hydraulic conductivity of 5.0
x 10 cm/sec at the conclusion of the cell
study.
The test site is located approximate-
ly 20 miles south of Cincinnati, Ohio.
The mean annual temperature is 54 F, with
the daily minimum temperature falling
below freezing on 111 days per year. The
mean annual precipitation is 43 inches.
The mean dally solar radiation is approxi-
mately 360 langleys.
Precipitation at the test site was
recorded once or more per week throughout
the study period. These values were used
to adjust daily precipitation records from
the nearest NOAA weather station, located
approximately 15 miles away at Covington,
Kentucky. Mean monthly temperatures were
taken from the Covington weather station.
Solar radiation values were the HELP model
default values for Cincinnati, Ohio.
119
-------
.60-
Jso-
f 40-i
cc ;
LU 30 ":
> ;
3 20~
i«.
o
OH
z 25
c5
I 20
-*
CO
? 15
S3
> 10
I 5
1 0
z 4
ul
o 2
s
o
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
DATE
Figure 3. Cumulative Comparison of HELP Simulation
and Field Measurements for Sonoma County
Cell without Pipe Inflow.
S 60-i
| 50-
1 40'
U4
S 30
20
3
E
3 10
0:
HELP SIMULATION
FIELD MEASUREMENT
540^
JAN
1974
Figure 4. Cumulative Comparison of HELP Simulation
and Field Measurements for Sonoma County
Cell with Pipe Inflow.
-------
Default soli texture 14 was chosen to
describe the two-foot top soil layer. This
texture matches the CL classification for
this top soil and has a hydraulic eoadue
tivity close to the average value which
was measured for this layer. In order to
model the circuitous path of clay liner
percolation, an equivalent liner thickness
of 12 feet was chosen. The area of the
gravel packing/clay liner Interface sur-
rounding the lower pipe was treated as a
leakage opening through a synthetic mem-
brane. The default soil texture for the
clay liner (texture 20) was chosen to
match the measured In-place hydraulic
conductivity.
The surface vegetation was assumed to
be fair grass based on reported observa-
tions at the end of the test cell study.
An evaporative depth of 7.5 Inches was
selected to correspond to this observa-
tion. A default runoff curve number of 86
was selected by the HELP model according
to soil and vegetation characteristics.
Measurable leachate was first produced
In September 1971, three months after
construction was completed. Cumulative
plots beginning in September 1971 of mea-
sured and predicted values are shown in
Figure 5. This figure shows very little
measured leachate until the end of 1972.
After that time, leachate drainage volumes
somewhat exceed those predicted by the
HELP model. Percolation volumes are neg-
ligible from both the field measurement
and the model prediction. Over the seven
year period, leachate drainage accounted
for 28.8 percent of the precipitation,
while the HELP model predicted 24.6 per-
cent. The model underpredlcted the drain-
age by 4.2 percent of the precipitation or
14.6 percent of the measured drainage.
The field assessment of the test cell
at the end of the Boone County study indi-
cated that secondary openings existed In
the soil cover through which relatively
rapid infiltration could occur. There-
fore, a second HELP model simulation was
conducted using the highest of the three
in-place hydraulic conductivity measure-
ments o£ the soil cover. This resulted In
a predicted leachate drainage of 26.2
percent of the precipitation, yielding an
error of -2.6 percent or —9.2 percent of
the measured drainage.
EVALUATION OF HELP VERIFICATION DATA
Measured runoff data existed for 8
cells at the University of Wlsconsin-
Madison and for 3 cells at Sonoma County.
In all cases an attempt was made to cali-
brate the runoff curve number In advance
of the stimulation by examining measured
rainfall-runoff data. Runoff was overpre-
dlcted for 5 cells by an average of 30
percent of the measured runoff and under-
predicted for 6 cells by an average of 20
percent of the measured runoff.
No suitable evapotransplratlon field
data from landfill sites was found for
model testing. This was not unexpected
due to the complexities Involved in col-
lecting this type of data. For those
cells which had runoff data available, a
surrogate variable for evapotransplratlon
was Identified, and comparisons were made
between measured and predicted results.
The variable consisted of the sun of the
water balance components which were not
directly measured. In the case of the
University of Wisconsin-Madison cells, the
variable was the sum of evapotransplratlon
and change in moisture storage, ET+AS.
For the Sonoma County cells, it was the
sum of evapotransplratlon, change In mois-
ture storage, and percolation, ET+AS4-PERC.
The ET+AS variable was found to be under-
predicted by an average of 5 percent of
the measured values, whereas the
ET+AS+PERC variable was underpredicted by
an average of 22 percent. It is obviously
rather complex to discern the meaning of
these results for evapotransplration since
It Is Interrelated to change in moisture
and percolation. However, the evidence
suggests that values chosen for evapora-
tive depths may have been too small.
Since measurements of barrier soil
percolation volumes and leachate ponding
depths were generally not available, the
lateral drainage and barrier soil percola-
tion submodel could only be evaluated
using measured drainage data. Lateral
drainage was overpredlcted by 11 percent
of the measured drainage In two cells
where very high leachate collection rates
were observed. In three cells where very
small quantities of leachate were col-
lected, lateral drainage was underesti-
mated by 97 percent of the measured drain-
age, although this difference only amount-
ed to 1.4 Inches per year. Of the remain-
Ing nine cells, lateral drainage was over-
121
-------
40
30
z
3
CC
I 10
o
0
jjzso-j
? 200-i
*+ 1J
ui 1M j
3 SO-
s ^
uJ on J
I— 8U*i
40 H
0-1
LEGEND
HELP SIMULATION
FIELD MEASUREMENT
JAN JAN
1971 1972
JAN
1973
JAN JAN
1974 1975
DATE
JAN
1976
JAN
1977
JAN
1978
JAN
1979
predicted by an average of 7 percent of
the measured drainage In five covered
cells and overpredicted by an average of
52 percent of the measured drainage J.n
four permanently uncovered cells with a
weathered waste surface that supported
dense vegetation.
CONCLUSIONS
The field data verified the utility
of the HELP model for estimating general
landfill performance. However, not all
model components were well tested due to
the limited field data available. It Is
concluded that a laboratory and field
monitoring program explicitly designed for
HELP model verification would be necessary
for further refinement of specific model
components.
The overall data base of long-term
water budget field measurements at land-
fills Is poorly organized and too small to
continually advance the state—of-the—art
In understanding landfill leachate genera-
tion and migration. More extensive moni-
toring activities are required to fill
this technology gap.
122
-------
Although the HELP model is primarily
oriented to landfill design, the simula-
tions in this study show the usefulness of
the model in evaluating existing condi-
tions at operating landfills.
The model was shown to simulate parti-
cularly well the leachate drainage from
landfills with relatively large Infiltra-
tion rates. The model did not simulate
well the leachate drainage when small
infiltration rates were significantly
underpredicted. Runoff predictions were
found to be within about plus or minus 25
percent of measured data. Evapotranspira-
tion verification data was lacking, but
the results tended to suggest that the
model underpredicted evapotranspiration.
The error was probably caused by the use
of small evaporative depths and high un-
saturated hydraulic conductivities. Col-
lection of evapotranspiration data should
be emphasized in future studies due to its
significant impact on the overall water
balance. Errors in HELP model estimates
of the various water budget components
were generally smaller than the standard
deviation of the measured values for the
components.
Improvement to the HELP model should
be made in the areas of winter runoff,
unsaturated hydraulic conductivities and
the selection of evaporative depths.
REFERENCES
1. Schroeder, P.R., J.M. Morgan, T.M.
Walskl, and 4.C. Gibson. The Hydro-
logic Evaluation of Landfill Perfor-
mance (HELP) Model, Volume I, User's
Guide for Version I. U.S. Environmen-
tal Protection Agency, Office of Solid
Waste and Emergency Response,
Washington D.C., EPA/530-SW-84-OQ9,
1984. 120 pp.
2. Schroeder, P.R., A.C. Gibson, and M.D.
Sraolen. The Hydrologic Evaluation of
Landfill Performance (HELP) Model,
Volume II, Documentation for Version
I. U.S. Environmental Protection
Agency, Office of Solid Waste and
Emergency Response, Washington D.C.,
. EPA/530-SS-84-010, 1984.
3. Ham, R.K. Decomposition of Residen-
tial and Light Commercial Solid Waste
in Test Lyslmeters. SW-190c, U.S.
Environmental Protection Agency, Of-
fice of Solid Waste, Washington, D.C.,
1980. 103 pp.
X
4. EMCON Associates. Sonoma County Solid
Waste Stabilization Study. EPA/530-
SW-65d.l, U.S. Environmental Protec-
tion Agency, Office of Solid Waste
Management Programs, Washington, B.C.,
1975. 283 pp.
5. Wigh, R.J. Landfill Research at the
Boone County Field Site, Final Report.
Municipal Environmental Research Labo-
ratory, Office of Research and Devel-
opment, tj.S, Environmental Protection
Agency, Cincinnati, Ohio, 1984. NTIS
PB 84-161546, 116 pp.
6. EMCON Associates. Field Assessment of
Site Closure, Boone County, Kentucky.
Municipal Environmental Research Lab-
oratory, Office of Research and Devel-
opment, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1983. NTIS
PB 83-251629.
123
-------
INTERFACIAL STABILITY OF SOIL COVERS
LINED SURFACE IMPOUNDMENTS
D. H. Mitchell and T. E. Sates
Pacific Northwest Laboratory
Richland, Washington 99352
ABSTRACT
The factors affecting the interfacial stability of soil covers on geomembranes were
examined to determine the maximum stable slopes for soil cover/geomembrane systems.
Several instances of instability of soil covers on geomembranes have occurred at tailings
ponds, leaving exposed geomembranes with the potential for physical damage and possibly
chemical and ultraviolet degradation. From an operator's viewpoint, it is desirable to
maximize the slope of lined facilities in order to maximize the volume-to-area ratio;
however, the likelihood for instability also increases with increasing slope.
Frictional data obtained from direct shear tests are compared with stability data ob-
tained using a nine-square-meter (m ) engineering-scale test stand to verify that direct
shear test data are valid in slope design calculations, Interfacial frictional data from
direct shear tests using high-density polyethylene and a poorly graded sand cover agree
within several degrees with the engineering-scale tests. Additional tests with other
soils and geomembranes are planned. The instability of soil covers is not always an
interfacial problem; soil erosion and limited drainage capacity are additional factors
that must be considered in the design of covered slopes.
This work was funded by the U.S. Environmental Protection Agency (EPA) and was per-
formed at Pacific Northwest Laboratory (PNL), operated by Battelle Memorial Institute for
the U.S. Department of Energy under Contract DE-AC06-76RLO 1830.
INTRODUCTION
Geomembranes (flexible membrane
liners) are required in hazardous waste
surface impoundments and landfills to pre-
vent migration of contaminants to the soil
and groundwater. Soil covers are some-
times placed over the membrane liners in
landfills and surface impoundments. The
principal function of a soil cover is to
protect the liner from damage by traffic
(vehicle, animal, and human), weather, gas
uplift, vandalism, and chemical degra-
dation. A soil cover can also provide
drainage, e.g., leachate collection in a
landfill.
Covered liners at impoundments and
landfills are typically on 3h:lv (hori-
zontalsvertical) slopes. It is desirable
to maximize the slopes to maximize the
facility capacity-to-area ratio. Slope
steepness may be limited due to
1) limitations of construction techniques,
2) stability limits of underlying soil,
3) tensile properties of the flexible
membrane liner, or 4) stability of the
soil cover.
Instability of soil covers on sloped
liners is a concern because loss of the
soil cover exposes the liner to damage by
the mechanisms previously mentioned. In
addition, the sloughing action of the soil
on the liner can tear the liner or cause
it to pull out of the anchor trench, pos-
sibly resulting in loss of containment.
A methodology to determine the sta-
bility of soil covers on geomembranes is
needed to evaluate proposed designs of
soil cover/geomembrane systems. This
paper discusses the interfacial stability
124
-------
of soil cover/geomembrane systems,
including occurrences of instability.
Interfaeial stability analyses and the
results of engineering-scale tests, used
to verify the stability analyses, are
presented.
PERFORMANCE OF SOIL COVERS
The published literature does not
suggest frequent loss of soil covers on
geomembrane-lined slopes. In fact, no
examples of loss of covers at hazardous
waste surface impoundments have been
reported. PNL conducted an informal sur-
vey of geomembrane manufacturing, design,
and hazardous waste industries and found
no reports of loss of cover soils on
liners. Typically, landfill slopes are
not covered until the fill height necessi-
tates a cover; hence, sloughing at land-
fills is not common. However, since slope
instability has been reported for other
applications of soil cover/geomembrane
systems (Table 1), it is prudent to
investigate the phenomenon as it may
relate to hazardous waste facilities.
Two cases of loss of soil cover at
lined tailings impoundments (6), with
3h:lv slopes, occurred during conditions
that may have quickly saturated the cover
soil. In the first case a rapid thaw
melted a snow cover; in the second case,
extremely heavy rainfall occurred. The
increased pore-water pressure in the soil
decreased interfacial friction, possibly
causing the apparent interfacial
instability.
PLANAR STABILITY ANALYSES FOR A LINED
SLOPE
The slope stability analysis for soil
masses is well developed, and three major
elements have been identified as necessary
to extend the conventional slope stability
analysis to cases of geomembrane-lined im-
poundment slopes (5):
I) data on the limiting shear strength
along interfaces between soils, geo-
membranes, and geotextiles,
2) the effect of tension in the liner
system (provided by the anchor trench) on
the overall slope stability, and
3) the effect of slippage between soils,
geomembranes, and geotextiles, and its
relationship to the general stress-strain
behavior of the materials.
TABLE 1. GEOMEMBRANE-LINED SLOPE INSTABILITIES
Cover Soil
Application Liner Type* Type Depth, m Slope
Problem
Canal liner PVC
Sand; 0.6
gravel
2h:lv Torn liners; slippage
between liner and
subgrade
Tai lings
pond
Tailings
pond
Tailings
pond
Lagoon
HOPE
alloy
PVC
CSPE
Vinyl
Sand
Silty
clay
Clay-
silt
Sand;
gravel
0.6
0.46
0.3
0.3
3h:lv
3h;lv
3h:lv
1.2h:lv
Sloughing of cover
Wave erosion
Sloughing of cover
Sloughing of cover
tearing of liner
soil
soil
soi 1 ;
* PVC = polyvinyl chloride; HOPE = high density polyethylene; CSPE = chloro-
sulfonated polyethylene.
125-
-------
Planar stability analyses for a soil
cover/geomembrane system have been pro-
posed in papers by Martin, Koerner, and
Whitty (5) and Giroud and Ah-Line (3). In
both papers, the slope stability analysis
follows the case of failure along a known
surface, once the values of friction are
known between the various interfaces
involved.
Martin, Koerner, and Whitty (5) de-
scribe a limiting equilibrium method of
analysis for a liner failure along a known
surface. If interfacial friction values
are known, a force polygon can be drawn
(Figure 1) and will consist of the fol-
lowing items:
W,
TAT» TA>
A»
NB
'NB
"NB
«SA, SNB,
» weight of the cover soil in
the active zone and neutral
block, respectively
= tensile strength of the
liner in the anchor trench,
active zone, and neutral
block, respectively
= possible resistance to
failure of a small wedge of
cover soil at the toe of
the slope in the active
zone and neutral block,
respectively
= frictional forces in the
active zone and neutral
block, respectively
= frictional angles in the
active zone and neutral
block, respectively.
After assuming a factor of safety,
FS» and applying it to 6n and fiNg, a force
polygon for the neutral block is drawn to
obtain a trial value for E^D. E^ is then
made equal to E^g and used to construct
the force polygon for the active zone.
Closure of the active zone polygon indi-
cates that the initially assumed factor of
safety was correct. If closure of the
active zone polygon is not achieved, suc-
cessive trials using different values of
6^ and SMR (adjusted by the factor of
safety) are required until a graph can be
drawn to accurately assess the actual
factor of safety. This usually entails
three or four trials.
The use of a factor of safety in sta-
bility analyses 1) allows for the margin
of error among the parameters used in de-
sign and those that may actually exist in
the field, and 2) limits strains. Many
.
Liner System
""////////M
NB
TNB
Neutral Block
Active Zone
Figure 1. Soil cover/geomembrane system
under incipient slippage fail-
ure with corresponding force
polygons (5).
soils experience relatively large plastic
strains as the magnitude of the applied
shearing stress approaches the shear
strength of the soil. Thus if the
ultimate strength of the soil is used in
design, a factor of safety greater than
1.0 assumes that the soil will maintain
strains within allowable limits.
Acceptable factors of safety depend
on 1) the degree of uncertainty of the
strength of the soil and the mode of fail-
ure in classical soil engineering, 2) the
potential danger involved if a failure oc-
curs, 3) the likelihood that certain
loading conditions will develop, and
4) the expected duration of loading. The
suggested minimum factor-of-safety values
for design of earthen slopes range from
1.1 to 1.5 (1).
126
-------
Giroud and Ah-Llne (3) identify two
causes of instability of earth covers on
geomembrane-lined slopes: a slide within
an earth cover and a slide at the inter-
face between any two components (soil
cover, geotextile, geomembrane). They de-
scribe a general method for evaluating the
risk of sliding at any soil/ geomembrane
interface and apply it to the critical de-
sign situations of rapid drawdown and fil-
ling of a reservoir. A wedge analysis
(Figure 2} is used to determine the force
per unit width in the liner (or geotextile
in this case). This value is then com-
pared to the maximum pull out resistance
provided by the anchor trench as shown in
Equation 1.
Sloughing of soil covers over liners
can be compared to shallow failure in the
surface stratum of slopes as found in
nature. The difference in permeability
between surface and substratum allows for
an increase of pore-water pressure in the
surface stratum with the percolation of
Geotextile 2 3
Figure 2. Slide along a geomembrane-
lined reservoir. 1» 2, 3,
and 4 refer to zones in the
subsequent equations (3).
rainwater, resulting in a collapse of part
of the surface. The shallow failure in
the surface stratum has been investigated
by Kobashi (4). Experiments were per-
formed using a small test frame to confirm
the conditions where shallow failures
occur when increases in pore-water pres-
sure cause piping and flow failure.
Neither of the previously described sta-
bility analyses account for the effect of
pore-water pressure on the sloughing of
soil covers over geomembranes; hence, a
method is provided to account for this
effect.
Consider the forces acting in the '
analysis of a slide on-a planar surface of
failure in an infinite slope (Figure 3).
In terms of effective stresses, the shear
strength mobilized under conditions of
Plane of
Weakness
(yZ cos2 0-M) tan f
Figure 3. Soil mass resting on inclined,
impervious layer.
cos 0 + sing tan (g -
(1)
w sine tan(g - 4.3) + cosg
(•5— tsn^-, + C,) (CO'SB - sins tan,j>,)
COSB •*• sing tan(e - 413)
where
-------
limiting equilibrium on that plane may be
expressed by
T = C1 + (crn - „) tan*' (2)
where T = shear strength
cohesion intercept
effective normal stress
angle of shearing resistance
pore-water pressure
C'
n
The factor of safety, F^, for this
system is given by the ratio of forces re-
sisting movement to those promoting move-
ment
± C'
(3)
where y = unit soil weight and
Z = average vertical soil depth
The frictional force is effectively re-
duced by the reduction in normal stress
due to pore-water pressure.
For a geomembrane-lined slope with a
multilayered cover as depicted in Fig-
ure 3, Equation 3 can be rewritten as
(4)
where
jij, v = pore-water pressure for zones
1 and 3, respectively
Zj, Zg, Zg = average vertical depth of
soil in zones 1, 2, and 3,
respectively
YJ, Y2> YS * unl* weight of soil in zones
1, 2, and 3, respectively
C|, C| = cohesion intercept for zones
1 and 3, respectively
Lp Lg = length of zones 1 and 3.
This equation is applicable when the toe
of the slope is required to produce a
stable cover.
The formulation developed accounts
for the change in pore-water pressure and
its effect on the normal and frictional
forces at the soil/geomembrane interface.
Once friction angles are determined for
the specific soil/geomembrane interface,
the methodology posed by Giroud and Ah-
Line (3) can be used to determine whether
pullout of the anchor trench or tearing of
the geomembrane will occur.
DETERMINATION OF INTERRACIAL FRICTION
VALUES
To use the stability analyses devel-
oped above, interfacial friction angles
are required. Limited experiments have
been reported on combinations of soil
cover/geomembrane systems, geotextile/
geomembrane systems, and soil cover/
geotextile systems (2,5,7).
The test apparatus used to measure
friction angles was a modified direct
shear apparatus, where the geomembrane was
attached to a solid block and soil was
placed above the geomembrane. The normal
stress applied varied from 13.8 to 206.7
kilonewtons per square meter (kN/m ) in
the above references. Displacement rates
of 0.13 to 0.76 millimeters per minute
(mm/min) were reported. In each case
described in the above references, the
friction between the soil and liner was
less than that of the soil itself. Inter-
facial friction was 60% to 80% of the soil
internal friction.
ENGINEERING-SCALE TESTS
Using the analyses for interfacial
stability and the experimental friction
angle from direct shear tests, it is pos-
sible to calculate a factor of safety for
interfacial stability of a given impound-
ment or landfill design. However, it is
necessary to verify that the interfacial
friction angles obtained with direct shear
tests are representative of field condi-
tions. If not, direct shear test proce-
dures may require modification, or more
conservative factors of safety will need
to be established. There are several pos-
sible reasons for bias of the direct shear
test:
1) The attachment of a liner to the di-
rect shear apparatus may create an uneven
surface that might bias the determined
friction angle.
2) The scale of the direct shear test
may induce edge effects that bias the
measured frictional angle.
128
-------
3) The direct shear test is performed at
normal stresses of 13.8 to 206.7 kN/m »
while typical cover stresses might range
from 6.9 to 20.7 kN/m .
4) The ability of less rigid liners to
conform to small bumps and disturbances in
the subgrade may effectively increase the
friction angle over that measured in the
direct shear test,
To determine the applicability of
direct shear tests for obtaining soil/
geomembrane friction angles, larger-scale
tests are being performed. The engineer-
ing-scale test stand depicted in Figure 4
is used to test the stability of soil
cover/geomembrane systems. A flexible
membrane liner (3 m by 3 m) is placed over
a soil base, attached to the inner edges
of the test stand, and covered with
soil. Pore-water pressure sensors are
buried in the soil at the top surface of
the liner. A rain simulator, which dupli-
cates the droplet size and velocity of a
heavy Midwest rainstorm, provides moisture
to increase the pore-water pressure of the
soil above the liner. Linear variable
displacement transducers (LVDTs), used to
detect movement of the soil, are located
at the elevated end of the test stand, and
sensing rods are buried in the soil as
diagrammed in Figure 5.
An experimental procedure has been
developed through a series of tests. Soil
is placed on the geomembrane at the de-
sired compaction as shown in Figure 5,
The density of the soil and its moisture
content are measured prior to initiation
of a test. The test stand is brought to
the starting angle of 14° unless prior
tests indicate that a greater starting an-
gle is permissible. Water is added to the
toe of the slope so that essentially no
stability is provided by the soil at the
toe, and then the rain simulator is start-
ed. When the soil in the lower half of
the test stand is fully saturated (mea-
sured by transducers), the slope is
increased in 0.5° to 1° increments at
Figure 4. Engineering-scale test stand.
129
-------
LVDT Sensor
Soil
Geomembrane
w
Pressure Transducers
FIGURE 5. Cross-section of the
engineering-scale
test stand.
30-minute intervals. Interfacial
instability is first detected by movement
of LVDT sensors, and within 30 minutes of
initial detection, a visible shear line
forms in the soil cover. A test is term-
inated when a shear line is detected.
Two cases have been investigated to
date, using poorly graded angular sand (SP
according to the Unified Soil Classifica-
tion System). The first case used HOPE
and the second used reinforced CSPE.
Three trials of the HOPE case result-
ed in Interfacial instability of the soil
cover at angles of 18.5°, 18.5°, and 19°
with a soil density of 15.1 to 15.7 kN/nr
(19.5 kN/nr when saturated). The soil
depth was 15 cm. Using Equation 3 and a
factor of safety of 1.0, the instability
corresponds to a measured soil friction
angle of 35°. Table 2 lists friction
angles determined from direct shear tests
with poorly graded angular sand. The
friction angle from the engineering-scale
tests is 3° to 5° greater than the fric-
tion angle determined in the direct shear
tests (if cohesion values are neglected).
Interlocking of the sand particles may be
responsible for the measured cohesion in-
tercept by direct shear tests. If this is
the case, a cohesion intercept would not
be expected when the shear direction is
reversed after the soil has been
disturbed.
Using reinforced CSPE, an interfacial
failure angle could not be determined with
the engineering-scale test stand because
surface erosion occurred. As the test
stand slope was increased above 20°,
surface erosion became severe. Piping
also occurred as inflow of water exceeded
the soil drainage capacity. Movement of
the soil was not detected by the LVDTs
during tests where erosion caused loss of
soil cover. Using the direct shear test,
two soil-testing laboratories determined
friction angles of 32° and 38° with this
soil on CSPE (Table 2), Additional shear
tests with these materials are required to
determine an average friction angle for
this combination of materials. The
erosion of soil in the engineering-scale
test stand will limit the cases of inter-
facial instability that can be investi-
gated with this equipment.
CONCLUSIONS
Existing interfacial stability analy-
ses for soils on geomembranes have been
extended to account for the reduction in
soil cover stability due to increased
pore-water pressure. The analyses can be
used to calculate the factor of safety of
a soil cover/geomembrane system. Factor
of safety analysis requires information on
soil densities, maximum expected pore-
water pressure, interfacial friction
angles for all contacting components, and
slope'design. Given the needed data, the
designer can determine whether a proposed
cover design will have the required factor
of safety at the soil/geomembrane inter-
faces. The adequacy of liner anchors can
also be assessed.
Several cases of reported instability
of soil covers on liners have occurred
during conditions when the soil cover was
quickly saturated with water. When this
occurs, the downward forces are magnified
by the water weight, and the soil cover/
liner frictional forces on the slope are
decreased due to lower effective stress on
the liner. Therefore, the designer must
consider the probability of extreme pre-
cipitation or rapid thaws (which may lead
to saturation of the soil cover) and de-
sign a system with the required factor of
safety under worst-case conditions.
Engineering-scale tests are being
performed to verify the proposed stability
analyses and to verify that friction
angles measured by the direct shear test
agree with those determined through
130
-------
TABLE 2. RESULTS OF DIRECT SHEAR TESTS USING POORLY GRADED SAND
Dry Soi
Laboratory Liner Height, kN/m
Friction
Angle, degrees
Cohesion. kN/V
A
B
A
B
HOPE
HDPE
CSPE
CSPE
18.1
16.5
18.1
16.5
32
30
32
38
4.1
4.1
0.0
2.1
larger-scale tests. In one case, the
friction angle determined using the
engineering-scale test stand was 3° to 5°
greater than the friction angle measured
by direct shear tests. In other words,
the direct shear data would result in a
conservative design. More engineering-
scale tests will be performed in 1986.
Interfacial stability is not the only
criterion for designing a stable soil
cover/geomembrane system. As the slope
increases, surface erosion can also become
a mechanism for loss of cover soils, even
though the system has a stable interface
as evidenced by engineering-scale tests
and in actual applications.
REFERENCES
1. Canada Centre for Mineral and Energy
Technology (CANMET). 1982. Waste
embankments. Pit SI ope Ha nu a 1j>
CANMET Report 77-01.
2. Gollios, A., P. Delmos, J. P. Gourc,
and J. P. Giroud. 1980. Experiments
on Soil Reinforcement with Geotex-
tiles.The Use of Geotextilesfor
Soil ImpToyement, ASCE NationalCon-
vention, Portland, Oregon, pp. 80-
.177.
3. Giroud, J. P. and C. Ah-Line.
1984. Design of earth and concrete
covers for geomemb'ranes. Proceedings
of the InternationalConference on
Geomembranes, Industrial Fabrics
Association International, St. Paul,
Minnesota, pp. 487-492.
4. Kobashi, S. 1971. Erosion and
surface stratum failure of steep
slopes and their prevention methods.
Proceedings of Fourth AsianRegional
Conference on Soil Mechanics and
Foundation Engineering, pp. 433-436.
5. Martin, J. P., R. M. Koerner, and
J. E. Whitty. 1984. Experimental
friction evaluation of slippage be-
tween geomembranes, geotextiles and
soils. Proceedings of the Interna- '
tional Conference on Geomembranes,
Industrial Fabric Association Inter-
national, St. Paul, Minnesota,
pp. 191-196.
6. Mitchell, D. H. 1984. Technology
for Uranium Mill Ponds Using Geomem-
branes, NUREG/CR-3890, available from
the NTIS, Springfield, Virginia.
7. Saxena, S. K., and Y. T. Wong.
1984. Frictional characteristics of
a Geomembrane. Proceedings of the •
International Conference on Geo-
membranes , Industrial Fabric Associa-
tion International, St. Paul,
Minnesota, pp. 187-190.
131
-------
DETERMINATION OF THE SOLUBILITY PARAMETERS OF FMLS FOR
USE IN ASSESSING RESISTANCE TO ORGANICS
Henry E. Haxo, Jr., Pearl Pick
Matrecon, Inc.
Oakland, California 94608
ABSTRACT
Results are presented of equilibrium volume swelling of 22 different polymeric FMLs
and six simple polymeric compounds of known composition on immersion in water and
in 23 organics ranging widely in dispersive, polarity, and hydrogen-bonding solubility
parameters. Preliminary determinations of these three component solubility parameters
of FMLs based on their equilibrium swelling characteristics in test liquids of various
known solubility parameters are made by two graphical methods. The three solubility
parameters of FMLs are being generated to use in developing a predictive model for
assessing FML-waste liquid compatibility. Results indicate the potential importance
of the solubility parameters of the FMLs with respect to predicting the absorption of
specific organic constituents from waste liquids.
INTRODUCTION
The absorption of even small amounts
of constituents from a waste liquid may
have adverse effects on the performance
of a flexible membrane liner (FML) and
other polymeric components of a liner
system in a waste storage or disposal
facility. Absorption may occur either
through contact with the waste liquid or
with vapor from the liquid. Possible
adverse effects include softening, loss
of structure, loss of tensile strength,
loss of tear and puncture resistances,
increasing permeability, creep and
eventual puncture, and in the case of
semicrystalline materials, environmental
stress-cracking.
The significant compositional fac-
tors that can contribute to the swelling
of a specific polymeric compound in a
liquid are:
- Solubility parameter of the polymer
relative to that of the liquid.
- Degree of crosslinking of the
polymeric compound.
- Degree of crystal!inity of the
polymer.
- Filler, plasticizer, and soluble
constituents in the total compound.
The work performed for this paper is
part of a project for developing a mathe-
matical model for predicting the compati-
bility and long-term effect of waste
liquids on FMLs. Such a model must con-
sider the compositional factors listed
above as well as the concentration of
individual dissolved constituents in a
waste liquid and other components of the
environment in which the FML is expected
to perform. This paper is a preliminary
report of experimental work performed to
develop the information necessary for
such a model. The solubility or absorp-
tive characteristics of the liner with
respect to the waste constituents are of
primary importance inasmuch as the
entrance of waste constituents into a
liner is the initial step in the degra-
dation of an FML in service. Determina-
tion of the solubility parameters of
polymeric materials is a first step in
developing such a model.
132
-------
This paper presents a continuation of
the work reported at the llth Symposium
and deals principally with the experi-
mental evaluation and study of solubility
parameters of polymeric FMLs.
OBJECTIVES
The four major objectives of this
project are:
1. To develop methodology that can
be used to predict the compati-
bility of polymeric membranes
with specific waste liquids and
can estimate their service lives
for lining hazardous waste
management facilities.
2. To explore the feasibility of
using solubility parameters to
measure the compatibility of
polymeric membranes with waste
liquids.
3. To prepare a predictive model
and test it against an existing
database.
4. To determine the effects of
swelling on mechanical and
permeability characteristics of
membrane liners in service.
This paper deals primarily with the second
objective.
TECHNICAL APPROACH
The total environment in which an
FML exists ultimately determines its
service life in a facility. Any model
for predicting the compatibility and
long-term durability of an FML with
respect to its use in a waste storage and
disposal facility will need to consider
the many factors that contribute to the
total environment in which the FML is
expected to perform. Some of the various
environmental factors that affect dura-
bility and service life are listed in
Table 1. Ultimately, the importance of
each of these factors for a given situ-
ation must be determined.
Of these factors, those given prime
consideration in this project are the
compatibility of the FMLs with waste
liquids and the various stress factors,
including both chemical and mechanical
stress. In reality, both stresses func-
tion simultaneously in the environment of
an FML in service. The liner is not
placed on a planar surface, but on three
dimensions with irregularities caused by
voids, uneven compaction, stones, settle-
ment, etc. When the FML is under mechan-
ical load, these irregularties, regardless
of size, can result in deformations,
bends, biaxial stressing, and, in the
proximity of welds or seams, in dif-
ferences in stresses in the FMLs. Thus,
in order to simulate accurately conditions
that exist where a liner is in service,
testing must be performed on FMLs that are
simultaneously under mechanical stress
and in the presence of waste liquids that
may be absorbed by the liner and result
in damage or changes in properties of the
FML.
TABLE 1. ENVIRONMENTAL FACTORS AFFECTING
DURABILITY AND SERVICE LIFE^
Factors based on compatibility with waste liquids:
Chemical
Physical
Stress factors:
Stress, sustained and periodic
Stress, random
Physical action of rain, hail, sleet, and snow
Physical action of wind
Movement due to other factors, e.g. settlement
Discontinuity at penetrations
Stress in the presence of chemicals
Weathering factors based on geographic location:
Solar radiation
Temperature
Elevated
Depressed
Cycles and fluctuations
Water — solid, liquid and vapor
Normal air constituents, e.g. oxygen and ozone
Freeze-thaw and wind
Use and operational factors:
Design of system, groundwork and installation
Operational practice
Biological factors
aFrom Haxo and Nelson (1984).
Our first concern in this study was
with the absorption by the FML of con-
stituents of the waste which may result
in softening, loss of tensile and other
strength properties, and cracking.
133
-------
Inasmuch as these effects begin with
the absorption of waste constituents, the
initial work explored the use of solu-
bility parameters for predicting organic
absorption by an FML. The solubility
parameter concept, as applied to FMLs, is
described in our paper presented at the
llth Symposium (Haxo et al, 19S5a). As
1s described in that paper, the disper-
sive, polarity, and hydrogen-bonding
solubility parameters (Hansen and Beer-
bower, 1971) of individual commercial
membranes are being determined experi-
mentally by measuring the equilibrium
swelling of the FMLs in solvents having
known values of these solubility para-
meters. Based on this swelling data, the
solubility parameters of individual FMLS
will be determined. These solubility
parameters can then be used to assess
the compatibility of FMLs with specific
organics, the parameters of which need to
be known.
The use of these parameters, however,
are only applicable to amorphous polymers
and to the amorphous phases of semi crys-
talline polymeric compositions, which con-
sist of both amorphous and crystalline
phases. The amorphous phases are between
the crystalline phases or spherulites.
These amorphous domains absorb organics.
On the other hand, the crystalline phases
of polymeric compositions resist the ab-
sorption of solvents at normal temper-
atures. Consequently, the degree of crys-
tal! inity, which ranges from <30% to >80%
in polyethylenes, for example, is a major
factor to be investigated in assessing
the swelling of a lining material.
Additional factors such as crosslinking
and compounding ingredients, particularly
plasticizers and fillers, also can con-
tribute to the level of absorption and
swelling.
DETERMINATION OF SOLUBILITY PARAMETERS OF
FMLS BY SWELLING IN SELECTED SOLVENTS
Design of the Experiment
The component solubility parameters
of the lining materials are being deter-
mined experimentally from the changes in
volume measured at equilibrium of speci-
mens Immersed in selected pure organics
with a broad range of established
solubility parameters.
It was recognized that FMLs are made
from a variety of polymers and mixtures
of polymers with additives, many of which
could affect the ultimate change in
volume of the lining material on contact
with waste liquid. Possible additives
include crosslinking agents which tend to
reduce swelling, and plasticizers which,
if they are monomeric, can be extracted
by the waste liquid; the liner would then
shrink and, in some cases, become stiff.
Such effects can seriously complicate the
determination of the solubility parameters
of the polymers in the FML. Consequently,
simple compounds of single polymers, with
slight crosslinking to prevent dissol-
ution, were included in the immersion.
In order to cover the range of
possible component solubility parameters
of the polymers used in the manufacture
of liners, particular attention was paid
to the selection of the solvents in order
to include a broad range of component
solubility parameters. The selection of
the organics, the FMLs, and the prepara-
tion of the simple polymer compounds are
described in the following sections.
Selection of Test Organics
The first task in this study was to
choose a set of test organics that met our
stated criteria for use in this project.
Basic requirements were that:
- The full range of values of the
different component solubility
parameters should be covered.
- Known common constituents found in
waste streams should be included.
- Organics known to be in the wastes
that were tested in our previous
work, and which will be used in
the verification of the final
model, should be included.
Other factors considered were the follow-
ing:
- Solubility of the organic in water.
- Density, melting point, and boiling
point of the organic.
- Vapor pressure of the organic.
134
-------
The most difficult criterion to
satisfy was to cover the full range of
solubility parameters. Investigation of
published data (Hansen and Beerbower,
1971) shows that consideration of the
Hi 1 debrand or total solubility parameter
alone is insufficient to assess the com-
patibility of two or more solvents and
of polymers with solvents. This was
confirmed for lining materials following
evaluation of data obtained at Matrecon
(Haxo et a!, 1985) of the swelling of
liner samples in various organics. In
these references, instances are reported
in which solvents and nonsolvents for a
given polymer have close or identical
Hildebrand solubility parameters.
Clearly, more data are required to
predict correctly liner compatibility
with waste liquids. Hansen proposed the
use of three solubility parameters based
upon the division of intermolecular
forces into dispersive (8$), polar (6p),
and hydrogen-bonding (6^,) components
(Hansen and Beerbower, 1971).
Choosing test solvents based upon
information on component solubility
parameters was more complex than making
the choice on the basis of Hildebrand
parameters alone. By selecting solvents
representing the full range of each of the
component parameters, interactions between
the three parameters could be analyzed.
The solvent selection was made by using a
computer to classify the solvents with
respect to component solubility parameters
and then applying the other selection
criteria to make the final choices.
The classification procedure was as
follows: The 131 solvents in the data base
from.Barton (1975) were listed in order
of increasing dispersive component (6
-------
TABLE 3. TEST SOLVENTS AND SELECTED PROPERTIES
Property3
Name
n-Octane
Isoamyl acetate
2-Ethyl-l-hexanol
Methyl Isobutyl ketone
Ethyl acetate
n-Propanol
Hltroethane
Methyl ethyl ketone
Methanol
Cyclohexane
01 ethyl carbonate
Cyclohexanol
D1(2-ethylhexyl)
phthalate
Cyclohexanone
Furfuryl alcohol
01 ethyl phthalate
N,H-01methylacetam1de
Ethylene glycol
Tetralln
Trl ch 1 oroethy 1 ene
m-Cresol
Tet rachl oroethy 1 ene
Qulnollne
Benzyl alcohol
Propylene carbonate
Butyrolactone
2-Pyrrolldone
Water
Isooctane
Xylenes (o,m,p)
Acetone
Solubility parametersb
7.6
8.4
9.9
8.3
8.9
12.0
11.1
9.3
14.5
8.2
8.8
10.9
8.9
9.6
11.9
10.0
11.1
16.1
9.8
9.3
11.1
9.9
10.8
11.6
13.3
12.9
13.9
23.4
7.0
8.8
9.8
aLange's Handbook of Chemistry
7.6
7.5
7.8
7.5
7.7
7.8
7.8
7.8
7.4
8.2
8.1
8.5
8.1
8.7
8.5
8.6
8.2
8.3
9.6
8.8
8.8
9.3
9.5
9.0
9.8
9.3
9.5
7.6
7.0
8.7
7.6
0
1.5
1.6
3.0
2.6
3.3
7.6
4.4
6.0
0.0
1.5
2.0
3.4
3.1
3.7
4.7
5.6
5.4
1.0
1.5
2.5
3.2
3.4
3.1
8.8
8.1
8.5
7.8
0
0.5
5.1
0
3.4
5.8
2.0
3.5
8.5
2.2
2.5
10.9
0.1
3.0
6.6
1.5
2.5
7.4
2.2
5.0
12.7
1.4
2.6
6.3
1.4
3.7
6.7
2.0
3.6
5.5
20.7
0
1.5
3.4
no,
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
, Revised 10th Edition,
Solvent
. and codec
HP,
Rp D I °C \ t
d Kp Kh \ ^/ \
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
1
1
2
1
1967.
1 1
1 2
1 3
2 1
2 2
2 3
3 1
3 2
3 3
1 1
1 2
1 3
2 1
2 2
2 3
3 1
3 2
3 3
1 1
1 2
1 3
2 1
2 2
2 3
3 1
3 2
3 3
3 3
1 1
1 1
3 2
HP is
-57
-78
-76
-80
-84
-127
-90
-87
-98
6.5
-43
22
-50
-47
-29
-3
-20
-13
-35
-87
10
-22
-15
-15
-55
-45
25
0
-107
Solubility in
BP, Density, 100 parts H20
°C) (g/cm3) (Temperature, °C)
125
142
183
114
77
97
112
80
65
81
126
160
384
155
170
298
165
196
207
87
203
121
237
205
240
204
245
100
99
-48-+13 138-144
-95
the melting
56
point
0.703
0.876
0.833
0.801
0.902
0.804
1.045
0.805
0.791
0.779
0.975
0.963
0.981
0.947
1.135
1.118
0.937
1.113
0.973
1.462
1.034
1.623
1.095
1.045
1.189
1.12
1.12
1.00
0.692
0.868
0.791
and BP
0.002 (16°C)
0.25 (15°C)
0.10 (20°C)
2.0 (20°C)
8.5 (15°C)
Miscible
4.5 (20°C)
37.0
Hiscible
Insoluble
Insoluble
3.6 (20°C)
4 x 10-5 (25°C)d
4.8 (30°C)
Hiscible
0.09 (25°C)d
Hiscible
Hiscible
Insoluble
0.10 (25°C)
0.50
0.02 (20°C)
0.6
3.8 (17°C)
Hoderate
Hiscible
Hiscible
...
Insoluble
Insoluble
Miscible
is the boiling
point of the solvent.
bBarton, 1975. SQ = Hlldebrand solubility parameter.
cEach code was made up of three digits representing the range values assigned to a solvent after the
ordering of solvents by each component solubility parameter as presented in Table 2. R^ is the range
value for the dispersive component, Rn the range value for the polar component, and Rn the range value
for the hydrogen-bonding component. Each solvent selected as a test solvent was assigned a unique number.
dEPA Report 440/4-81-014, December 1982, "Aquatic Fate Process Data for Organic Priority Pollutants."
136
-------
variations. One way to fulfill
this requirement is to select
materials of the same type as pro-
duced by different manufacturers.
5. Polyethylenes of different crys-
tallinity and type should be in-
cluded, i.e. ranging from low-
density to high-density.
6. Polyethylene alloy with an elasto-
mer to reduce environmental
stress-cracking should be included.
7. FMLs previously tested by Matrecon
should be included.
8. Simple compounds of known composi-
tion should be included. Simple
compounds contain a minimum
number and amount of compounding
ingredients and are usually
lightly crosslinked. These
compounds are useful for determin-
ing the solubility parameters of
polymers because they reduce the
effects of compounding ingredients
on the absorption of organics.
Also, they generally swell with-
out dissolving.
All the materials selected for testing
are listed in Table 4, which includes
data on extractables and specific gravity.
Table 5 presents the laboratory-prepared
compounds included in this study. These
compounds include four CSPE compositions,
a nitrile rubber composition, and a PVC
composition plasticized with di(2-ethyl-
hexyl) phthalate. Among the four CSPE
compositions are three gum compounds that
vary in the level of crosslinking. Gum
compounds contain no filler. The fourth
CSPE composition, which contains 100 parts
of carbon black, is similar to that used
in CSPE liner compounds.
Test Procedure
To determine the equilibrium swelling
of each polymeric material, the volume of
an immersed sample was monitored until
equilibrium was reached. This sample
consisted of three liner specimens, which
were placed in a 20 ml disposable scintil-
lation vial for each liner-solvent com-
bination. The inside of the vial caps
was lined with Teflon-coated aluminum
foil to prevent the loss of organics.
The 20 ml vial was satisfactory for most
liner-solvent combinations; however, in'
those cases in which the liner-solvent
combinations resulted in considerable
swelling, the specimens were transferred
to 70 ml vials. The specimens were hung
in the vials so they did not touch each
other.
The specimens, which measured 0.5 x
1.5 in., were cut from slabs or sheetings
with a die for cutting out ASTM environ-
mental stress-cracking specimens. How-
ever, the specimens were treated as a
unit, not individually, and were weighed
as a set before and after immersion.
This procedure allows more of a sample
to be immersed in a small container and
increases the sensitivity of the deter-
mination as it provides an averaging of
data for three specimens. The drawback
is the possibility of weighing errors.
This was not a significant problem as the
test was not terminated until the weights
on two consecutive weighings showed
reasonable agreement. Before immersion,
the weight of the sets of specimens
ranged from 0.5 to 5 g with most weighing
under 2 grams.
The specimens were weighed before
immersion and after 2 days, 1 week, and
2 weeks. Additional weighings were taken
approximately once a week if the specimens
had not reached equilibrium after the
second week of exposure. Once equilibrium
swelling had been reached, the equilibrium
swelling based on change in volume was
calculated from the equilibrium swelling
based on change in weight.
Test Results
Results to date have been obtained
with 24 solvents; testing is proceeding
with the remaining seven. Data on volume
change of eight of the 28 FMLs and poly-
meric compositions are presented in Table
6. The results are presented in the order
of increasing Hildebrand solubility para-
meters. Those compositions that swelled
more than 1000%, dissolved, or disinte-
grated are designated by "D" in the table.
Those materials that had high levels of
plasticizers lost weight and shrank.
Considerable variation in the volume
changes can be observed among the lining
materials. It should also be noted that
the magnitude of swelling in many solvents
137
-------
TABLE 4. POLYMERIC LINER COMPOSITIONS IN TEST
Polymer
Chlorinated polyethylene
(CPE)
Chlorosulfonated poly-
ethylene (CSPE)
Eplchlorohydrin rubber (ECO)
Ethyl ene propylene rubber
(EPDM)
Ethylene vinyl acetate (EVA)
Neoprene (CR)
Nltrile rubber (NBR)
Polyester elastomer (PEL)
Polybutylene (PB)
Polyethylene:
Low-density (LDPE)
Linear low-density (LLDPE)
High-density (HOPE)
it 11
11 H
HDPE/EPDM-alloy (HDPE-A)
Polyurethane (PU)
Polyvinyl chloride (PVC)
Elastlclzed polyvlnyl
chloride (PVC-E)
Polyvinyl chloride -
oil -resistant (PVC-OR)
ID
number
195
335R
378R
169R
174R
DOY-3C
DOZ-2C
DPOC
Dppc
178
232
308A
168
DPNC
316
323
221A
309A
284
184
263
305
181
351
153
DPQC
176R
144
Type of
polymer3
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
XL/AM
XL/AM
TP/AM
XL/AM
XL/AM
TP/CX/AM
TP/CX/AM
CX
CX
CX
CX
CX
CX
CX
TP/AM
TP/AM
TP/AM
TP/AM
TP/AM
Extract-
ables
(50
14.85
4.48
7.94
11.29
7.15
<1.0d
<1.0d
<1.0d
<1.0d
7.63
22.78
0.75
11.23
<1.0d
1.09
£0.6
3.68
1.85
0.65
0.73
£0.6
0.98
2.09
1.50
34.57
40.12
9.13
30.97
Property
Extraction
solvent13
n-heptane
n-heptane
n-heptane
acetone
acetone
• * •
* * *
* • *
* • *
MEK
MEK
ethanol
acetone
• * *
n-heptane
n-heptane
MEK
MEK
MEK
MEK
MEK
MEK
MEK
n-heptane
2:1 CC14:
CH3OHe
2:1 CC14:
CH3OHe
CHaOH
2:1 CC14:
CH3OH6
Specific
gravity
1.260
1.313
1.333
1.297
1.364
1.147
1.401
1.193
1.143
1.458
1.166
0.951
1.500
1.024
1.149
1.253
0.907
0.938
0.929
0.951
0.953
0.954
0.949
1.118
1.263
1.226
1.219
1.356
aTP * Thermoplastic; AM = Amorphous; XL = Cross!inked; CX = Partially
crystalline.
bMEK * Methyl ethyl ketone.
cLaboratory-prepared compounds, the composition of which is presented
Table 5.
^Calculated from compound formulation.
eBy volume.
138
-------
TABLE 5. COMPOSITION OF PREPARED COMPOUNDS OF CSPE,
NITRILE RUBBER, AND POLYVINYL CHLORIDE3
CSPE
Ingredient
DOY-3b DOZ-2b DPOb DPPb
Nltrlle
rubber
DPNb
PVC
DPQfc
CSPE (Hypalon 45) 100 100 100 100
Nitrile rubber
(Hycar 1052-30) ... ... ... ... 100
PVC (Geon 135) ... ... ... ... ...• 100
Di(2-ethylhexyl)
phthalate ... ... .... ... ... 50
MT black ... 100
MgO (Maglite D) 4 4 10 4
Peroxide (Varox
powder) ... ... 6 1.5
HVA-2C ... ... 2 0.5
Lead stearate *.. ... ... ... ... 2.0
Stearic acid (F300) ... ... ... ... 1.0 0.2
ZnO (Protox 168) ... ... ... ... 5.0
TMTDSd (Tuex) ... ... .... ... 3.5
Carbowax 1.5 1.5 ... ... ... ...
Pentaerythritol 3.0 ...
Tetrone Ae 2.0 ... ... ... ... ...
aFormulation in parts by weight.
bMatrecon identification code.
GHVA-2 = N,N-m-phenylenedimaleimide (DuPont) a curing adjuvant for CSPE.
dTMTDS = Tetramethyl thiuram disulfide.
eTetrone A = Dipentamethylene thiuram hexasulfide, accelerator or sulfur
source.
does not follow the Hildebrand solubility
parameter.
The volume swelling data on the com-
mercial PVC liner compound (Liner No. 153)
differed from that on the laboratory-pre-
pared PVC compound (DPQ). The differences
probably resulted from the different types
of PVC used in each. The laboratory-pre-
pared compound disintegrated more in the
solvents than the commercial liner
material.
Table 7 presents the results of the
swelling of four laboratory-prepared
CSPE compounds in the 24 solvents. The
variations in the compounding included two
peroxide-cured compounds without filler,
one with a low amount and one with a high
amount of peroxide curative. A third com-
pound without filler was with MgO cure,
which is similar to the cure used in CSPE
liner compounds, and the fourth was a CSPE
compound also with MnO cure and containing
37 volume percent of carbon black. With
the peroxide curing systems the number of
crosslinks generally is proportional to
the amount of peroxide added. The MgO
cure, which is used in potable-grade CSPE
lining materials, forms a different type
of crosslink which develops with time by
absorption of moisture.
139
-------
TABLE 6. EQUILIBRUIM PERCENT VOLUME CHANGE OF FMLS ON IMMERSION IN THE SELECTED SOLVENTS
Solvent
Isooctane (Ref. Fuel A)
n-Octane
Cyclohexane
Methyl isobutyl ketone
Isoamyl acetate
o-Xylene
Di(2-ethylhexyl) phthalate
Ethyl acetate
Methyl ethyl ketone
Trichloroethylene
Cyclohexanone
Acetone
Tetralin
Tetrachloroethylene
2-ethyl-l-hexanol
Cyclohexanol
N , N-di methyl acetami de
Benzyl alcohol
1-Propanol
Propyl ene-1 ,2-carbonate
2-Pyrrolidone
Methanol
Ethylene glycol
Water
Hildebrand
solubility
parameter
7.0
7.6
8.2
8.3
8.4
8.8
8.9
8.9
9.3
9.3
9.6
9.8
9.8
9.9
9.9
10.9
11.1
11.6
12.0
13.3
13.9
14.5
16.1
23.4
L1ner-polymer/IO number/ extractables
CPE
195
14.85a
4.24
6.08
26.6
De
D
41.2
362
D
D
D
D
103
D
123
4.57
2.14
D
35.2
2.13
23.0
112
3.56
2.21
3.99
CSPE
174R
8.71
12.9
99.7
40.8
45.4
153
29.2
16.5
26.8
D
101
13.3
181
161
3.12
5.66
20.9
8.79
3.13
5.63
11.2
4.58
3.22
4.44
EPDM
232
22.78C
68.6
86.8
126
-5.20
5.70
103
-4.16
-4.96
-8.33
135
-3.00
0.88
68.6
146
-4.83
-7.47
3.93
-0.38
-6.40
-0.08
1.29
1.59
0.49
1.18
HOPE
184
0.73C
7.06
8.49
11.8
3.62
4.24
12.6
0.46
2.63
2.53
10.5
3.03
2.42
7.67
13.7
0.52
0.86
1.67
0.94
0.77
0.28
0.87
1.40
0.41
0.62
HOPE-A
181
2.09C
15.9
18.6
34.7
4.36
7.72
28.8
3.12
3.27
3.42
32.1
5.40
1.41
16.9
41.4
1.44
1.82
2.02
0.61
1.34
0.09
0.54
2.86
0.43
1.60
PEL
323
1.09a
0.86
2.88
4.46
10.5
10.5
17.1
0.72
11.6
12.9
25.1
15.8
11.2
24.7
14.5'
1.51
2.41
16.5
17.1
7.28
3.30
4.68
4.72
1.67
2.13
PVC
153
34.57d
-21.7
-19.9
-19.9
D
245
7.92
176
147
D
17.0
D
172
111
-2.64
-12.8
-11.2
D
-11.8
-19.0
11.9
278
-17.7
3.43
1.58
DPQ
40.12d
-22.1
-20.3
-16.4
D
D
-17.3
143
D
D
276
D
D
D
0.19
3.0
0.37
D
-9.5
-19.4
27.3
D
-11.7
4.9
4.2
aWith n-heptane.
bwith acetone.
cWith methyl ethyl ketone.
dWith 2:1 mixture of CC14 and CH3OH.
eD = dissolved or disintegrated.
-------
TABLE 7. EFFECT OF COMPOUNDING VARIATIONS ON THE PERCENT VOLUME
SWELLING OF CSPE COMPOUNDS^ IN SELECTED SOLVENTS
Solvent
Isooctane (Ref. Fuel A)
n-Octane
Cycl ohexane
Methyl isobutyl ketone
Isoamyl acetate
o-Xylene
Di(2-ethy1hexyl) phthalate
Ethyl acetate
Methyl ethyl ketone
Trichloroethylene
Cyclohexanone
Acetone
Tetralin
Tetrachl oroethy 1 ene
2-Ethyl-l-hexanol
Cyclohexanol
N,N-dimethylacetamide
Benzyl alcohol
1-Propanol
Propylene-1 ,2-carbonate
2-Pyrrolidone
Methanol
Ethyl ene glycol
Water
Hildebrand
solubility
parameter
(«0)
(cal cm-3)l/2
7.0
7.6
8.2
8.3
8.4
8.8
8.9
8.9
9.3
9.3
9.6
9.8
9.8
9.9
9.9
10.9
11.1
11.6
12.0
13.3
13.9
14.5
16.1
23.4
Compounds without filler
Peroxide
Low
DPPC
21
28
273
119
136
751
100
40
68
821
359
28
766
718
3.5
4.5
39
10
3.3
2.7
13
15
3.5
5.3
cure
High
DPOC
18
24
193
93
103
448
65
32
53
509
241
22
523
446
2.4
3.0
33
8.2
2.0
1.9
6.4
9.5
1.5
6.0
MgO cure
DOY-3C
23
32
137
96
105
291
74
37
58
326
190
25
280
263
3.3
9.2
52
37
14
12
25
50
2.3
15
CBb filled
compound with
MgO cure
DOZ-2C
12
15
Dd
64
89
D
52
25
36
D
205
17
D
D
1.5
1.8
30
11
2.3
1.3
8.9
8.9
0.6
3.0
aFormulations for these compounds are presented in Table 5.
b36.7% volume, carbon black.
cMatrecon identification code.
^D = dissolved or disintegrated.
In comparing the data between the
compounds of high and low peroxide con-
tents, the swell of the compounds with
the high peroxide is lower in practically
all cases, though not proportional to the
anticipated crosslink density. The gum
compound with the MgO cure varies con-
siderably from the peroxide cure with
respect to the swell in the different
solvents and may reflect the changes that
took place in the crosslinking with time.
The carbon black compound generally swells
less than the corresponding MgQ-cured gum
compound, due to the lower CSPE content
of the black compound. After Immersion
in those solvents that caused all of the
compounds without filler to swell con-
siderably, the carbon black-containing
compound actually swelled excessively and,
in some cases, disintegrated. These sol-
vents probably match very closely the
solubility parameter of the CSPE, which
will be determined when all swelling tests
have been completed. The solubility
parameters determined from the data on
each compound should be the same.
141
-------
TREATMENT OF SWELLING DATA TO OBTAIN
SOLUBILITY PARAMETERS OF FMLS
Several methods of treating volume
swell data to obtain component solubility
parameters of polymeric compositions have
been proposed (Hansen and Beerbower, 1971;
Garden and Teas, 1976; Van Krevelen and
Hoftyzer, 1976; Lyman et al, 1985). Since
the data being generated in this project
are not complete at this time, we have
only made preliminary determinations of
the individual component solubility
parameters of FMLs using two graphical
methods.
In the first method, we plot the
volume swell data for an individual
polymeric material after immersion in the
test solvents versus each of the four
solubility parameters, i.e. the three
component solubility parameters ($4,
6p, and 6fo) and the Hildebrand solu-
bility parameter (SQ), of each of the
test solvents to generate four separate
graphs. Enveloping curves that represent
the relationship between swell and a
solubility parameter were constructed by
hand on each of the graphs. The solu-
bility parameter value corresponding
to the highest point on these curves is
taken as the respective solubility
component for that polymer. By noting if
the greater swelling values occur in sol-
vents with similar values for a specific
solubility parameter, this method of
treating the swelling data can be used to
determine whether any single solubility
parameter has more or less importance in
determining the degree of interaction of
a polymer with solvents, i.e. the degree
of swell, than the other parameters. If
this is the case, it might be possible to
develop a very simple compatibility model
using only one waste solubility parameter
as an input. If this is not the case,
these plots will provide perspective.
Two examples of plotting the data in
this manner are presented in Figures 1 and
2 for HDPE-alloy (Liner No. 181) and a
polyester elastomer (Liner No. 323),
respectively, for swelling in 24 solvents.
These graphs show peaks for the Hildebrand
solubility parameter (6g) and the three
component solubility parameters (6
-------
component solubility parameters. The
results indicate that solvents with sol-
ubility parameters similar to a specific"
polymer should be avoided in sites lined
with compositions based on that polymer.'
We have also found promising the use of
an iterative computer process that fits
Lorenztian and Gaussian curves to data
for developing'the swelling versus solu-
bility parameter curves.
5
I
£ e
I
O ®
£ E
i
25
20
15
10
5
0
25
20
15
10
5
0
25
20
15,
10
5
0
25
20
15
10
5
0
'I I 1 Jf I O
Hildebrand
Parameter
«o
Hydrogen Bonding
• Component
Sfi
024 6 8 10 12,14 16 18 20 22 24
Value of Parameter
Figure 2. Volume .swell of a polyester
elastomer (Liner No. 323} as a
• function of .the. •Hildebrand.. and.
each, component solubility pa.-..:
. rameter for 24 solvents. The
respective, solubility parame-
ters for the liner are graph-
ically estimated based on the ,
results with 24 solvents (7
more wi 1 1 be added 1 ater as
the work is completed). . ,
Another method that has been sug-
gested by Garden and Teas (1976) uses
triangular plots of the component solu-
bility parameters which are converted
into fractions on the following basis:
Thus, each fraction is the square of the
respective solubility parameter divided
by the sum of the squares of the three
component solubility parameters. The
resulting three values serve as coordi-
nates for a point on the triangular plot.
The fractional component solubility pa-
rameters for the 31 solvents selected
for this study were plotted on a single
graph. Two examples of this graph are
presented in Figures 3 and 4 with volume
swell data for the HDPE-alloy (Liner No.
181} and the polyester elastomer (Liner
No. 323). For each figure, the equi-
librium swelling values for a single
polymeric material after immersion in 24
of the solvents were written next to the
respective points representing the three
component solubility parameters of each
solvent. Solvents, the parameters of
which fall in those areas of the plot
which cause high swelling values, have
solubility parameters close to that of
the respective polymer and should be
avoided in sites lined with compositions
based on that polymer. Organics in areas
that have low swell are likely to be
compatible with the FML in test. The
boundaries of the compatibility and non-
compatibility areas will be determined
more specifically when all of the data
have been generated.
DISCUSSION
In determining the component solu-
bility parameters of FMLs, we have dealt
primarily with pure liquids and their
effects on the properties of the various
FMLs. Such determinations are relevant
to secondary containment of hazardous
materials in tanks; however, in the case
of hazardous waste storage and disposal
facilities, the waste liquid or leachate
that is generated is generally a dilute
solution of organics in water. The re-
lationship between the solubility pa-
rameters of the FMLs and the absorption
of organics from aqueous solutions, i.e.
143
-------
leachates, must be quantified to aid in
developing a predictive model for liner-
waste compatibility.
.2,9,4.5.3,9
119.35.16
\
AAAAAAAAAA
10 20 30 40 50 60 70 80 80
Increasing Hydrogen Bonding »
fh
Figure 3. Volume swell of an HDPE-alloy
(Liner No. 181) in water and
in 23 organics having different
component solubility parameters
located on a triangular plot of
fractional solubility parame-
ters. The volume swell in
each solvent is shown next to
the location of the solvent.
Those seven points without
swelling data are for those
organics which are in the test
program and for which data are
being obtained.
We have reported (Haxo et al, 1985b)
on the swelling of a variety of polymeric
materials in a dilute aqueous solution of
a single organic, i.e. tributyl phosphate,
and have found the effect varied with the
lining material. Based on the preliminary
determination of the Hildebrand solubility
parameters of the FMLs, the degree of
swelling appears to be determined by the
relationship between the solubility pa-
rameters of the lining materials and the
Hildebrand parameter of the tributyl
phosphate.
10 20 30 40 50 60 70 80 90
Increasing Hydrogen Bonding »•
Figure 4. Volume swell of a polyester
elastomer (Liner No. 323) in
water and in 23 organics
having different component
solubility parameters located
on a triangular plot of frac-
tional solubility parameters.
The volume swel1 i n each
solvent is shown next to the
location of the solvent.
Those seven points without
swelling data are for those
organics which are in the test
program and for which data are
being obtained.
Work is now underway to assess the
swelling of FMLs in aqueous solutions of
mixtures of organics. Preliminary tests
have been performed on HOPE liners with
aqueous solutions of mixtures of acetone,
methyl ethyl ketone, 1,1,1-trichlorethane,
trichlorethylene, benzene, toluene, and
xylenes. Headspace chromatography (Haxo,
1983) was used to measure absorption of
the organics by the HOPE liner from the
aqueous solution in which it had been
immersed. Results show that both the
solubility of the organics in water and
the solubility parameters of the organics
with respect to that of the HOPE con-
tribute to absorption by the HOPE liner.
144
-------
Values ranging from 7.6 to 8.6 have
been reported for the solubility parame-
ters of HOPE. The analyses of the im-
mersed HOPE show that essentially no
acetone or methyl ethyl ketone were
absorbed by the liner. These two sol-
vents have high solubility in water and,
at the same time, high solubility pa-
rameters compared with the HOPE (Van
Krevelen and Hoftyzer, 1976). On the
other hand, the aromatic and chlorinated
solvents are absorbed. They have low
solubility in water, and are close in
solubility parameters to the HDPE.
CONCLUSIONS
Based on the results obtained to
date, it appears that additional immer-
sion testing is needed to establish
the individual solubility parameters.
Additional immersion tests on an increas-
ed number of organics will be performed
on a limited number of FMLs. We tenta-
tively conclude that the component -
solubility parameters will be useful" 1n
developing predictive models of the
effects of specific waste liquids
on specific lining materials. However,
since waste liquids or leachates from
hazardous waste storage and disposal
facilities are generally dilute solutions
of organics in water, partitioning of the
dissolved organics between water and
exposed liners needs to be thoroughly
investigated before any predictive
model can be developed. In the meantime,
the use of compatibility tests such as
the EPA Test Method 9090 should be con-
tinued in order to develop the necessary
database to confirm the effectiveness of
the solubility parameter concept for
predicting the compatibility of all
polymeric liner types.
REFERENCES
1. Barton, A.F.M., 1975. Solubility
Parameters. Chemical Reviews,
75:6, pp 731-753.
2. Garden, 0. L. and J. P. Teas, 1976.
Solubility Parameters. In; Treatise
on Coatings, Characterization of
Coatings: Physical Techniques. (R.
• Myers and J. S. Long, eds.). Marcel
Dekker, New York, Vol. 2: Part II,
pp 413-468.
Hansen, C. M. and A. Beerbower, 1971.
Solubility Parameters. In; Kirk-
Othmer Encyclopedia of Chemical Tech-
nology. 2nd Edition, Supp. Volume.
John Wiley and Sons, New York, pp
889-910.
Haxo, H. E., 1983. Analysis and
Fingerprinting of Unexposed and
Exposed Polymeric Membrane Liners.
In: Proceedings of the Ninth Annual
Research Symposium: LandDisposal,
Incineration, and Treatment of
Hazardous Haste. EPA-600/9-83-018.
U.S. EPA, Cincinnati, Ohio, pp
157-171.
Haxo, H. E., N. Nelson, and 0. A.
Miedema, 1985a. Solubility Parameters
for Predicting Membrane-Waste Liquid
Compatibility. In: Proceedings of the
Eleventh Annual Research Symposium:
Land Disposal ofHazardous Waste.
EPA/600/9-85/013. U.S. EPA, Cincin-
nati, Ohio, pp 198-212.
Haxo, H. E., R. S. Haxo, N. A.
Nelson, P. D. Haxo, R. M. White, and
S. Dakessian, 1985b. Final Report:
Liner Materials Exposed to Hazardous
and Toxic Sludges. EPA-600/2-84-169.
U.S. Environmental Protection Agency.
Cincinnati, Ohio, NTIS No. PB-85-
121333.
Haxo, H. E. and N. A. Nelson, 1984a.
Factors in the Durability of Polymeric
Membrane Liners. In: Proceedings of
the International Conference on Geo-
membranes, Denver, Colorado. Volume
II. Sponsored by Industrial Fabrics
Association International, St. Paul,
Minnesota, pp 287-292.
Lyman, W. J., A. D. Schwope, J. P.
Tratnyek, and Peter P. Costas, 1985.
Solubility Parameters for Predicting
Membrane-Waste Liquid Compatibility.
In: Proceedings of the Eleventh Annual
Research Symposium: Land Disposal of
Hazardous Haste.EPA/600/9-85/013.
U.S. EPA, Cincinnati, Ohio, pp 283-
295.
Van Krevelen, D. W. and P. 0. Hoftyzer,
1976. Properties of Polymers—Their
Estimation and Correlation with Chem-
ical Structure. Elsevier Scientific
Publishing Co., Amsterdam.
145
-------
FIELD DEMONSTRATION OP IN SITU BIOLOGICAL TREATMENT
OF CONTAMINATED GROUNDWATBR AND SOILS
Roger S« Wetzel, Donald H. Davidson, Connie M.. Durst, Douglas J. Sarno
Science Applications International Corporation
McLean, Virginia 22102
ABSTRACT
In situ biological degradation of organic contaminants was demonstrated at a waste
disposal site at Kelly Mr Force Base (AFB), Texas. A s.mall-scale soil and groundwater
treatment system was used to circulate groundwater by pumping and gravity injection.
Oxygen, in the form of stabilized hydrogen peroxide, and specially formulated nutrients
were introduced to the subsurface in order to stimulate nticrobial degradation of the
organic contaminants. Extensive background studies, were performed to characterize the
site geology and hydrology, determine the contaminant profile,.and demonstrate treat-
ability in the laboratory. This project was the first- field application of in situ bio-
logical treatment at a site contaminated with a complex mixture of both organic'and
inorganic wastes. «». : •
The injection/extraction treatment system was operated for a period of eight months
during which hydrogen peroxide was added to the subsurface. Nutrients were added for a
period of approximately five months. Routine monitoring of the soil and groundwater was
conducted to provide data for evaluating the system's performance and controlling its
operation. Decreases in hydrocarbon compounds and chlorobenzenes were observed in the
contaminated groundwater over the operating period. Similar results were shown in
laboratory treatability studies.
Problems associated with operation of an in situ biological treatment system and
useful monitoring parameters are discussed. The demonstration project at Kelly AFB has
provided a better understanding of the capabilities and limitations of the technology
for treatment of organic contaminants in soils and groundwater.
INTRODUCTION
In situ biological treatment of soils
and groxindwater contaminated with organic
compounds is based on stimulating the in-
digenous subsurface microbial population
to degrade the organic contaminants.
Conditions for contaminant biodegradation
are optimized by providing the nutrients
and oxygen which may be limiting factors
for the growth of aerobic microbes in the
subsurface.
In situ treatment offers advantages
over conventional methods such as "pump
and treat" technologies which pump
groundwater to the surface for treatment.
Because the active treatment zone is in
the subsurface, in situ biological treat-
ment has .the potential to remove contami-
nants sorbed to the soil matrix, in addi-
tion to treating the contaminated ground-
water. Pump-and-treat methods treat only
the groundwater, allowing for clean
groundwater to become contaminated as de-
sorption of the pollutants occurs when
the groundwater contacts the untreated
soil. In situ biological treatment is
also more desirable environmentally than
excavation, removal, and disposal of con-
taminated soils which merely transfers
contaminants to a more secure disposal
area without treatment. In situ bio-
logical treatment offers treatment of
organic-contaminated soils and ground-
water and is often less expensive than
146
-------
conventional treatment or disposal
methods.
A field demonstration of the tech-
nology was conducted at a waste disposal
site at Kelly AFB, Texas. The site was
originally used as a disposal pit for
chromium sludges and other electroplating
wastes. Prior to being closed in 1966,
it had been used as a chemical evapora-
tion pit for chlorinated solvents, ere-
sols, chlorobenzenes, and waste oils.
The Kelly AFB waste site was selected
for demonstration of in situ biological
treatment for a number of reasons. The
subsurface contained biodegradable.or-
ganics and a highly adaptive and sub-
stantial population of microbes. The
perched aquifer present was not used as
a supply of drinking water and ground-
water temperatures were near optimum for
biological -treatment. The project at
Kelly APB was the first field application
of the technology at a site contaminated
with a complex mixture of both organic
and inorganic wastes.
BACKGROUND
Background studies were conducted
prior to design and implementation of the
treatment system at Kelly AFB. These
studies included site hydrologic and geo-
logic characterization, identification of
the contaminants present, enumeration of
the microbial population, laboratory
treatability studies, and an examination
of the effects of nutrient and hydrogen
peroxide addition on soil permeability.
Site Location and Geology
The Kelly APB demonstration site
(Figure 1) encompasses a 2800 ft^ area
situated on an elevated peninsula-like
landform that is surrounded on three
sides by surface drainage channels. The
demonstration site is shown in Figure 1
at the location of the pumping and in-
jection wells and represents the zone
where treatment was conducted. Three
monitoring wells are located outside the
demonstration area (Figure 1). Leon
Creek borders the site on the west and
an unnamed intermittent surface drainage
ditch borders the site on the east and
south. The storage yard, depicted in
Figure 1, represents the location of the
former waste disposal area at Kelly AFB.
The site geology is highly variable
and was difficult to characterize. Non-
homogeneous deposits near the surface are
comprised of gravels and sands in a silt
and clay matrix. Soils below the aquifer
consist of clay and marl, characteristic
of the Havarro Formation. A perched
aquifer is present which fluctuates
seasonally and ranges from 4 to 8 feet
in thickness (4). Groundwater movement
in the saturated zone is directed toward
Leon Creek at a slow rate because of the
predominance of clay in the alluvial
materials which are present. Fine-
grained soils and sludges are interlayer-
ed with gravel lenses, producing highly
variable rates of groundwater movement
within the demonstration site. Hydraulic
conductivity of the perched aquifer with-
in the treatment zone ranges from 0.11 to
9,26 ft/day (3).
Contaminant Profile
The aquifer is contaminated by a wide
variety of organic and inorganic com-
pounds. Complete characterization of the
contaminant profile indicated the pre-
sence of 1,2-dichlorobenzene and 1,4-
dichlorobenzene at concentrations greater
than 10 parts per million (ppm) in the
soil samples. Total hydrocarbons were
found in soil sampled from the saturated
zone of Borehole 1 (Figure 1} at a con-
centration of approximately 200 ppm.
Borehole 1 is located within the area of
the treatment system. Tetrachloroethy-
lene (PCE), trichloroethylene (TCE),
cis-1,2-dichloroethylene, and total
hydrocarbons were found in the ground-
water at concentrations greater than 1
ppm. Antimony, chromium, lead, nickel,
silver, thallium, and zinc were detected
at concentrations greater than 10 ppm in
soil samples taken from the demonstration
site. Concentrations of metal ions pre-
sent in groundwater samples were low and
did not exceed 1 ppm.
Microbiological Studies
Microbiological studies were also con-
ducted on subsurface samples collected
from the site. Direct and viable cell
counts were performed by Memphis State
University (1). Direct plate counts
ranged from 7.6 x 10^ to 1.7 x 108
cells/gram of sample and viable cell
counts ranged from less than 100 to
7 x 106 cells/gram of sample. Seven
147
-------
6A
AAD
11 A|
BBD
A4
A5
,A3
Leon
Creak
V
\
a
A
Monitoring Well
Borehole
ceo
A12
P2
P9,
M2Q
O Injection Wall
• Pumping Well
• — — Storaga Yari COOVBIS former waste site!
10 A
0 50
S=ESS
Scale iFeet)
Unnamed
Drainage
Dhch
Figure 1. Site Plan and Location of Treatment System.
148
-------
types of substrate media were used for
enumerating bacteria and all media
yielded similar numbers of cells for any
one sample. This lack of selectivity
observed between media types indicated
the presence of highly adaptive bacteria
populations at the site.
Laboratory Biodegradation Study
laboratory biodegradation studies were
performed to investigate the potential of
in situ biodegradation of the organic
contaminants by indigenous bacteria (1).
Closed microcosms of 240. milliliter (ml)
volumes were prepared with soil and
groundwater sampled from the site. Aer-
obic microcosms were amended with nutri-
ents and either hydrogen peroxide or oxy-
gen. Anaerobic microcosms were also pre-
pared in addition to aerobic and anaer-
obic sterilized controls.
Three microcosms were selected in-
itially (day 0) and again at 25, 50, and
100 days after the study commenced, and
samples were analyzed for volatile or-
ganic hydrocarbons and base/neutral and
acid extractable organics. Results from
the biodegradation study indicated that
significant degradation of aliphatic
hydrocarbons (n-alkanes) and halogenated
aromatics such as chlorobenaene occurred
when oxygen and nutrients were supplied
to aerobic microcosms of contaminated
soil and groundwater from the Kelly &FB
site. Results from the anaerobic micro-
cosm studies showed degradation of
chlorinated aliphatic compounds such as
tetrachloroethylene, trichloroethylene,
and trans-1,2-dichloroethylene (1).
Treatment System Configuration
The treatment system that was de-
signed, installed, and operated at the
site is shown in Figure 1. It consisted
of nine extraction wells and four injec-
tion wells arranged in a grid-like pat-
tern within a 60-foot diameter circular
area. The extraction wells pumped
groundwater to a central surge tank;
the water was then released at a con-
trolled rate to a distribution box.
Nutrients and peroxide were pumped at
controlled rates into the line between
the surge tank and distribution box. An
in-line baffle pipe provided mixing of
the nutrients and peroxide with the
groundwater. The flow was then directed
from the distribution box to each of the
four injection wells and injected by
gravity. One upgradient and two down-
gradient monitoring wells were used to
determine the effects of the system out-
side the treatment zone.
SYSTEM OPERATION AND MONITORING
Construction of the treatment system
began in April 1985. Development of the
pumping and injection wells took place in
May 1985, followed by start-up of the
system in early June. Groundwater was
circulated for approximately two weeks
prior to nutrient addition. Nutrient
addition began in mid-June, followed by
hydrogen peroxide addition in late June.
The nutrient solution used was Restore™
375K Microbial Nutrient, supplied by FMC
Aquifer Remediation Systems, Princeton,
New Jersey. A 35 percent hydrogen per-
oxide solution, Restore™ 105 Microbial
Nutrient {FMC Aquifer Remediation Sys-
tems ), was added as the source of oxygen.
Hydrogen peroxide was selected as the
source of oxygen because it can provide
approximately five times more oxygen to
the subsurface than aeration techniques.
Other in situ bioreclamation projects
have been conducted for treatment of
gasoline-contaminated aquifers, using
hydrogen peroxide as a source of oxygen
(5). Hydrogen peroxide can be toxic to
bacteria at high concentrations; however,
studies indicate that it can be added to
soil or groundwater systems at concentra-
tions up to 100 ppm H202 without being
toxic to microbial populations (2). Con-
centrations as high as 1000 ppm l^Oj can
be added to microbial populations without
toxic effects if the proper acclimation
period is provided for the bacteria (2).
Hydrogen peroxide was initially added
at the Kelly APB demonstration project to
maintain a concentration of 100 ppm I^Og
in the groundwater. The concentration
was increased by increments of 100 ppm
every two weeks in order to acclimate the
bacteria to the hydrogen peroxide solu-
tion. The concentration was maintained
thereafter, at 500 ppm H2O2» until system
operations ceased in late February 1986.
Soil and groundwater samples were col-
lected from the demonstration site prior
to system construction in order to deter-
mine initial levels of microbial popula-
149
-------
tiona and contaminants present. Back-
ground levels of groundwater chemical
parameters listed in Table 1 were also
determined in order to establish a base-
line for evaluating changes in the
groundwater chemistry.
ft detailed monitoring schedule was im-
plemented to routinely sample groundwater
from the wells in the system and soil
from the treatment zone in order to moni-
tor the performance and effectiveness of
'the system. The groundwater monitoring
schedule is outlined in Table 1. The
groundwater and soil sampling program was
implemented prior to start-up of the sys-
tem in Hay 1985 and continued throughout
the project.
Eroblems Associated with System Operation
A variety of problems were encountered
during the period the system was in oper-
ation. The treatment system was origi-
nally designed to operate continuously
and provide uniform pumping of ground-
water and injection of nutrient and hy-
drogen peroxide solutions to the aquifer.
Following construction and development of
the injection and extraction wells, slug
tests were performed to determine the hy-
draulic conductivity and expected dis-
charge of each extraction well. Results
from the slug tests indicated that the
hydraulic conductivity of the aquifer
ranged from 0.11 ft/day to 9.26 ft/day
(3). Due to the variable soil perme-
ability, uniform pumping and injection
rates could not be maintained throughout
all areas of the demonstration site. In
an attempt to reduce this problem, pump-
ing rates were lowered to maintain a con-
stant injection rate without flooding of
the injection wells. In addition, pump-
ing was reduced to half of the pumps
working only a portion of the day on a
rotational basis. The decrease in
groundwater circulation rate observed
following system start-up is shown in
Figure 2.
fiquifer heterogeneity is one of the
problems inherent to many contaminated
waste sites. Therefore, system design,
well placement, and equipment selection,
such as pumps and flow meters, should in-
clude enough flexibility in order to
allow for adjustments in the pumping and
injection scheme and still maintain hy-
draulic isolation of the treatment area.
This was important at the Kelly ftFB site
in order to avoid migration of untreated
contaminants downgradient of the site.
Another problem encountered during
system operation was chemical precipita-
tion which occurred following system
start-up. & thick, white precipitate
of calcium phosphate developed in the
distribution box and injection wells.
Addition of nutrients was changed to a
batch mode and the concentration of
nutrient addition was changed in order to
reduce the precipitation problem. Efforts
were taken to rehabilitate the injection
wells, including manual brushing of the
well screens and air surging, followed by
pumping and bailing. These activities
removed a large amount of the precipi-
tate. Injection capacity normally in-
creased immediately following these
operations but returned to the slower
rate shortly thereafter.
Grpundwater Monitoring
Groundwater monitoring was a critical
component of the treatment system opera-
tion at Kelly AFB. The groundwater moni-
toring schedule used at Kelly &FB is
shown in Table 1. Routine analysis of
groundwater chloride, ammonium, and phos-
phate ion concentrations.was important
for determining if nutrients were suc-
cessfully transported through the sub-
surface and made available for microbial
growth. Regular monitoring of these
parameters indicated that nutrient trans-
port was variable throughout the site due
to contrasts in the hydraulic conduc-
tivity of the aquifer.
Modifications in system operation were
made during the project, based on ground-
water monitoring results, in order to
promote transport of nutrients to the
areas which had been left untreated.
Groundwater was pumped from, and in-
jected into, only those wells which had
not shown evidence of nutrient transport
in order to enhance microbial degradation
and treatment of the contaminants within
those areas.
Carbon dioxide (C02> concentrations
were measured in groundwater samples from
each of the wells as a means of monitor-
ing biological activity and system per-
formance. C»2 is an end product of
150
-------
TABLE 1. GR008DWATER MONITORING SCHEDULE
Parameter
Temperature
Conduct! vi ty
pH
Dissolved Oxygen
Carbon Dioxide
Ammonia
Phosphate
Chloride
Hydrogen Peroxide
Nitrate
Sulf ate
Acidity
Alkalinity
Total Hardness
Chromium
Lead
Extraction
Wells
2/Week
2/Week
2/Week
2/Week
2/Week
2/Week
2/Week
2/Week
Weekly
2/Month
2/Month
2/Month
2/Month
2/Month
Monthly
Monthly
Sampling Frequency
Injection
Wells
2/Week
2/Week
2/Week
2/Week
2/Week
Weekly
Weekly
Weekly
Weekly
2/Month
2/Month
2/Month
2/Honth
2/«onth
_ —
Monitoring
Wells
Weekly
Weekly
Weekly
Weekly
Weekly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Oil and Grease
Ttttal Hydrocarbons
(Alkanes)
Priority Pollutants
(Volatile Organics
and Metals)
Microbial Plate
Counts
Groundwater
Elevations
Monthly Monthly Monthly
Quarterly Analysis of 10 Wells
Quarterly Analysis of 10 Wells
Monthly
Daily
Monthly
Daily
Monthly
Daily
151
-------
ndcrobial degradation and chemical oxida-
tion of organics. Changes in groundwater
COj values can indicate changes in the
activity of the microbial population.
and nutrient breakthrough was high and
suggests an area where biological degra-
dation of organic compounds would be ex-
pected to be greatest.
In order to facilitate analyses of
system performance, CO2 levels were
measured in the groundwater samples and
are shown in Figure 3. Wells were
grouped according to well type and the
relative magnitude of CC>2 concentrations
detected* Background concentrations of
CC>2 were determined in the upgradient
monitoring well (M-1) and varied from 5
to 40 ppm. Levels of CC>2 in M-1 were
consistently lower than those detected in
the pumping wells, injection wells, and
downgradient monitoring wells. Pumping
well P-9 showed the highest levels of COj
within the system; weekly averages ranged
from 160 to 210 ppm COj. Pumping well
P-9 corresponds to an area where chloride
COKCLOSIONS
Preliminary laboratory studies indi-
cated that organic compounds at Kelly AFB
could be treated by in situ biological
degradation. Laboratory results showed
biological degradation of aliphatic
hydrocarbons and chlorinated aromatics.
Microbiological investigations confirmed
the presence of an adaptive microbial
population.
Field investigations indicated that
the limiting factors for successful site
remediation were the heterogeneous nature
of the subsurface and the low hydraulic
conductivity of the aquifer. Declining
2000-
"D
Q.
3
11500-
£
I
O
1
•o
c
1
O
a>
tB
s
I
1000-
500-
10 15 20
Weeks of Operation
25
30
250
200-
150-
E
Q.
§ 100-
c
0)
o
c
o
U
50-
B AVG M2 and CC Q
A AVG PI, P2, P3, P6, P9 O
A AVG P4, P6, P7, P8 *
P-9
Ml
AVG Injection Wells
20
25 30
Weeks of Operation
35
Figure 2. average Groundwater Circu-
lation Rates for Kelly AFB
Treatment System.
Figure 3. Kelly RFB Treatment System
Groundwater Carbon Dioxide
Concentrations.
152
-------
groundwater recirculation rates were
attributed to the stratified nature of
the subsurface and chemical precipitation
of calcium phosphate from the treated
groundwater. These factors resulted in
slow transport of nutrients to some areas
of the demonstration site.
Modifications can be made in the
treatment system pumping and injection
patterns in order to maximize transport
of treatment solutions and promote bio-
degradation of organic contaminants in
the less permeable areas of a treatment
site. System design, well placement, and
equipment selection are important factors
for achieving operational flexibility of
the treatment system in a heterogeneous
geologic setting.
Another important aspect of operating
an in situ biodegradation treatment sys-
tem is the development of a comprehensive
monitoring program to assess system per-
formance and treatment effectiveness.
Frequent on-site monitoring of chemical
parameters such as nutrients, chloride,
carbon dioxide, dissolved oxygen, and pH
are useful to the system operator for
determining concentrations and volumes of
treatment chemicals to be injected.
Groundwater monitoring is also critical
for determining if modifications in pump-
ing and injection patterns are required.
Results from on-site groundwater moni-
toring showed transport of nutrients in
areas of high soil permeability. High
groundwater carbon dioxide concentrations
corresponded to those areas of the demon-
stration site where maximum transport of
nutrients occurred. These results indi-
cate that microbial degradation of or-
ganic compounds can be expected and that
demonstration of the technology has been
successful.
The pilot project at Kelly AFB has
provided useful information about opera-
tion of an in situ biological treatment
system in a heterogeneous geologic set-
ting. A better understanding of the
capabilities and limitations of the
technology have resulted from this
project.
ACKNOWLEDGEMENTS
This effort was jointly sponsored and
funded by the U.S. Environmental Protec-
tion ftgency and the Air Force Engineering.
and Services Center. The authors wish to
acknowledge the assistance of Stephen C.
James, U.S. Environmental Protection
Agency, and Captain Edward Heyse, U.S.
Air Force.
REFERENCES
1. Science Applications International
Corporation, 1985. Interim Report:
Field Demonstration of In Situ Bio-
logical Degradation. For Air Force
Engineering and Services Center, Tyn-
dall AFB, Florida and U.S. EPA Haz-
ardous Waste Engineering Research
Laboratory, Cincinnati, Ohio.
2. Texas Research Institute, Inc., 1982.
Final Report: Enhancing the Microbial
Degradation of Underground Gasoline by
Increasing Available Oxygen. For the
American Petroleum Institute, Washing-
ton, D.C.
3. Wefczel, R. S., C. M. Durst, D. J.
Sarno, P. A. Spooner, S. C. James, and
B. Heyse, 1985. "Demonstration of In
Situ Biological Degradation of Contam-
inated Groundwater and Soils." Pro-
ceedings of the Sixth National Con-
ference on Management of Uncontrolled
Hazardous Waste Sites. Hazardous
Materials Control Research Institute,
Washington, D.C., pp. 234-238.
4. Wetzel, R. S., S. M. Henry, P, A.
Spooner, S. C. James, and E. Heyse,
1985. "In Situ Treatment of Contami-
nated Groundwater and Soils, Kelly Air
Force Base, Texas." Proceedings of the
Eleventh Annual Research Symposium on
Land Disposal, Remedial Action, Incin-
eration, and Treatment of Hazardous
Waste. U.S. Environmental Protection
Agency, Cincinnati, Ohio, pp. 97-106.
5. Yaniga, P.M. and W. Smith, 1984.
"Aquifer Restoration via Accelerated
In Situ Biodegradation of Organic Con-
taminants." Proceedings of the NWWA/
API Conference on Petroleum Hydrocar-
bons and Organic Chemicals in Ground
Water-Prevention, Detection and Re-
storation. Houston, Texas, pp. 451-
472.
153
-------
STRINGFELLOW LEACHATE TREATMENT WITH RBC
E. J. Opatken, H. K. Howard, and J. J. Bond
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
A study is underway to determine if a rotating biological contactor (RBC) can be
effectively employed to treat leachate from a hazardous waste site. In this case a
pilot sized RBC was installed at the U.S. EPA's Testing and Evaluation (T & E)
facility in Cincinnati. The RBC was scaled down from full size to 1/10 of standard.
Prior to the leachate studies a series of kinetic runs will be made on primary
effluent from the Cincinnati Mill Creek Sewage Treatment Plant to establish the
capability of the RBC to convert soluble organic compounds into innocuous products.
Nine runs are scheduled with Stringfellow leachate to determine:
° The rate of organic removal/conversion
0 The efficiency of organic removal/conversion
0 The effect of increased speed of rotation on organic removal
0 The reaction kinetics.
This paper reports on the results from these experiments and the effectiveness of
an RBC to adequately treat leachate from a hazardous waste site.
INTRODUCTION
Stringfellow is a hazardous waste
site that is located in Glen Avon,
California which is near Riverside,
CA. Leachate is generated at an
approximate rate of 20,000 gallons/
day and contains high concentrations
of metals and organics. A leachate
treatment facility is operating at
the site that consists of lime treat-
ment, followed by clarification, sand
filtration, and granular carbon treat-
ment. The effluent from the carbon beds
is trucked to an Orange County inter-
ceptor sewer located about 12 miles from
the site, for disposal and treatment by
the Orange County wastewater treatment
facility.
Project Description
154
-------
A pilot sized RBC was installed at
the EPA Testing and Evaluation Facility
(T & E) in Cincinnati for testing the
biochemical treatability of Stringfellow
leachate. The pilot sized RBC contains
11,000 square feet of surface area or
approximately 10% of the size of a full
scale RBC. The diameter of the pilot
unit is identical to the full scale
RBC, 12 feet. The length is less than
3 feet; whereas the full scale RBC is
25 feet. The Cincinnati treatment
plant's primary effluent (PE) was used
to develop a biological population and
to obtain basic kinetic data. The RBC
Facility was designed to operate in a
batch mode with leachate that was
trucked from California to Cincinnati
for experimentation. The experiments
were designed to operate in a batch mode
for the following reasons:
0 eliminate flow controls
0 minimize spillages
0 minimize accidental releases into
the T & E sewer system
0 improve mass balance analyses
0 obtain reaction kinetics data
0 control final disposal
0 avoid overloading of the shaft
0 can be directly scaled for the
Stringfellow Site
Objective
The prime objective of this project
is to determine whether the Stringfellow
leachate can be cost effectively con-
verted into an innocuous waste by
biochemical treatment with a rotating
biological contactor (RBC).
Methodology
The first few runs are being made
with increasing ratios of leachate/
primary effluent to allow the biomass to
acclimate. When 100% leachate is
reached, the methodology will be to
pump 1000 gallons of Stringfellow
leachate from the storage tank to a
1000-gallon calibrated stainless steel
tank. The volume of leachate is
determined and then pumped into the
RBC. The batch experiment is designed
to operate for 24 hours at a constant
speed of 1.5 rpm, unless the disappear-
ance of Dissolved Organic Carbon (DOC)
is unsatisfactory, then the treatment
will continue until the quality is
within the limits established by the
Metropolitan Sewer District (MSD) and
the Contingency Plan.
When the DOC drops below 30 mg/L,
the treated leachate is then clarified
and stored in another calibrated tank.
A chemical analysis is made to determine
whether the leachate is within the
limits prescribed before draining the
leachate into the T & E sewer system.
The parameters that are monitored during
an experimental run are:
1. Volume of leachate treated in the
RBC (initial & final volumes)
2. Temperature (every 2 hours)
3. Speed of rotation (every 6 hours)
4. Visual comments on thickness of
biomass on the discs (every
2 hours)
5. Dissolved oxygen (every 2 hours)
6. pH (every 2 hours)
Kinetic Parameters
During the experiment, samples are
obtained from the RBC tank every hour
for the first 12 hours. During the
final 12 hours samples are withdrawn
every 2 hours to obtain kinetic data on
the disappearance of the soluble gross
organics. The soluble material is the
liquid phase that passes through a
Uhatman 934AH filter. The operational
control used during the experiment is
to determine the Dissolved Organic
Carbon (DOC) on site and obtain informa-
tion on the progress of the reaction.
155
-------
Soluble Gross Organic Analyses
0 Biochemical Oxygen Demand, SBOD, mg/L
0 Chemical Oxygen Demand, SCOD, mg/L
0 Organic Carbon, DOC, mg/L
Performance Analyses
The chemical analyses required
for characterizing the waste influent
and effluent consisted of three samples
per batch. The samples were obtained
on the influent to the RBC, the effluent
following treatment by the RBC, and the
effluent following clarification. A
sludge sample from the bottom of the
clarifier was also obtained to estimate
the fate of the specific organics. The
characterization involved the following
analyses.
Conventional wastewater analyses
Total Gross Organics
BOD, mg/L
COD, mg/L
TOC, mg/L
Total Suspended Solids, TSS, mg/L
Volatile Suspended Solids, VSS, mg/L
Ammonia-Nitrogen, NH^-N, mg/L
Nitrite, Nitrate-Nitrogen, N02-N03-N,
mg/L
Total Kjeldahl Nitrogen, TKN-N, mg/L
Total Phosphorus, mg/L
Major Organic Contaminants
Para-chlorobenzene sulfonic acid
1, 2 Dichlorobenzene
o-Xylene
Chloroform
Ethylbenzene
2-Hexanone
Tetrachloroethylene
Metals analyses were required on
the influent to the RBC to confirm the
effectiveness of the lime treatment
operation conducted at the Stringfellow
site and to comply with the pretreatment
levels required by the Metropolitan
Sewer District of Greater Cincinnati.
The metal analyses for Stringfellow
leachate are:
Metals
Arsenic
Cadmium
Chromium
Copper
Cyanide
Lead
Mercury
Nickel
Silver
Zinc
Biomass Buildup
The T & E Facility has available
primary effluent (PE) from the Mill
Creek Treatment Plant, which is part
of the Metropolitan Sewer District of
Greater Cincinnati (MSDGC). The PE
contains a high concentration of soluble
organics, at times reaching 180 mg/L of
SBOD. This soluble organic concentration
is more than twice the normal concentra-
tion of a municipal treatment plant that
contains little, if any, industrial
contribution. The Mill Creek raw waste
water is estimated to have a 50%
industrial portion which accounts for the
high organic concentration.
The PE was used, to develop a bio-
logical growth on the RBC discs.
Approximately 1000 gallons of PE were
pumped into the RBC from the mix tank
and were treated for 6 hours. The con-
tents were then drained and the RBC
refilled and the process repeated until
growth of biomass was seen. Samples
were then taken of the influent and
effluent during the day shift and
analyzed for
156
-------
0 SBOD
0 SCOD
0 DOC
After several days kinetic data were
gathered by taking samples for soluble
organic analyses at the 0.5, 1,0, 2.0,
4.0 and 6.0 hours reaction period. This
was then followed by two more runs with
PE based on the procedure that was
developed for Stringfellow leachate to
test the effectiveness of the operating
procedure.
Stri_ngfellpw Treatment
Batch I 1
After the buildup of biomass the RBC
began its first run with Stringfellow
leachate. The leachate was combined
with PE and consisted of 2 parts treated
PE to 1 part leachate to avoid shocking
the biomass.
The DOC was monitored every hour
from the start of the experiment.
After 6 hours there was no significant
drop in DOC and the sampling frequency
was changed to once per day. The DOC
results are shown in Table I, along with
pH.
Table I shows that no significant
drop in DOC was evident from the start
of the experiment until the 7th day,
when a drop of 41 mg/1 occurred in the
DOC along with a 0.9 drop in pH. During
the next several days, 3.8% caustic was
added in 200 ml increments to raise the
pH. After 13 days it appeared that the
DOC level had stabilized and the experi-
ment was concluded by clarifying the
RBC effluent. The reductions experienced
during the 1st experiment are shown in
Table II.
The rotational speed was reduced by
0.1 rpm each day after the first full day
of operation until a final speed of 1.0
rpm was obtained on the RBC. The opera-
tional change was made to prevent
stripping the biomass from the disc.
Table I shows that 3 samples were
analyzed 3/26/86. The first sample was
unfiltered and the 2nd and 3rd samples
were filtered through 0.45 u millipore-
filter. The Table shows that the
filtered sample was lower than the
unfiltered sample by 13 mg/L. To
indicate the filtration of the sample, an
F follows the date signifying that the
sample was filtered thru a 0.45 u
membrane filter. On April 3rd the filter
was changed from the 0.45 u membrane
filter to a Whatman 934AH because of
indicated organic contamination from the
0.45 u filter. The effect is shown on
4/3/86 as a drop of 8 mg/L when Whatman
934AH was used. The letters FP follow-
ing the sample date indicate that
filtration was with a Whatman 934AH
filter paper.
157
-------
TABLE I. DOC CONCENTRATION WITH TIME
Date
3/25/86
3/25/86
3/25/86
3/25/86
3/25/86
3/25/86
3/26/86
3/26/86F
3/26/86F
3/27/86F
3/28/86F
3/29/86F
3/30/86F
3/31/86F
4/1/86F
4/2/86F
4/3/86F
4/3/86FP
4/4/86FP
4/5/86FP
4/6/86FP
4/7/86FP
Time
(hour)
0840
0940
1040
1140
1240
1440
0840
0840
1140
0840
0840
0840
0840
0840
0840
0840
0840
0840
0840
0840
0840
0840
Retention
(hours)
0
1
2
3
4
6
(days)
1
1
1.1
2
3
4
5
6
7
8
9
9
10
11
12
13
DOC
(rag/L)
130
123
125
124
130
130
115
102
104
106
109
116
116
108
67
59
59
51
52
53
53
56
PH
7.4
* * *
7.7
• • *
7.8
8.0
7.9
• » *
7.8
7.7
7.5
7.5
7.2
6.7
5.8*
•5.9*
5.7*
• » *
5.6*
5.7*
6.5
6.1
* Caustic added to RBC contents
158
-------
TABLE II. REMOVALS FROM EXPERIMENT S-l
SBOD mg/L
SCOD mg/L
DOC mg/L
TBOD mg/L
TCOD mg/L
TOC mg/L
TSS mg/L
VSS mg/L
TKN-N, mg/L
Influent
82,
360,
116.
160.
360.
140.
98.
52.
Eff1uent
4.0
156.
53.
198.
680.
78.
520.
400.
8.0
Clarified
21
259
Organics
MeCl2, ug/L 159.
CHC13, ug/L 30.
1, 1, 1 Trichloroethane, ug/L 211.
Tetrachloroethane, ug/L n. d,
1, 2 Dichlorobenzene, ug/L 74.
o-Xylene, ug/L <10.
Ethyl benzene, ug/L <10.
p-Chlorobenzene sulfonic acid, mg/L 196.
Vapor Space Organics* •••
Total Organic Halogen, mg/L •••
n. d.
n. d.
n. d,
n. d.
n. d.
* * •
n. d.
n. d.
26
*A Metropolitan Sewer District of Greater Cincinnati in-house method.
These reductions were satisfactory
for disposal to the T & E sewer with the
exception of Total Organic Halogen. The
limit was set at 5 mg/L by the MSD which
is well below the reported level of
26 mg/L. The plans are to reprocess the
batch in the RBC to lower the total
organic halogens to an acceptable level.
Batch S-2
The 2nd experimental batch was made
by blending 1 part of Stringfellow
leachate with 1 part from Batch S-l
(2 parts PE/1 part Stringfellow). Again
the contents were measured by volume in
the mix tank, stirred, and fed to the
RBC. The RBC operated at a speed of 1.5
rpra. The disappearance of DOC was again
monitored to follow the reaction. The
results are shown in Table III.
159
-------
TABLE III. DISAPPEARANCE OF DOC WITH TIME
Reaction Time
(days)
0.
.04 (1 hour)
.08 (2 hours)
.12 (3 hours)
1.
2.
2.2
3.
4.
5.
6.
DOC
(mg/L)
176 •
174
173
177
169
144
132
93
89
88
88
DO
(mg/L)
7.4
• * •
7.4
8.0*
7.9
7.7
7.8
7.9
7.3
8.6
&
6.8
* » »
7.3
* * *
7.3
6.8**
6.4
6.7
6.9
6.7
6.3
*Reduced RBC speed to 1.0 rpm
**Started 3.8% caustic addition
The reduction in DOC started after
the second day rather than the 7th day
as it occurred in the first batch. The
other removal data are not available
at this time. The third batch is
currently being treated and the drop in
DOC started after 1 day of operation.
The blend for batch 3 consisted of 3
parts leachate to 1 part treated
effluent from batch 2.
The results are encouraging and give
indications that Stringfellow leachate
can be treated by an RBC. Further
experimentation is required to reduce the
reaction time to less than 12 hours and
to improve the removal efficiencies.
160
-------
NONDESTRUCTIVE TESTING (NOT) LOCATION OF CONTAINERS BURIED IN SOIL
Arthur E. Lord, Jr. and Robert M, Koerner
Drexel University
Philadelphia, PA 19104
ABSTRACT
This paper describes work involving the detection and delineation of buried steel and
plastic containers using a wide variety of NOT (remote-sensing) techniques. Seventeen
techniques were considered and evaluated, and the four most promising were used in the
work to be described herein. These four are:
. electromagnetic induction (EMI)
, metal detection (MD)
. magnetometer (MAG)
. ground penetrating radar (GPR).
The containers, varying in size from 5-gal to 55-gal were buried in known distributions in
a wide variety of soils, and also some were submerged under water. Five sites were used:
. Site One - a dry sandy soil with little man-made interference
. Site Two - a saturated silty clay soil with much metal interference nearby
Site Three - a river of varying electrical conductivity (containers submerged)
. Site Four - a sandy soil of varying electrical conductivity (steel containers)
Site Five - a sandy soil of varying electrical conductivity (plastic containers)
As a result of the work at the five sites, a relatively complete picture has emerged
concerning the strengths and weaknesses of the four major NOT subsurface container loca-
tion techniques. This paper will describe the individual NOT techniques, the experimental
work involved and will give a summary of the findings.
Briefly, it can be stated:
. GPR is the only reliable method to detect plastic containers, but it has limita-
tions.
GPR, EMI and MD all suffer severe loss of detection ability when the background
electrical conductivity exceeds 40 mi Himhos/meter.
In a dry sandy soil EMI, GPR and MAG are all capable of picking up a single 55-gal
steel drum to a depth of at least 10 feet.
. The MAG method works well for steel under all subsurface conditions.
. GPR can usually pickup the side walls of the excavations where waste is dumped.
INTRODUCTION
The investigation of subsurface ob-
jects can be approached in two very differ-
ent 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 or sub-
sequent testing, such methods have certain
drawbacks in identifying and locating
buried containers. Some disadvantages of
these destructive test approaches are:
. The information obtained is discon-
tinuous over the area investigated.
161
-------
. Permission to enter the properties
in question and, especially to ex-
cavate therein, may be troublesome
or impossible to obtain.
. Access for excavation equipment may
be very difficult at the site in
question.
. Costs are generally high, e.g., the
cost of small excavations can
easily be {300/cu. yd. and boring
costs of $15/ft are not uncommon.
. There is a danger to the environ-
ment due to such methods (e.g.,
leakage from the containers if they
have been ruptured or pierced.)
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 Been used or
seem to have certain applicability: seis-
mic reflection, electrical resistivity,
electromagnetic induction, induced polari-
zation, metal detector, magnetometer, con-
tinuous microwave, pulsed radio frequency
(ground penetrating radar), infrared radia-
tion, and sonar (pulse echo acoustics).
All of the above methods are not equally
suited for identifying and locating buried
containers, but the interest in these NOT
approaches to the problem of subsurface in-
vestigation is increasing.
Based on our past work (which will be
described later), only four of these NOT
techniques have general applicability in
the detection and location of buried con-
tainers. These techniques are: metal de-
tector (MD), electromagnetic induction
(EMI), ground penetrating radar (GPR) and
magnetometer (MAG). Each method will be
described briefly.
DESCRIPTION OF METHODS
The MD and EMI methods are both induc-
tive methods. A transmitting coil sends a
continuous electromagnetic signal to a re-
ceiving coil. As a simplified description,
the signal arrives at the receiver essen-
tially through two major paths. One path
is through the air and does not change with
the search position. The other path is
through the subsurface material and is af-
fected mainly by the local electrical con-
ductivity 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 accordingly. Refer to
Lord, et al. (1982a) and McNeil! (1982),
where the commercially-available instru-
ments used in these studies are described
in detail.
The GPR method operates on exactly the
same principle as ordinary aircraft radar.
A short pulse of electromagnetic radiation
is beamed into the ground by a special
highly-damped antenna and reflections occur
from any subsurface discontinuity in di-
electric constant. The reflected pulse
arrives back at the receiving antenna and
a display 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 MAG method measures the local mag-
netic 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 mea-
suring 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 protons
is an organic liquid and the generator and
associated frequency 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
(commercially-available) is described in
more detail in Tyagi, et al. (1983).
RESULTS OF THE VARIOUS SITES
Each of these methods will operate
optimally under a prescribed set of soil
types and man-made interference. The typ-
ical sites where most waste material
containers are buried are far from these
"ideals". Rather than burial in dry granu-
lar soils, drums are usually dumped in
swamps, mudflats, water and the like.
Furthermore, the most successful methods
we have worked with are based on measuring
electrical or magnetic effects. High
electrical conductivity areas, e.g., near
equipment storage areas, junk yards or
ocean water, can severely influence the
162
-------
techniques. Soil homogeneity and soil
water conductivity are major issues. Quan-
tities of ferromagnetic material (e.g., any
steel objects) can severely affect the MAG
method.
With these thoughts in mind, a series
of test sites were obtained, containers of
various sizes were carefully placed at
different depths, geometric arrangements,
etc., backfilled.and then located using the
various NOT methods. Due to lack of space,
only a few results will be given here for
each site.
The first site was in a nearly ideal
dry sandy soil in an open field free of
man-made interference (Lord, et a!.,
1982b). This site provided an excellent
starting point and essentially narrowed our
thoughts (after careful literature review)
from seven of the possible NOT methods to
four of those mentioned previously. Steel
containers buried to 10-ft. depths were
accurately located and could possibly have
been located deeper if stable burial pits
could have been excavated. Various steel
container arrays and the boundaries of a
"metal trash dump" were accurately lo-
cated. Some plastic containers were also
located, but with poorer results. Figures
1-4 show some of the results for Site 1 for
each of the major NDT techniques. (The
actual detail of each situation is given on
the figure and in the figure caption.)
The second site was much more formida-
ble. (Koerner, et a!., 1982) Here a satu-
rated silty clay soil overlying shallow
shale rock was used. Detection depths were
much.shallower, approximately 4-ft., and
the results were influenced by the large
amount of background metal in the areas
(e.g., trailers, equipment, fences, etc.).
Of the seven methods attempted, only four
gave reasonable indications of the presence
of the buried drums. Those four methods
(EMI, MD, MAG and GPR) are those areas that
are currently used in our survey work. The
MD results at Site 2 are shown in Figure 5.
Recognizing that containers are some-
times 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). Containers were sub-
merged in water and placed on the bottom
sediments at four different sites. The
salinity of the water ranged progressively
from fresh to ocean. To depths of 3 ft. of
water above the containers, the detection
and delineation results were "excellent"
to "no good" in direct proportion to the
increase in water salinity, i.e., electri-
cal conductivity of the water. Figures 6
through 7 show some results of this under-
water detection.
Bearing directly on these three stu-
dies is the extent to which ground water
salinity influences the detection capability
of these NDT methods. To this end studies
were made at a fourth site with steel con-
tainers buried in a soil of varying elec-
trical conductivity. (Koerner, et al.,
1984) The ocean was used as an electrical
conductivity extreme and the conductivity
decreased substantially as the survey moved
inland. The soil was a medium to fine,
granular sand indigenous to the coastal
area. The sand density ranged from loose
(near the surface) to intermediate (at a
depth of 6 ft.). It was found that back-
ground conductivities greater than 40
millimhos/meter seriously impaired the use
of those methods based on electrical con-
ductivity measurements, e.g., MD, EMI and
GPR. The MAG method worked much better,
for it is a method based on magnetic mea-
surements and not on electrical conductivi-
ty. However, certain limitations were ap-
parent. The boundaries of a "trash dump"
containing metal objects were observed with
all methods even though the background con-
ductivity varied from 25-60 millimhos/
meter. Figures 8 and 9 indicate some of
the findings at Site 4.
Site 5 was the same location as Site 4
but, in this case,plastic containers were
used instead of steel. Figures 10 and 11
show some of the results of this study.
The MD, EMI and MAG did not detect any of
the plastic containers. Note the ability
of the GPR to pick up the water table as
well as the containers (see Fig. 11).
SUMMARY AND CONCLUSIONS
Table 1 presents the results obtained
at all five field sites. Some additional
remarks are in order to help in assimilat-
ing all the results in this study.
In a dry, granular soil with minimum
interference individual typical steel con-
tainers can easily be seen to a depth of at
least 10 ft. with all methods except MD,
which detects to 6 ft. Deeper detection
163
-------
is probably possible, but 10 ft. was the
limit of our burial ability. As the soil
water electrical conductivity becomes lar-
ger, the detection ability of the MD, EMI
and 6PR methods suffers. When the back-
ground conductivity rises to 40 millimhos/
meter or above, the detection ability is
very seriously impaired. The MAG method
works well under all granular soil condi-
tions, for it is not affected by high back-
ground electrical conductivity. However
an anomaly exists in our data, namely, the
inability of the MAG method to detect steel
drums in sea water saturated sand. This
point will be researched further.
In cohesive soils, there are definite
problems with MD, EMI and 6PR due to the
usual high water content and soil inhomo-
genieties. Again, a problem arises with
respect to the MAG data, since work in
cohesive soils was performed in the pre-
sence of much magnetic interfering material
(trucks, fences, etc.). Research should be
conducted in an interference-free cohesive
soil using the MAG method. Also of inter-
est would be the utility of MD, EMI and GPR
in relatively uniform, dry cohesive soils.
When steel containers were submerged
under water, the MD, EMI and GPR methods
are only of value in relatively fresh water.
When the water conductivity rises above
about 100 millimhos/meter, the three meth-
ods are quite useless. The MAG method
functions well in water of all conductivi-
ties.
Plastic containers are more difficult
to detect than steel containers. The MD,
EMI and MAS methods are quite useless in
detecting buried plastic containers. The
GPR method works well for typical size
plastic containers, especially if these are
filled with electrically-conductive mate-'
rial. However, the method still works with
non-conductive contents. These results for
plastic containers only apply for granular
soils with relatively low electrical con-
ductivity. If the granular soil has high
conductivity material in its voids or if
the soil is a wet, non-uniform cohesive
material, then the same limitations apply
to GPR as were mentioned earlier.
While this is a systematic and compre-
hensive study of NOT methods, it is not
complete and a few additional situations
still remain to be studied.
As a brief bottom line it can be
stated:
. MD, EMI and MAG all work extremely
well in detecting buried steel con-
tainers in dry, granular soil to
any typical depth.
. The MAG method works well under all
subsurface conditions.
. MD, EMI and GPR all suffer severe
loss of detection ability when the
soil's electrical conductivity rises
above about 40 millimhos/meter.
The same type figure also applies to
the detection ability for containers
submerged under water.
. GPR is the only reliable method to
detect buried plastic containers.
For a preliminary survey of a metal-
container dump site, the MD (instru-
ment cost =$400} is a good first
method, followed closely by the MAG
method (cost =$4000). More detailed
surveys can use the more expensive
instruments (EMI =$8000 and GPR
=$30,000).
. GPR can "see" excavation boundaries
(Fig. 12).
ACKNOWLEDGEMENTS
We thank the U. S. Environmental Pro-
tection Agency for continued funding of
this work through Cooperative Agreement No.
CR 807777020. Special thanks are due Dr.
John E. Brugger, the project officer on
this Cooperative Agreement, for his constant
interest, advice and encouragement. A large
number of fine Drexel University Graduate
Students deserve credit for their coopera-
tion and enthusiasm in this work.
REFERENCES
1. Bowders, J. J., Jr., Koerner, R. M. and
Lord, A. E., Jr., "Buried Container
Detection Using Ground Penetrating
Radar," Journal of Hazardous Materials,
Vol. 7, 1982, pp. 1-17.
2. Koerner, K. M., et al., "Use of NOT
Methods to Detect Buried Containers in
Saturated Clayey Silt Soil," Proceed-
ings of Management of Uncontrolled
Haste Sites, Nov. 29-Dec.' 1, 1982,
Wash., DC, HMCRI, pp. 12-16.
3. Koerner, R. M. and Lord, A. E., Jr.,
"NOT Location of Containers Buried in
Saline Contaminated Soils," Proceedings
of Managementof Uncontrolled Waste
Sites. Nov. 7-9, 1984, Mash.. DC, HMCRI
164
-------
Pub!., Silver Spring, MD, 20910.
4. Lord, A. E., Or., et al., "Use of NOT
Methods to Detect' and Locate Buried
Containers Verified by Ground Truth
Measurement, Proceedings Hazardous
Materials Spills Conference, April 19-
22, 19825, Milwaukee, Misc., HMCRI,
pp. 12-16.
5. Lord, A. E., Or., Koerner, R. M. and
Arland, F. 0., "The Detection of Con-
tainers Located Beneath Water Surfaces
Using NOT (Remote) Sensing Techniques,"
Proceedings of Hazardous Hastes and
Environmental Emergencies, March 12-14,
1984, Houston, HMCRI, pp. 392-395.
6. Lord, A. E., Or., Koerner, R. M. and
Freestone, F. 0., "The Identification
and Location of Buried Containers Via
Non-Destructive Testing Methods,"
Journal of Hazardous Materials, Vol. 5,
1982a, pp. 221-233.
7. McNeil!, J. D., "Electromagnetic Resis-
tivity Mapping of Contaminant Plumes,"
Proceedings of Management of Uncon-
trolled Hazardous Waste Sites, Nov. 29-
Dec. 1, 1982, Wash., DC, Hazardous
Materials Control Research Institute,
Silver Spring, MD, pp. 1-6.
8. Tyagi, S., Lord, A. E., Jr. and Koerner,
R. M., "Use of Proton Precession Mag-
netometer to Detect Buried Drums in
Sandy Soil," Oournal of Hazardous
Materials. Vol. 8, 1983, pp. 11-23.
9. Weston Consultants, "Ground Penetrating
Radar Survey Elizabeth River, New
Jersey," Final Report of U. S. Coast
Guard, New York, 30 June 1981, 14 pgs.
J
55
•
GAi
1.
10
§
GAL
ALL «JtH 3.5' 6F CDV£R
S5*L 2 GAL
.0. ,1
(a)
5TRQH&
S
WEAK
t
T reft
Alt 3D GAL
t
r
1.
0 10 20 30 HO 50 60 70
POSITION < FEET)
FIG. 1 MD results for Site 1 (dry, granu-
lar soil):
(a) various size steel containers
with 3.5 feet of cover
(b) 30-gallon steel drums -
various covers.
Actual drum positions are shown
as black circles.
165
-------
600
A 400
*f>
I
laob
a.
a.
20 40
DISTANCE (FT)
60
FIG. 2 MAG results for Site 1 with indi-
cated steel containers. Labels on
curves indicate traverses with off-
set (from directly overhead) in
feet. ATI containers at 3.5 feet
of cover, (y = 105 gauss)
Cff.)
FIG. 4 GPR results for Site 1. Inverted
parabolas indicate the buried
steel containers. All containers
at 3.5 feet of coyer.
55 GaUON METAL DRUMS - SINGi-i AND MULTIPLE
30 40
DISTANCE (It)
FIG. 3
EMI results for Site 1 with indi-
cated steel containers. All con-
tainers at 3.5 feet of cover.
= PLASTIC
FIG. 5 MD results for Site 2. Typical
burial was 30-gallon steel con-
tainers at about 3 feet of cover.
Note: None of the plastic con-
tainers could be detected with MD;
all steel were detected.
166
-------
DISTANCE-IFTO
«0 SO M
FIG.
BOTTOM
6 Ground penetrating radar (6PR) scan
at Site 3. Clearly seen are the 55-
gallon steel drum (with 2 feet of
water cover) at 60 feet, and the
three 5 gallon drums (with 3 feet
of water cover) at 20, 20 and 35
feet.
>|60
CONDUCTIVI
lillimhos/ n
> s s
„ -
S' 55380
UJ
u. 55130
£54980
z
z (
FULL
o
1 HALF
a NULL
(
-•• .
t i i
} 20
> *
) 20
,
) 20
) . 20 .
EMI
i i i
40 60
MAG
40 60
MD
i ii
40 60
DISTANCE ( feet)
«0 , 6p
EM INDUCTION RESULTS
(H*.tp',2l*.H')
JO JO ~ 40 5O SO TO
•FIG. 7 The electromagnetic induction (EMI)
scans from the four locations at
Site 3. The spatial distribution
and water cover for the drums are
indicated on the figure.
FIG. 8 Results for EMI, MAG and MD in
delineating the boundaries of a
"metal trash dump" as Site 4,
which is an ocean sand of varying
electrical conductivity.
so'
.FIG. 9 Results of GPR survey of 30-gallon
steel containers buried at depths
from 9 inches to 50 inches at
Site 4.
167
-------
__ mte
>m
: nun
»?
r *««»
UAC MFTMOB
J J 1 , I
FIS. 10 Electromagnetic induction (EMI)
and magnetometer (MAG) surveys
over 35-gal. plastic containers
buried under 26" of sand cover 25'
apart. Results apply for the
following container contents: air,
fresh water and salt water, as
Indicated. Burial was at "Site
5", which was the same location
as Site 4.
3$ Ml.
FIG. 12 Two GPR scans over a metal trash
dump. Note the downward slopes at
the ends of each trace (at the
arrows), indicating sloping side
walls of filled-in excavation.
FIG. 11 Ground penetrating radar (GPR)
scan over 35-gal. plastic con-
tainers buried under 26" of sand
cover and placed 25' apart. Re-
sults apply for the following
container contents: air, fresh
water and salt water, as indi-
cated. Burial was at "Site 5",
which was the same location as
Site 4.
168
-------
Table 1. - General Acceptability of Using Various NDT Methods to Locate Typical Sized Buried
Containers under Conditions Listed (Maximum Penetration Depth in Parentheses)
Steel Containers
Soil Type
(Reference)
Granular
Cohes i ve
Water
Saturation
01 - 201
10% - 501
50% - 100%
501 - 100%
1001
low
100Z
Type of
Void Water
fresh
intermediate
ocean
fresh
fresh
intermediate
ocean
Metal
Detector
excellent (61)
excellent (2')
excellent (4'+)
excellent (3'+)
poor
no good
Electromagnetic
Induction
excellent (ID1)
average (41)
no good
poor (41)
excellent (3'+)
no good
no good
Ground Penetrating
Radar
excellent (10 ')
excellent (31)
poor (21)
excellent (4'+)
excellent (4'+)
no good
no good
Magnetometer
excellent (10'}
excellent (4'+)
no good*
poor (4')+
excellent (3'+)
excellent (3'+)
excellent (3'+)
Plastic Containers
Granular
10% - 50%
50% - 100%
intermediate
ocean
no good
no good
no good
no good
excellent - if
contents con-
ductive (4'+)
fair - if con-
tents non-con-
ducting
poor
no good
no good
*Not understood - Needs further research
Many interfering magnetic objects
-------
A SELECTION GUIDE FOR MOBILE/PORTABLE TREATMENT TECHNOLOGIES
FOR THE REMOVAL OF VOLATILE ORGANICS FROM WATER
Jeffrey L. Fleming
IT Corporation
Knoxville, Tennessee
and
Michael D. Royer
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Edison, New Jersey
ABSTRACT
This paper describes the salient features of a Guide for evaluating the appli-
cability to hazardous substance cleanup operations of volatilization technologies,
such as surface sprayers, surface aerators, bubble columns, cooling towers, and
steam strippers. Unaided evaporation from a pond is also addressed. The Guide
enables the user to assess performance and cost under a variety of operating condi-
tions (e.g., temperature, influent concentration, allowable liquid and gas effluent
concentration, and flow rates) for "representative" equipment designs that could be
transported on a trailer 8 feet wide, 45 feet long, and with a maximum height of
13.5 feet. The designs are used as a basis to calculate representative contaminant
removal efficiency, treatment rates, air emissions, and treatment costs of each
technology. A key parameter used in assessing these technologies is the Henry's
Law Constant (Hc). A tabulation of available values of Hc is provided for CERCLA-
designated volatile hazardous substances, and methods for estimating Hc are also
described. Qualitative guidance is provided on other factors that should be con-
sidered during site-specific assessments of the technical and economic feasibility
of volatilization technologies.
170
-------
INTRODUCTION
A Guide is being prepared for the EPA
on the preliminary selection of treatment
technologies for removing spilled volatile
organics from water during removal actions.
The Guide assists the user to assess the
technical and economic feasibility of re-
moving volatile organic contaminants using
several volatilization technologies. The
Guide enables the user to assess perfor-
mance and cost under a variety of opera-
ting conditions (e.g., temperature, influ-
ent concentrations, allowable liquid and
gas effluent concentration, and flow
rates).
This Guide will aid On-Scene Coor-
dinators (OSC) by reducing duplication of
effort, accelerating the production of
cost and performance estimates for de-
cision-makers, and promoting consistency
in estimation procedures. Technical per-
sonnel who support the EPA OSC by develop-
ing cost and performance estimates for
water treatment options are the principal
target users for this Guide. In order to
effectively use this Guide, college level
training in chemistry and thermodynamics
is recommended. Only limited experience
in spill cleanup activities is required.
The Guide is, of necessity, written
about "representative" types of equipment
and about selected situations. Although
the final selection of equipment should
take into account the factors cited in the
Guide, it will be necessary to give con-
sideration to the individual character-
istics of the equipment and the situation
in which it will be applied.
The impetus for developing a Guide
stems from involvement of the Hazardous
Waste Engineering Research Laboratory's
Releases Control Branch in technical assis-
tance activities that require assessment
of the feasibility and cost of various
treatment options. It was recognized that
EPA OSC's and their technical support per-
sonnel are often faced with changing or
uncertain conditions that could affect the
cost and feasibility of removing volatile
substances from water. As conditions
change or as some of the uncertainties are
resolved, the OSC's technical support per-
sonnel are called upon to revise their
estimates accordingly. It was recognized
that the OSC's and their technical support
staffs did not have a concise guide on the
subject of volatilization technologies and
their application to spill cleanup opera-
ations.
Some situations that OSC's have faced
include the following:
Hide potential variation in allowable dis-
charge limits. In one case, the favored
option was to pump spill-contaminated
groundwater to a publicly-owned treatment
works (POTW). The POTW declined to accept
the waste stream due to concern about the
flammable nature of the contaminant. The
remaining option was to pump the effluent
back into the ground, which required treat-
ment to less than 50 parts per billion
(ppb) of the contaminant.
Wide variation in the influent concentra-
tion.At the incident mentioned above,
samples from the contaminated groundwater
that were taken at various times indicated
that the influent concentrations to the
treatment system could range from 200 to
1500 parts per million (ppm).
Potentially widevariation in the ambient
temperature.Itis often difficult to
accurately predict how long it will take
to progress from the planning stage to the
operating stage. Hence, the treatment
may not occur during warm weather, which
favors volatilization. It is often desired
to estimate the effect on efficiency and
cost if the treatment is conducted during
unfavorable ambient temperatures.
Potential air emissions from the treatment
system. Volatilization technologies that
were attractive from the standpoint of
simplicity and comparatively low operating
costs were found unacceptable in one inci-
dent due to the air emissions problem that
they caused.
171
-------
Assembling technology specific information
and developing a selection rationale.For
each Incident, it was necessaryto assem-
ble information on various technologies,
and then determine a suitable selection
approach. This information collection
effort was time consuming—when time was a
critical element. It was also recognized
that a decision-making rationale could be
developed that would permit faster, more
comprehensive, and more uniform perfor-
mance and cost estimates.
Assembling chemical specific data. One
problem that has repeatedly been encoun-
tered is finding the Henry's Law Constants
(Hc) for the compounds of interest. The
data that do exist are located in a var-
iety of sources. The Hc data are reported
in a variety of units, and these units are
not always readily interconverted, which
adds to the difficulty of making perfor-
mance estimates.
SCOPE OF THE GUIDE
The Guide specifically addresses vol-
atile organic liquids that are also clas-
sified as hazardous substances under the
Comprehensive Environmental Response Com-
pensation and Liability Act (CERCLA). A
Materials Property Table for 74 volatile
compounds is provided in the Guide that
includes data [e.g., water solubility,
vapor pressure, Henry's Law constant (the-
oretical and empirical)] that are commonly
used in estimating performance of volatil-
ization technologies. (See Table 1.)
The Guide is targeted for use in sup-
port of removal actions that require rela-
tively rapid mobilization and short set-up
times. Therefore, the-technologies that
are considered are limited to those that
could be legally transported (i.e., no more
than 8 feet wide, 45 feet long, and 13.5
feet high) by truck. For each class of
technology selected, the largest transport-
able design is considered for performance
comparisons between technology types.
The types of volatilization approaches
that are addressed 1n the Guide are solar
evaporation from a pond, surface spraying.
high speed surface aeration, bubble column
air stripping, spray column air stripping,
counter-current packed tower air stripping,
cooling tower air stripping, and steam
stripping. By addressing these technolo-
gies in the Guide, the reader is provided
with a wide range of options in terms of
capital and operating costs, removal effi-
ciencies, treatment rates, complexity,
availability, and air emissions.
A variety of other technologies, such
as cross-flow air stripping and a propri-
etary activated carbon/stripping hybrid
technology, are not included because of
their similarity to other technologies or
because systems are not available for wide-
spread use. Due to resource limitations,
the Guide does not address mobile or read-
ily portable technologies for treating the
off-gases from the volatilization pro-
cesses that are described.
DEVELOPMENT OF PERFORMANCE ESTIMATES
System performance is estimated in the
Guide for material removal efficiencies,
treatment time requirements, and emission
rates. Cost data are also provided.
For each type of technology, a repre-
sentative design is chosen that has the
largest available treatment capacity and
also meets the transportability require-
ments described above. For each represen-
tative design, several operating condi-
tions are selected, performance estimates
are made, and the results are tabulated or
plotted. Organic removal efficiency and
time to treat a model impoundment are cal-
culated for each design case system and
operating condition.
Organic Removal Efficiency Estimation:
The air stripping and steam stripping sys-
tems are compared based on the calculated
organic removals that may be achieved using
a representative design on a single-pass
basis. A total of 32 performance curves
(Hc vs. perce.nt removed) are presented in
the Guide for these column systems. The
performance curves are generated for a
range of configurations ([single unit, par-
allel, and series], one-pass vs. recycle
172
-------
TABLE 1. MATERIAL PROPERTIES TABLE Material property data are provided in the
Guide for 74 compounds.
CERCLA
Compound
(Synonym)
Acetic Acid
(Ethanoic
Acid)
Acetic Acid,
Ethyl Ester
( ethyl
acetate)
Acetic
Anhydride
Acetone
(Propanone)
Acetone
Cyanohydrin
Chemical
Formul a
CH3COOH
CH3-
COOC2H5
(CH3CO)20
CH3COCH3
(CH3)?-
C(OH)CN
Solubility*
(in H20)
PPM
misc
89,800
Hydro-
lyzes
(forms
acetic
acid)
misc
vs
Hyd ro-
lyzes
(forms
HCN)
°C
25
Vapor
Pressure*
mm Hg
760
10
760
95
760
214
5.3
760
227.1
200
15
760
"C
118
17
77
25
139
100
25
56
25
25
81
95
Henry's Law Constant*
Theoretical
P = 1 atm
1.29
0.01
6.3
—
M_
—
—
0.28
6.8 x 106
«-
--
Units
none
none
none
none
atm~m3
mole
UC
100
17
25
25
Experimental
P = 1 atm
0.73
_ _
—
__
—
—
15.5
— -*
--
Units
none
none
"C
100
100
* The source of the data are also reported in the Guide.
173
-------
treatment of an impoundment), and opera-
ting conditions (i.e., gas and liquid flow
rates, number of theoretical stages, and
packing). The performance computations
are based upon accepting that the effect
of air flow rate, liquid flow rate, Hc,
and number of theoretical stages is de-
scribed mathematically for a continuous
isothermal air stripper in the Kremser
equation:
1 - (G/L)
WT6/L)(H)J
Where:
f * fraction of material left in
liquid phase;
G = molar flow rate of gas [in
moles/min, for air, G = 0.0026
x flow rate (cfm)]
L = molar flow rate of liquid [in
moles/min, for water, L = 0.46
x flow rate (gpm)]
H * Henry's law constant of strip-
pable component {mole fraction/
mole number)
N * Number of stages in column.
Although steam stripping is a compara-
tively expensive and complex technology,
the use of appropriate vent controls can
reduce air emissions below those produced
by treatment technologies such as air
stripping, which discharge the contamin-
ants directly to the air. However, caution
must be exercised, since a concentrated
product from steam stripping can also pose
health and flammability hazards.
The column diameter of the model system is
limited by the physical size and weight of
the auxiliary system equipment required to
operate the column. The largest standard
column diameter (and consequently heat
duty) that can be placed on a flatbed
trailer is 1.5 ft. Random packing is cho-
sen over trays because of ease of cleaning,
and over rigid packing because of avail-
ability. The height of the packing in the
model steam stripper is 25 feet, which is
chosen as a realistic size for a single
flatbed trailer. Performance is estimated
for four different boil-up ratios (3%, 5%,
10%, and 30%). The treatment rate varies
from 96 to 24 gallons per minute. The
organic removal efficiency for evaporation,
surface aeration, and surface spraying is
considered only as a function of operating
time.
Treatment Time Requirements:
A model impoundment with a volume of
750,000 gallons {i.e., 100 ft x 100 ft x
10 ft) is used as the basis for comparing
the treatment time required to remove or-
ganic contaminants from water for the de-
scribed systems. It is assumed that there
is no net flow into or out of the impound-
ment during the treatment period {i.e.,
batch system). The model is presented as
representative of a typical body of con-
taminated water and is adequate in size to
accommodate available commercial-sized
mechanical agitation equipment.
The criteria for comparison is the half-
life of the organic contaminant in the
model impoundment. The half-life {tn) is
the operating time required to reduce the
concentration to 50% of its original
level. Values for t^ for the surface
sprayer, surface agitator, and solar evap-
oration are calculated from the following
equation, which is derived in the text:
tn = 0.693/[KL (a)]
K-J is the overall mass transfer rate coef-
ficient {m/hr), and a is the specific
surface area of the Tiquid phase (m^/m^).
The methods and assumptions employed to
compute K|_ are described in the Guide.
Values of tj, vs. Hc are plotted for both
surface spraying and surface aeration for
two sizes of commercially available units.
No attempt is made to quantify effects of
other incidental variables such as climatic
conditions of wind and temperatures, dif-
ferences in a particular equipment design,
and quality of the contaminated water.
Instead, reasonable average values for
the key variables are estimated based on
probable field conditions.
174
-------
The operating time required to obtain a
desired removal efficiency can be deter-
mined by multiplying th by the number of
half-lives (n) required to determine the
desired percent reduction (R) in the con-
taminant concentration. A plot of R vs.
n is provided for the convenience of the
reader.
The continuous systems of air and or
steam stripping could be used to augment
volatilization from an impoundment. In
this case, the discharge from the treat-
ment system could either be placed back
into the impoundment or sent to off-site
disposal. For the case where the dis-
charge is placed back into the impound-
ment, the half-life in the pond will be
governed by the equation:
= 0.693
where:
th =
V =
L =
f =
L (1-f)
half-life of organic in the
impoundment;
volume of the impoundment;
liquid flowrate of the treatment
unit; and
fraction or organic remaining
after treatment, based on the
Kremser equation.
The results of the above equation have
been plotted in Figures 12 through 19 of
the Guide for the design case systems and
operating modes in treating the model
impoundment.
These performance estimates neglect the
volatilization that would naturally occur
from the impoundment. The actual half-
life, including this or any other compet-
ing removal mechanism, may be obtained
using the following equation:
where:
th half-life of the organic in the
Impoundment; and
tnx half-life of the organic consider-
ing mechanism X.
Examination of the half-life figures allows
the conclusion that surface aerators will
normally be the best option to augment vol-
atilization from an impoundment. Howeverj
operational constraints of surface aeration,
the desire to control organic emissions,
or the unavailability of a surface aerator
may require the use of one of the other
units for this service.
A Table is also provided in the Guide that
compares the design case continuous treat-
ment units based on the assumption that
off-site discharge is available. In this
case, the treatment time is purely a func-
tion of liquid flow rate. For the conven-
ience of the reader, the percent removals
for each of four Henry's Law Constants
are also given.
Emission Rates and Costs:
Tables are also provided that contain es-
timates of the flowrate and concentration
of air emissions; and capital, mobiliza-
tion, demobilization, and operating costs.
USING THE GUIDE
The technology evaluation approach in
the Guide progresses through four stages:
(1) site characterization; (2) calculation
of basic material properties; (3) technol-
ogy evaluation; and (4) equipment selec-
tion. Decision points are noted where
technologies may be eliminated from con-
sideration based upon inability to ade-
quately remove contaminants, insufficient
treatment rate, excessive cost, or lack
of availability.
Characterization of the site is situ-
ation dependent, so a checklist is pro-
vided to assist the reader to collect and
consider the pertinent information.
Following site characterization, it
is recommended that the user determine the
pertinent properties of the spilled mate-
rials. For CERCLA volatile substances, a
tabulation of data is provided. For data
175
-------
not found in the tables, estimation meth-
ods are provided, as well as pertinent
references.
Next the reader is advised to perform
system evaluations. It is suggested that
removal rates be considered first, then
flowrates and time requirements for treat-
ment, and finally emissions that will be
generated by treatment. Technologies that
remain after the performance evaluations
should then be compared based on an anal-
ysis of costs. Costs for pretreatment,
disposal of treated water, emission con-
trols, and water polishing units should be
added.
Through the evaluation process in the
Guide, the user will be able to identify
promising types of volatilization technol-
ogies and will also find the information in
the Guide useful for more in-depth evalua-
tions of cost and performance of specific
technologies. However, the Guide is not
designed to be the sole reference for mak-
ing the final selection of a treatment
system. There are situation specific con-
siderations that are beyond the scope of
the Guide. As examples, addressing the
problems caused by poor water quality
{e.g., salts, solids, biological material);
evaluating the significance of differences
between equipment of the same type; or per-
forming pilot tests, cannot be adequately
addressed by reference to the Guide. A
flow chart of the approach recommended in
the Guide for evaluation of volatilization
technologies is provided in Figure 1.
SUMMARY
The Guide provides valuable informa-
tion and data in one place on representa-
tive volatilization technologies. It is
expected that this volume will be useful
to OSC's and their technical support staffs
in conducting preliminary evaluations of
the technical and economic feasibility of
removing spilled volatile substances from
water.
This work was performed by IT Corpor-
ation under EPA contract 68-03-3069 for the
Releases Control Branch, Land Pollution
Control Division, Hazardous Waste Engineer-
ing Research Laboratory.
176
-------
A GUIDANCE MANUAL FOR THE SELECTION MO USE OP
FOR I3QUID HAZARDOUS
S-6eve Gibson
Combustion Engineering, Inc.
• Newbury Park, California 91320
Bobert Melvold
Rockwell International Corporation
Palmdale, California 93550
W. Ellis
JRB Associates
McLean, Virginia 22102
Bobert Scarberry
U.S. Environmental Protection Agency
Washington, D.C. 20460.
Mike Boyer
U.S. Environmental Protection Agency
B3ison, New Jersey 08837
ABSTRACT
This study was conducted to provide information for the selection and use of
sorbents for cleanup or control of liquid hazardous substances. Literature reviews,
sorbent manufacturer data and experiences of on-scene coordinators were reviewed in
conjunction with laboratory studies. These laboratory studies determined the com-
patibility and sorption capacity of selected representative hazardous liquid-sorbent
pairs. The combined experimental and literature data were used to prepare A Guidance
Manual for the Selection and Use of Sorbents for Hazardous Substance Releases' (Manual).
On-Scene Coordinators and their technical support personnel are the primary target
audience for the manual.
To utilize the Manual, the user must first identify the spilled liquid. If it is
one of the 200-f- liquid hazardous substances addressed in the Manual, a reference is
provided to one of 25 "Sorbent Selection and Use Guides" ("Guides"). Each of the
"Guides" enables the user to rapidly identify generic sorbent classes, physical forms,
and methods for application and collection that are most suitable for each of four
different chemical release-control scenarios: 1) a spill onto land; 2) a floating re-
lease into •water; 3) a non-floating release into water; and, 4) immobilization for
landftiling.
The Manual contains "Sorbent Data Sheets" for 13 generic classes of sorbents
other than activated carbon. These data sheets contain information on manufacturers,
acquisition costs, buUc density and sorbent limitations. The sorption capacity of 190
sorbent-chemical pairs was determined and recorded in the Manual. The Manual also in-
cludes cost and estimation procedures, test methods, hazardous liquid physical proper-
ties, and a description of the rationale for the release-control scenarios.
177
-------
Spills and releases of liquid haz-
ardous substances pose a severe threat
to the public and the environment. These
substances may be released at fixed sites
or during transportation accidents and
generally require control and cleanup.
Sorbents are potentially effective ma-
terials for cleaning up and controlling
many releases of liquid hazardous sub-
stances. She Manual described in this
paper •was designed to facilitate the se-
lection and use of appropriate sorbents
for treating releases of hazardous liq-
uids. 3he Manual is targeted primarily
to assist Federal On-Scene Coordinators
(OSC's) and their technical support
staffs, but is also applicable to emer-
gency response, spill cleanup and in-
dustrial personnel who respond to re-
leases of hazardous liquids and im-
nobilize such substances prior to dis-
posal.
Development of the Manual involved
a review of the literature, sorbent man-
ufacturers' data, OSC experiences, and
laboratory studies. Die laboratory
studies determined the compatibility and
sorption capacity of selected hazardous
liquM-sorbent pairs. To establish a
framework for the Manual, four liquid
release-control scenarios were studied:
1) a spill onto land? 2) a floating re-
lease onto water; 3) a non-floating re-
lease into water; and 4} immobilization
for landfilling. "Sorbent Selection and
Use Guides" were developed which list
sorbents for each release-control sce-
nario. Sorbent Data Sheets were also
prepared for each 13 classes of sorbents.
The Manual also provides information on
sorbent application, collection and dis-
posal, cost estimation procedures, test
methods, hazardous liquid physical prop-
erties, and a narrative to enable se-
lection of appropriate sorbents. Ihe
Manual is designed to provide information
for rapid decision making and for con-
ducting thorough evaluations of alter-
native sorbent-use strategies.
PROCEDURES
Ihe Manual development project was
completed in three phases. Ihe approach
and results of each phase are described
below.
Phase 1; Information Collection and Test.
Plan Development
Development of a List of Applicable
Hazardous Liquids
A list of hazardous liquids for
which the Manual would be applicable was
developed from the substances regulated
by the Comprehensice Bwironmantal Re-
sponse, Compensation and Liability Act
of 1980, EL96-150 (CERCIA). The CERCIA
hazardous liquids of concern are 213 neat
hazardous substances that meet the
liquid criteria designation (i.e., that
a substance possess a melting point at or
below 77°F (25°C) and a boiling point at
or above SOT" (10°C). Waste mixtures
were not considered.
These QSRCIA-regulated liquid
hazardous substances were placed into
27 classes according to functional groups
(Herrick et al, 1982). Since it was not
possible to experimentally evaluate all
of the CERCLR. hazardous liquids, repre-
sentative hazardous liquids were selected
from each chemical class based on aqueous
solubility, specific gravity, liquid
surface tension, and viscosity.
Collection of Information
Tto complete the "Sorbent Selection
and Use Guides" and the "Sorbent Data
Sheets," pertinent data were collected.
The physical data gathered for each
liquid included specific gravity, aqueous
solubility, surface tension, and vis-
cosity. Bor physical data that was un-
available from technical reference ma-
terials and journals, the data were
solicited from chemical manufacturers
of the liquids.
A literature review was also con-
ducted to identify available sorbents
and their manufacturers, and to obtain
published data on sorbent properties and
performance. Based on information de-
ficiencies noted in the available liter-
ature, data tsheets were requested from
sorbent producers. Limited sorbent prop-
erty and procurement information, and
sorbent-hazardous liquid pair data such
as sorption capacity, retention capacity,
sorbent compaction and expansion data,
performance parameters and safety pre-
cautions were obtained.
178
-------
Test Plan Formation
Compatibility Testing
For each chemical class, a repre-
sentative chemical was chosen. Similarly,
a representative sorbent was chosen from
each of the 13 generic sorbent classes.
Those representative sorbent-hazardous
liquid pairs for which no data were avail-
able became primary candidates for test-
ing. The selection of a hazardous liquid
for testing considered the probability
that a substance from one of the 27 chem-
ical classes would be released and the
degree of hazard that it would present to
public health or the environment. Prob-
ability of a release was based on the
annual production volume, while the de-
gree of hazard was based on the proposed
reportable quantity (RQ) of the CERCLA
liquid as promulgated by the Environ-
mental Protection Agency in 48FR23552-
23605 and subsequent revisions. The sor-
bents were assigned a priority for test-
ing based on 1) sorbent availability and
2) sorbent applicability for ameliorating
a spill similar to one or more of the
four cleanup scenarios. The following
list presents the 13 generic sorbent
categories identified by the project:
Sorbent day
Diatomite
Wood Fiber
Treated Wood Fiber
Expanded Mineral •
Foamed Glass
Polyurethane
Polyethylene :
Polypropylene
Cross-linked Polymer
Feathers
Treated Clay/Treated Natural Organic
Treated Expanded Mineral/Treated
Wood Fiber
' . The sorbent-hazardous liquid pairs
chosen for testing were from the top
priority groups, resulting in selection
of a maximum of 250 pairs from 23 chem-
ical classes and 10 generic sorbents.
PHASE 2; Compatibility and Sorptioh
Capacity Testing
Laboratory experiments and esti-
mation procedures were employed in
Phase 2 to generate or estimate sorbent
performance data that were not available
from the literature. The Phase 2 work
included: ...
A Standard Operating Procedure (SOP)
was prepared that describes the proce-
dures employed in conducting the com-
patability tests. When a gross incom-
patibility was observed or degradation of
the sorbent occurred, the sorbent-hazard-
ous liquid pair was excluded from sorption
capacity 'testing. Also, when the- sorbent
was not wetted by the hazardous liquid,
the sorbent-liquid pair did not qualify
for sorption capacity testing. The sorp-
tion capacity measurements of 190 sorbent-
liquid pairs were conducted in Phase 2
of the project. The 10 sorbents tested
for sorption capacity with the represen-
tative hazardous liquids were also tested
for their sorption capacity with water.
The data obtained were used in determin-
ing hazardous liquid/water preference
indices for the sorbents.
Sorption Capacity Testing
Sorbent testing protocol consists of
three procedures for three different sor-
bent forms: (1) A decanting procedure for
particulate sorbents, in which after a
2-hour exposure in a graduated cylinder,
the hazardous liquid is poured out through
a stainless steel screen leaving behind
the loaded sorbent for measurement; (2)
A volume measurement procedure for finely-
divided particulate sorbents, wherein the
sorbent is allowed to settle and the
volume of the sorbent layer (and the
liquid contained therein) is determined?
and (3) A pad/mat immersion procedure, in
which a standard-sized sorbent specimen
is immersed in and then removed from a
hazardous liquid and weighed after drip-
ping ceases.
Sorbent/Hazardous Liquid Data Estimation
Project funds did not permit experi-
mental determination of all sorbent-
hazardous liquid sorption capacity values
of interest. Procedures were established
for estimating sorption capacities for
cases where the untested sorbent-chemical
pair was sufficiently similar to pairs .
for which data were available. The esti-
mation procedures are not rigorously de-
fensible from a scientific standpoint,
but it was concluded that On-Scene Co-
ordinators and their technical support
personnel would be forced to develop
similar estimation procedures when faced
179
-------
with hazardous liquid releases for
which sorption capacity data did not
exist, therefore, the estimates were
made, included in the Manual, and clearly
marked as estimates.
Phase 3; Manual Preparation
This phase of the project involved
the following activities:
o Development of the "Sorbent
Selection and Use Guides"
o Preparation of the "Sorbent Data
Sheets"
0 Production of ancillary data
tables:
- sorption capacity
- hazardous liquid/water
preference
- physical properties
- test methods
- equipment/sorbent costs
o Preparation of the handbook test
Description of these elements follows in
the Results and Discussion section.
RESULTS AND DISCUSSION
A document, A Guidance Manual for
the Selection and Use of Sorbents for
Liquid Hazardous Substance Releases has
been prepared, which contains eight
sections. The first three sections,
"Sorbent Selection and Use", "Sorbent
Data", and "Technical/logistical In-
formation", are essential to users who
require a condensed source of sorbent
selection and use guidance. The other
five sections, "Cost Estimation Pro-
cedures and Data", "Test Methods",
"Spill Scenario Rationale", "Sources of
Information", and "CERCLA Liquid Chemical
Information," explain the rationale used
to develop the condensed guidance in the
first three sections and also provide
infonnation that enables the user to
acquire and use sorbent information for
specific needs. A synopsis of each of
the eight sections is provided below.
Section A - "Sorbent Selection and Use"
Two CBOA Liquids Indexing Tables
are an integral part of Section A. The
first indexing table lists CERdA des-
ignated hazardous liquids, the chemical
class and guide number to which each
liquid has been assigned, the CAS number
for each liquid, the hazards in addition
to toxicity, and the behavior of the
liquid in water. The second indexing
table lists the 4-digit DOT ID number
used in cormercial shipping, and supple-
mental information similar to that in
the first indexing table. Information
is provided for all 213 CERCIA liquids
that were identified.
Section A of the guide also contains
the "Sorbent Selection and Use Guides".
Each "Guide" presents information on the
use of sorbents relevant to the four
scenarios: Landspill, Floating Spill,
Non-Floating Spill, and Landfill. Chemi-
cal functional classes are generally re-
garded to be descriptive of the chemical
reactivity of the CERCIA liquids? con-
sequently, one guide was prepared for
each chemical class for which data were
available. Each guide lists generic
sorbents, in recommended order for con-
sideration for each of the four scenarios.
An example of a "Sorbent Selection and
Use Guide" is in Table 1. Application
and collection procedures are listed in
each "Guide" based on assessment of the
optimal procedures for use with each
generic sorbent. For each sorbent listed
in each "Guide", the page number of the
"Sorbent Data Sheet" is also included.
Section. B - "Sorbent Data"
This section contains 1) an index
of sorbent manufacturers, trade names,
and generic sorbent categories, 2) 13
"Sorbent Data Sheets", 3) the sorption
capacity data, and 4) the hazardous
liquid/water preference indices.
A "Sorbent Data Sheet" was prepared
for each generic class of sorbents.
Each data sheet lists the generic sorbent
class, conniercial trade-names, manu-
facturers data on the types of sorbents,
their cost, bulk density and storage re-
quirements. The sorbent types generally
available includes particulates, pillows,
pads, mats, and boons. The data sheets
also report limitations on the use of
each generic sorbent for landspills,
floating spills and non-floating spills.
180
-------
Table l.._ SORBEHT SELECTION AND USE GUIDE
GUHE mmm 10
CHMCOL (USS: Aromatic Hydrocarbons
E MZAHXJB LIQUID: Toluere
For jieleases onto land, reecdmsided sorbetts are listed to prtcrttlasd order uxSer tha Landapm Scenario, for water,
under the Floating Spill or Non-Floating 3(031 Scenarios, and, for landfill applications, mder Landfill Scenario. Fbr
additional sa-bent friorlttatlon iBforaatien, see the Sorter* Priorttiiaticn taM.es en pagjs 1-11 thru 1-16, Section I of
the handbook. Also, for additional scrbent specifics turn to the Sorbent Data Sheet fouid on the pags listed in the
prioritized sorbent oolum. Pertinent logistical information, such as application and collection BBthodoles', Is given in
Section C of the handbook.
utnroi. - a-wi.
Priori, tlaad
Sorbents/lype/tte
(DFeattera/pl/B-20
(2)ClP/p/B-19
(2)C3J>/pl/B-19
(3)Sorbent Oay/p/B-10
JippllcaU.cn
Ihrow
aovel
Throw
Shovel
Iftrcw
Ccaieetlon
Pitchfork
Shovel
Pitdifork
Shovel
Shovel
Pitchfork
Limitations
DCE.RT
R.W.SS
R*I,p'
Prlorltl;gd
Sorfaents/Type/Pg
(DOP/p/B-19
(3)Sorbent Clay/p/B-10
( 3)Polypropylere/p/B-l 8
(3)Feathers/piyB-20
(*)a HbBral/p/B-14
CjU.eoti.on Limitations
Saplceder R/W/SS
StdploaJer R/I
Sdploafer H,I,P
Sanlcaler W,SS,DOC
Stdplcader DQ3,Rr
Sdploader
FLMTI1C SPILL
HOMLOffiNG - SMftLL FOB)
NCK-fLCHIMG - UUtE LAKE
Prforttizgd
Sorbents/I^pe/Rg
(1)CLP/pl/&-19
(2)Fsathere/pl/B-20
(8)W Fiber/py&-13
Throw
Colleotlon Limitations
Spear
Spear
Spear
Spear
(1)Polypropylens/iB/B-l8
(J()Polyethylene/in/B-17
FLOmS - SMALL CHEEK
Priori tlzsd .- . • .
Sorbents/Type/Pg , Application Collection Limitations
(1)Polypropylens/b/B-18 Knd Hand
(2)Polvurethane/b/B-16 and Hand
fend Band
tend ttnd
(3)Feathers/b/&^0
(*)TO Fiber/b/B-13
C )
PrioritlsEd
Sgrbents/Type/Pg
(2)Feathera/pl/B-20 BoatvTlirow
(3)TO Flber/pl/B-13 Bcat/Ihrow
(1)Polypro[yleia/wB-l8 Baat/Throw
{M)polyurathane/m/B-l6 Bsat/Throw
Jpplfcatkn Collection Limitations
ftjat/Throw Boat/Spear R
Boat/Spear
Boat/Spaar
Boat/Spear
{5)Polyethylem/n/B-17 Boat/Throw Boat/Spear
- LMGE RIVER
Prtorttiaed
Sorbents/Type/Fg Appllcatioi Colleotlcn Umltations
(1)Folypropl«ieA>/B-18 Baat/tend ftat/Wlnch
(2)Polyurethane/b/B-l6 Boat/Hand Bat/Utah
(3)Faattera/b/B-20 Ebot/lfand Boat/Winch
tt)TM FibeiVb/B-13 Baat/Hand
UMKffiL
OCUMJ
BOTCH TffiSBENT
Bstjmatad*
Priori tlzEd
Sorbenja/Type/Pg Limitations
( 1)PolypropylenB/p/B-l8
PriorltiZEd
Sorbenta/lype/PE
( 1)Feathers/pl/B-20
( 1)ftjlyethylene/qvB-17
( DPolyurethane/m'B-16
(DCLP/pl/B-19
(2)Pdyp«t5pylen9/o'B-l8
Limitations
R
R,!W
Prtoritlzgl
Sorbents/Iype/Ps
(1)Sc Mlnsral/p/B-W
(3)Sorbent Clay/p/B-10
C4)Wood Flber/p/B-12
(5)FbaEwd Olass/p^-15
{6)Polvpropylene/p/B-l8
Costs($)
1175
1358
I67f
2017
32T1
3852
Legend:
b = Boom . pi
CLP = Opss lilted poljiaer R
DOC = Not effective Hhere ground cover Is dense BM
Ex Mireral = EScpanded mineral RT
I - Not inclnarable SS
B = vat. TC
p = Partioalate 1W Fiber
P n Effectlvenass redusjed s*en rainy W
= WllOB
= Hot reusable
= Ranovabillty (rreo batch contactor) is difficult
= Hot effective Where terrain la rossed
= K>t for use wlthta envirotnsnlally sensitive sites
= "Bleated day/treated nattral
= Heated wxxJ fiber
= Effectiveness reduced
-------
Hazards relative to gross inccrapatibil-
ities, health, safety, and environmental
concerns are delineated and the sorbent
density is also given. The format allows
for updating of the Sorbent Data Sheets
as additional sorbents become available.
An example of a Sorbent Data Sheet is
shown in Table 2. Sorption capacity data
for the representative sorbant-hazaidous
liquid pairs are also tabled in Section B.
toother table in Section B contains
hazardous liquid-water preference in-
dices, which are the ratio of the grams
of hazardous liquid absorbed to the grams
of water absorbed by the same quantity
of sorbent. The preference index pro-
vides an indication of a sorbent's per-
formance in a waterspill scenario. The"
larger the hazardous liquid/water pref-
erence index, the greater tie hydro-
phobic quality of the sorbent and the
greater the likelihood that it will
preferentially sorb hazardous liquid in-
stead of water in an aqueous medium.
Section C - "Technical/logistical In-
formation"
This section presents guidance on
equipment and techniques to use for: 1)
the application and collection of sor-
bents in a spill cleanup situation, 2)
the innobilizatLon by sorbents of hazard-
ous liquids for disposal in a landfill,
3) reuse of sorbents, and 4) the disposal
of used sorbents. The text describes
equipment and manpower requirements,
limitations, and safety precautions for
the application and collection of the
three basic sorbent forms: 1) particu-
late, 2) pillow, pad, or mat, and 3)
sorbent boons.
Section D - "Oast Estimation Procedures
andData"
This section provides data for esti-
mating the cost of using a sorbent. A
discussion of general cost categories
for cleanup with sorbents is presented.
The major categories are equipment, ex-
pendable materials, labor, disposal, and
transportation. Cost data are given in
1983 dollars for estimation purposes.
GuMe Sections B and C should be con-
sulted to determine specific equipment,
materials, labor, disposal, and trans-
portation requirements.
Section E - "Test Methods"
Section E describes test methods
available for evaluating sorbent per-
formance characteristics such as sorption
capacity and release rates. Brief des-
criptions of the test methods, including
purpose, applicability, limitations, and
data and resource requirements are pro-
vided. These sorbent test methods are
discussed in the Manual to enable selec-
tion of an appropriate test method if
required. Manufacturers' data on sorption
of hazardous liquids by sorbents generally
do not cite test procedures. The ab-
sence of sorbent performance data derived
from a common test method makes compar-
ison of results difficult. The test
methods cited were developed by the Amer-
ican Society for Testing and Materials
(&STM), General Services A3ministration
(GSA), Vfestinghouse Corporation, Rensse-
laer University, U.S. Coast Guard, Envi-
ronment Canada, and Rockwell Corporation.
Section F - "Spill Scenario Rationale"
Section F presents the rationale for
selecting and defining the four cleanup
scenarios considered in this Manual; 1}
spills onto land, 2) floating spills into
water, 3) non-floating spills into water,
and 4) disposal in landfills. These
scenarios represent situations where sor-
bents are likely to be a viable method of
spill treatment.
Section G - "Sources of Information"
Sources of information on CERCIA
liquids and sorbents are presented in
Section G. References such as reports,
handbooks, publications, and vendors are
identified.
Section H - "CERCIA Liquid Chemical In-
formation"
Section H presents physical property
data on specifi gravity, solubility in
water, viscosity, and surface tension for
the CERCIA liquids. The physical data
tabulation enables identification of
CERCIA liquids with similar physical
properties, which is useful information
for estimating sorption capacity for
similar liquids. The hazardous liquid
specific gravity, in conjunction with
182
-------
Table 2, SORBENT DATA SHEET EXAMPLE
GBM1BIC CUSSs Sorbent Clay
SKD-G, SKG-M
o Oil Dry Corporation
Chicago, IL 312-321-1515
o Costs Participate - $0.05/lb
o Bulk Density! 29-36 Ib/ft3
o Store: Dry
Clean Dpi, Quick Sorb?_ Absorbs-It
Clean U£ J_
o Excel-Hineral Co.
Goleta, CA 805-683.5321
o Cost: Particulate - |0.12Xlb
o Bulk Density: 31-38 Ib/ft3
o Stores Dry
Safestep, Friction
o Andesite of California, Inc
o Costs Partioulate - $O.H9/lb
o Bulk Density: 37 Ib/ft3
o Store: Dry
Lowe's Safety Absorbent
o Lowe's, Inc.
South Ben, In 219-231-8191
o Cost: Particulate $0.0l/lb
o Bulk Densitys 12 Ib/ft3
o Store: Dry
Waverly
o Waverly Mineral Products Co.
Bala Cyiwyd, PA 215-668-2808
o Cost: particulate - '$1.50/lb
o Bulk Density: 30-10 lb/ft3
o Store: Dry
i-« NOTE: Sorbent Clays are commercially available as kitty litter
CO
LIMITATIONS
Landspill
Bot reusable or incinerable, effectiveness reduced when
rainy
Floating Spill -. Non-Flowing Mode
Not for use on spills into water
Not for use on spills into water
INCOMPATIBILITIES
o Acidic compounds, inorganic
o Halides, inorganic
o Cyanates and isocyanates
o Heavy metals
HEALTH/SAFETX CONSIDERATIONS
Wear dust respirator and eye protection when handling
particulate and other protective equipment appropriate
for the ha2ardous liquid
ENVIHQNMENTAI, GOHSIDEMTIOHS
Does not biodegrade, but is a natural component of the
environment
Hon-Floating Spill ' -
Not for use on spills into water
SORBENT DENSITY
137.2 pounds/cubic foot
PERFORMANCE
Sorption capacity data and hazardous liquid/water preference index information are given in Tables B1 and B2, respectively
If the sorbent/hazardous liquid pair is either known, suspected, or observed to be incompatible, an "I" is given in Table B1,
or if the sorbent is not wet by the ha'^-dous liquid, an "NM" is used.
-------
sorbent specific gravity and sorption
capacity, can be used to calculate the
volume of loaded sorbent that will be
generated and will subsequently require
disposal. The data are organized alpha-
betically by chemical class followed by
the hazardous liquid CERCL& name and the'
CAS Registry number.
CONCLUSIONS
A Guidance Manual for the Selection
and Use of Sorbents for Liquid "Hazardous
Substance Releases provides a concise
compilation of information and data per-
taining to the selection, acquisition,
application, collection, regeneration,
and disposal of sorbents. de inclusion •
of a chemical index (by chemical name and
DOT number), which is cross-referenced
to "Sorbent Selection and Use Guides" and
"Sorbent Data Sheets," enables the user
to quickly locate pertinent sorbent in-
formation for the liquid hazardous sub-
stances addressed in the Manual. The
Manual includes 190 sorption capacity
measurements that were made as part of
the project and which were unavailable
in the literature prior to this effort.
Ihe availability of a condensed source
of information on sorbents promotes their
efficient use by assisting decision-
makers to estimate and corpare the effec-
tiveness and cost of various sorbent-use
strategies. Although the primary target
audience for the Manual is Federal On-
Scene Coordinators and their technical
support staffs, the Manual is also
applicable to the needs of spill clean-
up managers representing State and local
agencies as well as the private sector.
REFERENCES
Herridc, E.D., D. Carstea, and
G. Goldgraben. Sorbent Materials for
Cleanup of Hazardous Spills, BPA-600/2-
82-030, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1982.
184
-------
IH SITU TREATMENT OF SODIUM HYDROXIDE SPILLS
IN QUIESCENT WAT1R BODIES -
AH EXPEBIMENTAL PROGRAM
MICHAEL BORST & JOSEPH SHUM
MASON & HANGER-SILAS MASON CO., IMC.
P.O. BOX 117
LEONARDO, HEW JERSEY OT737
ABSTRACT
In situ treatment of soluble hazardous material spills in quiescent water
bodies has been advocated as an alternative to a withdrawal-treatment-replaeement
approach to avoid the necessity of transferring large volumes of water to a treat-
ment facility. To document the practicality of this type of treatment, the U.S.
Environmental Protection Agency (SPA) conducted tests at the Oil and Hazardous
Materials Simulated Environmental Test Tank (OHMSETT) in Leonardo, New Jersey. The
program assumed from the start that the chemistry of the in situ treatment was not
under examinations therefore, the selection of the contaminant and treatment cheni-
cals was based on experimental needs and not requirements of a field (end-use)
situation.
Acid/base chemistry was selected for this testing. The testing used sodium
hydroxide as the contaminant because of cost and relative frequency of actual
spillage. The treatment chemical selected was acetic acid because it is available
in high concentrations, has a relatively low cost, and because the principal reac-
tion product, sodium acetate, is highly soluble at the expected water temperature
and biodegrades,
The first full-scale testing conducted was the generation of a baseline for
comparison of future equipment test results. These used 0.5-m3, 1.0-m3 and 2.0-m3
additions of 5$ solution into the 10,000-m3 test tank. In these tests, the treat-
ment chemical- was injected at the spillage point of the contaminant without applying
external mixing. The results of these tests were used for comparison with the
results of tests using the designed equipment to determine if a substantial reduc-
tion in neutralization time occurred. Design constraints required that the treat-
ment would occur from a small boat and that no special equipment would be avail-
able. The system would require that the treatment chemical be a liquid, water
soluble or slurryable solid. To meet these requirements, a series of mixing and
flow eductors powered from a single'pump was chosen.
Tests were conducted using the mixing eductor in a stationary mode. The tests
showed that the neutralization period within the monitored area was substantially
reduced by using the equipment. Although the results varied from test to test, the
neutralization period was reduced by as much as two-thirds of the baseline. The
tests also demonstrated the importance of considering side reactions of both the
spilled and treatment chemicals with the chemical species occurring naturally in the
waterbody.
185
-------
INTRODUCTION
Insitu treatment of spills of soluble
hazardous materials has an obvious advan-
tage over withdrawal-treatment-replaeement
approaches: no need exists either to es-
tablish a treatment plant or to transport
huge volumes of contaminated water for
treatment. The contaminated waterbody, in
essence, becomes the treatment facility.
Disadvantages also exist. The single
largest disadvantage is the loss of con-
trol and the ability to establish treat-
ment conditions. Volumes or boundaries
are no longer well defined, and treatment
must occur under ambient conditions.
The U.S. Environmental Protection
Agency (EPA) elected to conduct
controlled-eondition tests at the Oil and
Hazardous Materials Simulated Environmen-
tal Test Tank (OHMSETT) facility in
Leonardo, New Jersey, to determine if in
situ treatment of spilled hazardous
materials was feasible and whether the
treatment could be accelerated. The
program required that the treatment would
occur in quiescent water using very highly
soluble chemical species. The treatment
process would be chemical in nature.
Physical treatment such as air stripping
or carbon adsorption was not considered.
Additional criteria were established
for the tests. First, tests would be con-
ducted in the OHMSETT test tank. This
above-ground, concrete tank has a working
volume of 10,000 cubic meters (m3) of salt
•water. An electronic measurement tech-
nique must be available to measure the
concentration of the contaminant. The
water within the tank must meet discharge
requirements at the end of the testing.
Neither the contaminant nor the treatment
chemicals could be radioactive.
Heavy metal salts and acid/base
chemistry were considered for testing.
Soluble heavy metal salts can be treated
with sulfides, carbonates, or phosphates
to form an insoluble precipitate (Drake,
et al., 1976). Treatment of the heavy
metal precipitates would require major
modifications to the existing water treat-
ment plant at OHMSETT. The required
modifications were estimated to cost in
excess of BQ% of the total project funds.
A capital outlay of this magnitude eould
not be justified until some testing had
been successfully completed.
Elimination of heavy metal precipita-
tion left only acid/base chemistry tests.
A brief literature examination revealed
that chemicals of this type were fre-
quently spilled (see Buckley and Veiner,
1978). Research reports suggested several
possible treatment schemes for spills of
this type (Drake, et al.., 1976; Bauer, et
al_._, 1975). Pew treatment attempts were
documented, however.
This EPA program was carried out in
the five steps listed below and described
in the sections following.
1. Assemble data collection system
and develop software
2, Conduct intermediate-scale indoor
tests
3. Conduct full-scale baseline "hot
spot" tests
k. Design and fabricate treatment
chemical application system
5. Evaluate application system at
full-scale.
DATA COLLECTION AND SOFTWARE
An array of pH probes was used in the
intermediate scale and all full-scale
tests to locate the spilled and treatment
chemicals and to obtain the data needed to
evaluate the success of the application
system. The probes were connected to a
data collection interface ISAAC 2000
manufactured by Cyborg Corporation. The
ISAAC included temporary RAM storage, an
analog-to-digital (A/D) converter and a
l6-channel high-speed multiplexer. The
ISAAC was connected to an IBM XT personal
computer. The software was written in Ad-
vanced BASIC and compiled using the BASIC
compiler for the IBM XT.
The probes and transmitters vere
manufactured by Great Lakes Inc. of
Milwaukee, Wisconsin. Each probe con-
sisted of a.glass pH electrode encased in
a. vinyl-ester housing with electronic
amplifiers, a reference cell, and a tem-
perature compensation circuit. The sensor
was connected to a k-20 ma transmitter.
This device changed the form of the signal
from millivolt levels to a current flow
directly proportional to the pH of the
solution surrounding the sensor.
The current loop included a DC power
supply and a 250-ohm precision resistor.
186
-------
A 12-bit A/D converter housed in the ISAAC
measured the voltage drop across this
resistor. The value produced by the A/D
converter was stored in the ISAAC RAM
space until the ISAAC was instructed to
transfer the data value to the IBM XT per-
sonal computer. The data value vas trans-
ferred as ASCII decimal equivalents of the
binary coded data (BCD).
The pH transmitters were zero and span
adjusted. After the analog portion of the
calibration was complete, the entire data
collection system was calibrated. This
was accomplished using a series of buf-
fered solutions from 2 to 11. While the
probe was immersed in a given buffer, data
was logged on the channel of the ISAAC
corresponding to the sensor under
calibration. The logged data was sub-
jected to a least-squares linear regres-
sion to obtain the calibration function
needed to convert ISAAC output (BCD) to
physical units (pH).
IHTIBMEDIATE SCALE INDOOB TESTS
Once the data collection hardware was
in place and the software was written, in-
termediate scale tests were performed.
These tests were conducted to gain ex-
perience with the hardware, observe the
sequence of events in testing and de-bug
the software developed for the program.
In brief, it was felt that prior to the
twenty-million fold increase from beakers
to the full-scale tank tests an inter-
mediate step was appropriate.
For this testing, a children's wading
pool was selected as the waterbody. The
pool held roughly 200 liters of water. A
200-mesh stainless steel screen was
selected to serve as a distribution
manifold. It was hoped that this dis-
tribution method would yield a diffusion
pattern similar to a theoretical line
source. The screen was formed into a rec-
tangular tank and placed into the pool at
the geometric center. Five equiangular
radii were drawn on the bottom with grease
pencil. Three concentric circles were
drawn at 305 millimeter (mm) diametric
intervals. One pH sensor was placed at -
each intersection of the radii and circles
(see Figure l).
The pool was filled with tap water.
After one hour had passed to allow for
energy dissipation from filling, glacial
Figure 1. Systematic sensor layout
used for intermediate scale tests.
acetic acid was siphoned into the cen-
ter of the distribution screen through
a 3-meter (m) long 6-mm ID plastic
hose from a 1000 milliliters graduated
cylinder on a work table 1-m above the
pool surface. Siphoning typically
required 20 to kQ seconds(s).
Several observations resulted
from this testing. First, the five-
pointed star arrangement of sensor
placement worked out nicely. A solid
bottom was fitted into the screen so
that the heavier-than-water acetic
acid did not run through voids between
the screen edge and pool bottom. Ad-
ding methyl red to the acetic acid al-
lowed visual inspection of the
dispersion.
As a result of these tests, it
was decided that full-scale tests
would use the same arrangement for
sensor location, the same screen
materials, and a strong monobasic
chemical as the contaminant chemical.
Acetic acid was selected as the treat-
ment chemical. Hote that acetic acid
is not the preferred treatment
chemical. • It was selected because, of
those suggested (see Drake, et al.,
187
-------
19T6)» it was the cheapest when pur-
chased in bulk in commercial grade.
Sodium hydroxide was selected as the
contaminant based on a cost analysis
Of monobasic compounds with a high in-
cidence of reported spills,
PULL-SCAM BASELINE "HOT-SPOT" TESTS
These baseline tests were two-
step tests: a contamination of the
tank followed by treatment using
natural dispersion of the treatment
chemical. The treatment chemical was
applied at the suspected point of
highest concentration. This approach
was dubbed hot-spot treatment.
When the treatment chemical is
applied, it, like the contaminating
chemical, begins to disperse
throughout the waterbody, with one
major difference. The treatment
chemical reacts with the contaminant
wherever they come in contact. The
contaminant and treatment chemicals
annihilate each other to produce a
non- or less-hazardous substance.
This assumes that (l) the contaminant
and treatment chemicals react only
with each other and (2) the kinetics
of the reaction under the given condi-
tions enable the reaction to proceed
at, as a minimum, the diffusion rate.
This mutual destruction of the two
chemicals creates a large concentra-
tion gradient at the interface of
treated volume and contaminating
volume. The concentration of the con-
taminant in the treated volume goes to
zero, which helps drive additional
contaminant into the already treated
volume to increase the reaction rates.
Similarly the concentration of the
treatment chemical outside the treated
volume is negligible.
In order to establish a con-
taminated waterbody for treatment, the
test tank was intentionally con-
taminated with the sodium hydroxide.
It was decided that so long as there
was a need to contaminate the test
tank, data would be logged to measure
the nattiral dispersion rate of the
sodium hydroxide solutions. For these
tests, nominal 5J5 solutions were
produced by diluting the supplier's
50j£ stock solution with tap water.
The diluted solution was mixed with an
electric mixer and sampled. The
diluted solution was prepared in 0.9-
m3 polyethylene barrels and held for
distribution into the main tank.
These tests were run using 0.5 and
1.0-m3 of the diluted sodium hydroxide.
The time of the sodium hydroxide
addition was designated as "time
zero." Although occurring over a
definite time interval, the contamina-
tions were considered instantaneous
since the distribution period was less
than one data scanning interval (six
minutes}. The application occurred
120 to 150 minutes into the data log-
ging period. This two to two and one-
half hour period was used to obtain
precontamination data and prepare the
test solutions. Data logging con-
tinued at six-minute intervals until
the concentration of sodium hydroxide
in the test volume appeared to
stabilize, as indicated by the pH
measurements.
When the test was concluded (a
somewhat arbitrary decision), data
files were backed up and preparations
were made for the next tests. The
next test consisted of the treatment
of the sodium hydroxide spill with the
acetic acid. For these tests, glacial
acetic acid was transferred to the
holding tanks and then added to the
contaminated water through the dis-
tribution manifold.
Data was collected in the same
manner as during the contamination
test. The time required for the test
volume to return to the original pH
level was regarded as the treatment
time. This step in the testing also
restored the tank to the pretest con-
ditions in preparation for the next
test.
The contamination step in each
test did not provide the desired end-
use data. This step did, however,
give an insight into•dispersion in
quiescent water. The application of
the sodium hydroxide gave pH-vs-time
records as shown in Figure 2 for a ,,•
1.0-m3 spill. The data shows a dis-
persion pattern not predicted by any
of the theoretical models. The two
spikes immediately after the addition
could be attributed to the large tur-
188
-------
bulent eddies from the addition
process. There is a third spike at
approximately 20 hours after the addi-
tion resulting in a pH fluctuation of
0.3 unit. Also note that the baseline
(after complete dispersion had
occurred) is elevated. This is an ex-
pected result in a non-infinite
waterbody. If it were not for the
buffering effect of the salt water,
the final pH would have been 10.1.
pH VS TIME
0.
Sodium Hydroxide Added
40
BO 120
[TNousDnisj
BJPSD WE (SOUNDS)
Figure 2. Typical pH-vs-time record
from contamination of the test tank
with sodium hydroxide.'
The hot—spot treatment tests gave
similar results when the treatment
chemical is considered as an added
contaminant to the "naturally high pH"
waterbody. Typical results from a
151.^-liter glacial acetic acid addi-
tion are shown in- Figure 3. The
single data point collected from each
test couplet is the time required for
the contaminated volume to return to
the precontamination concentration of
the contaminating species as shown by
the arrow in Figure 3.
pH VS TIME
S.O
1 1 1 r
BO 120 160
flhtusnk)
a>PSED TIME (SEC)
T—r
200 240
Figure 3. Typical pH-vs-time record
from hot-spot treatment. The time
required for treatment was 115,000
seconds (32 hours).
Both the contamination and the
treatment test results show pH fluc-
tuations occurring about 20 hours into
the dispersion process and the fluc-
tuations would last for approximately
another 10 hours. At the end of these
tests, a layer of white precipitate
was found at the tank bottom. No
definitive chemical reaction process
was identified for the precipitate.
It is not known if the formation of
the precipitate and its subsequent
reaction with the contaminating/
treatment chemical could have con-
tributed to the pH fluctuations.
Nevertheless, the Observed pH fluctua-
tions may be significant in a real
spill, since aquatic lives are
189
-------
generally more susceptible to fluctuations
than a slowly changing environment.
TREATMENT CHEMICAL APPLICATION SYSTEM
In order to apply the treatment
chemical more effectively, proper amount
of chemical should be placed at the
desired location and mixed vigorously. In
order to achieve this goal, three distinct
steps must occur. First, the average con-
centration of the contaminant in the
volume to "be treated must "be determined.
Second, the proper amount of treatment
chemical must he applied to that volume;
last, the treatment and contaminating
chemicals must be mixed.
The equipment selected to perform
these operations is shown in Figure k.
This equipment consists of four parts.
The sole moving part is the centrifugal
pump. The pump discharge is split into
three lines. One line (U-inch) carries
the majority of the pump discharge to
operate the main mixing eductor. A smal-
ler line (2-in) feeds a smaller eductor,
which draws the treatment chemical from
the treatment chemical storage tank. In
the event that the treatment chemical is a
solid, a ij-inch discharge hose may be
used to feed a small mixing eductor in the
treatment chemical storage tank. This
mixing eductor provides both the water and
mixing energy to dissolve the solid. The
equipment also contains instrumentation to
monitor the flow of treatment chemical.
Measurement of the contaminant con-
centration is clearly necessary for in
situ treatment. In these tests, however,
it was considered to be outside the scope
of the program. The quantity of treatment
chemical needed to neutralize the entire
contaminant spill was metered into the
system at an arbitrary rate, 1.1 m3/hr.
The treatment chemical was glacial acetic
acid diluted to O.i m3 with tap water.
The treatment chemical was educted from
the storage tank and the low concentration
treatment chemical was discharged adjacent
to the intake ports of the mixing eductor.
APPLICATION SYSTEM EVALUATION
Tests of the equipment were conducted
in essentially the same manner as the
baseline tests previously described. The
contaminant was added to the tank through
the distribution manifold and allowed to
disperse while measuring the pH over time.
The next day, the eductor was started to
complete the dispersion of the contaminant
throughout the waterbody. After circula-
tion the eductor was stopped and induced
currents were given a 2-hour dissipation
period.
Pump
•**•
E doctor
Tank
Mixing Eductor
. I—CK3
Main Mixing Eductor
Figure k. Schematic diagram of the equipment for mixing and delivering
the treatment chemical.
190
-------
After the dissipation period the educ-
tor was restarted vith metered addition of
the treatment chemical; acetic acid.
These tests were conducted sequentially.
Figure 5 shows the results for a 1.0-m3 of
5$ sodium hydroxide spill followed by Q.U-
m* of 20$ acetic acid treatment. The data
for the natural dispersion of the con-
taminating chemical has essentially the
same features as obtained in the earlier
"hot-spot" tests. A drop in the pH oc-
curred at 22 hours when the mixing e.ductor
was- started to accelerate the dispersion
of the contaminant prior to treatment.
The mechanical mixing from the eductor
during treatment results in an immediate
pH drop followed "by a steadily rising pH.
The treatment results with external mixing
do not show the spikes characteristic of
the "hot-spot" tests. The time required
for complete treatment is dramatically
reduced, as shown in Figure 5» from 32
hours to 11 hours (Figure 3).
pH VS TIME
9.0'
as
as
' u-
&%•
5,0
71
7.6
7.4
73
7.0
as
as
a<
az
ao
Acetic Acid Added
40,000s
Sodium Hydroxid
Added
0 20 40 BO BO 100 120 140 1BO
(Itausxxic)
Figure 5. Typical pH-vs-time record for
neutralization test using the treatment
equipment. The treatment period for this
test was 1(0,000 seconds (11 hours), 35? of
the time required for hot-spot treatment.
Figures 6 and 7 show typical pH-time
plots for sensors located along the same
radial direction and at the same circum-
ference respectively. The chemicals used
in this test were approximately 33% of the
quantities used in the other tests. The
end point pH shows the test did not run
sufficiently long enough to reach the
final equilibrium condition.
pH VS TIME
20
40 60 80
BARS) THE (SEC)
Figure 6. Typical pH-vs-time record for
sensors 13, lk and 15 (see Figure l).
CONCLUSIONS
The study demonstrated the feasibility
of insitu treatment using the designed
application apparatus. Although testing
was limited to acid/base chemistry, the
study nonetheless presents a treatment al-
ternative for one type of frequently
reported spills. The application ap-
paratus gave a greatly reduced neutraliza-
tion period. The heart of the data
gathering system, the pH probes and
transmitters, functioned well. These sen-
sors and transmitters can be powered by a
portable DC power supply such as automo-
191
-------
tive batteries. They -will operate over a
long time span 'because of the limited
power drain. They vill readily lend them-
selves to field use and are light enough
for small boat use.
The test data shovs the natural dis-
persion pattern departs from the theoreti-
cal model prediction. The observed pH
fluctuations may be significant in a real
spill and response situation. Before ap-
plication of the treatment chemical the
possibility of competitive side reactions
should be evaluated at a small scale.
Samples of uneontaminated water should be
mixed with, both the contaminating chemi-
cals and treatment chemicals to determine
this possibility. The thermodynamics of
the desired neutralization must be con-
sidered under ambient conditions.
pH VS TIME
oil incidents, there is limited field
applicability. The faster-than-expected
dispersion measured in this testing indi-
cates that in reasonably high-energy-level
water bodies, the base would have time to
diffuse to the natural limits prior to as-
sembling treatment chemicals and equip-
ment. Use of the apparatus to spills
other than acid/base may require a more
versatile monitoring system to be
developed. Several additional methods for
in situ treatment of hazardous material
spills into waterways have been suggested.
Precipitation and chelation, for example,
have been theoretically proposed but have
not progressed beyond the preliminary
laboratory theoretical studies. A proven
ability to advance to full scale testing
of these concepts and the mechanical
equipment necessary for implementation and
application of the theory now exists.
Let's get started!
BEPEB1MCBS
Bauer, W. H., 1. Drake, D. Shooter, W.
Lyman, L. Davidson. Agents, Methods
and Devices for Amelioration of Dis-
charges of Hazardous Chemicals on
Water, U.S. Coast Guard Beport No. CG-
0-38-76, August 1975, 28k p.
Drake, E., W. H. Bauer, D. K. Borton, and
J. J. Buloff. A Feasibility Study of
Response Techniques for Discharge of
Hazardous Chemicals that Disperse
Through the Water Column, II. S. Coast
Guard Report No. CG-D-16-T7, July
1976, 255 p.
Buckley, J. L., and. Wiener S.A., Hazardous
Material Spills: A Documentation and
Analysis of Historical Data, U.S. En-
vironmental Protection Agency Report
No. 600/2-78-066, 2l*3 p.
no
Figure 7« Typical pH-vs-time record for
senaors 1, k, 1, 10 and 13 (see Figure l),
RECOMMENDATIONS
These tests were, by design, chemi-
cally simple. Although spills of acids
and bases are among the most common non-
192
-------
CHEMICAL AND M1CROBIAL STABILIZATION TECHNIQUES FOR REMEDIAL ACTION SITES
Janet lickabaugh, Sara Clement, Janette Martin, Mark Sunderhaus
University of Cincinnati
Cincinnati, Ohio 45221
Ronald P. Lewis
QRB-HWERL-US EPA
Cincinnati, Ohio 45268
ABSTHACT
Highly contaminated soils at remedial action sites may require multiple technologies
for successful on site clean-up. One such example is the Chem-dyne site in Hamilton,
Ohio. The site is located above a major aquifer which provides the water supply to many
communities in Southwestern Ohio. While this site was active, a large variety of chemi-
cals were received in various containers. Subsequently, leaking containers and spillage
caused heavy contamination to the soil in many locations on the site. Soil collected from
one location on the site which contained hydrocarbons, chlorinated hydrocarbons, and
organo-chloride pesticides was used in studies of biodegradation, surfactant scrubbing,
followed by photolysis, and reverse osmosis. This paper discusses each technology and
presents preliminary results from each study.
INTRODUCTION
Soils at remedial action sites can be
contaminated with a variety of compounds
which will respond in different ways to
clean-up techniques. Therefore, single
stabilization techniques can reasonably be
expected to produce somewhat incomplete
results. The goal of this study was to
look at several technologies which could
be used in series and/or parallel on a
single site for maximum destruction of
contaminants. Soil was collected from the
Chem-dyne site in Hamilton, Ohio and sub-
jected to laboratory evaluations of con-
taminant removal by biodegradation, sur-
factant scrubbing followed by photolysis,
and reverse osmosis treatment of leachate.
Two batch studies of biodegradation com-
pared the native miorobial population to
bioaugmentation products under mixed oxy-
gen conditions for removal of pesticides
and chlorinated hydrocarbons from soils.
Surfactant enhanced scrubbing was examined
as a technique to improve aqueous separa-
tion of organic contaminants from the
Chem-dyne soil in batch studies and
continuous flow through studies. Surfact-
ant solutions were subjected to photolysis
for destruction of chlorinated hydrocarbon
compounds. Leachate obtained by passing
water through Chem-dyne soil was subjected
to reverse osmosis to evaluate its use as
a final polishing step to follow other
clean-up techniques.
EXPERIMENTAL
1• Biodegradation
Batch studies were set up in 14 shal-
low pans with 1 to 1.5 kilograms of soil
to simulate the first 25.4 to 50.8 mm of
tilled soil. Four units were aerobic for
6 weeks followed by limited oxygen condi-
tions for another 6 weeks; a second set of
4 units had the conditions reversed, while
the remaining 6 units were kept under
anaerobic conditions for the full 12
weeks. Half the samples under each oxygen
condition contained only native microbial
cultures while the other half had bioaug-
mentation products added on a weekly
193
-------
basis. Sufficient nitrogen and phosphor-
ous were added to produce a carbon:
nitrogen: phosphorous ratio of 100:15:3»
and moisture content was held between 10
and 20? or approxinately 80J of field
capacity. The systems were monitored
weekly for a variety of parameters,
including standard plate counts and
chlorinated hydrocarbons and pesticides by
gas chromatography (GO. In addition,
ohromatography/mass spectometry (GC/MS)
was performed initially and twice during
each study to identify the contaminants
and any products of microbial degradation.
Hiorobial Results. Standard plate counts
(SPC) were used to monitor bacterial
growth in a general fashion without iden-
tifications for specific organisms. Fig-
ure 1 shows the microbial activity under
mixed oxygen conditions for both the na-
tive populations and the bioaugmentation
population. Due to the weekly addition of
the bioaugmentation culture, the popula-
tion in these test units had 1 to 2 log
greater SPC than the native culture units.
Anaerobic conditions produced approxi-
nately $ log lower SPC than mixed oxygen
conditions for both the native and bioaug-
mented miorobial cultures.
Heekly GC analyses following soxhlet
extractions indicated that some chemicals
were reduced better than others. Figure 2
shows the behavior of 1,2-dichlorobenzene >
typical of the chlorinated hydrocarbons
during this study. Efforts made to obtain
representative soil samples from the test
pans were not always successful as indi-
cated by somewhat erratic GC data. How-
ever, in general the results indicate that
the native population performed as well or
better than the bioaugwentatlon products
even with lower plate counts. The oxygen
conditions tended to have produced the
major differences. Aerobic conditions
followed by limited oxygen resulted in
more pollutant decline than anaerobic or
limited oxygen conditions followed by
aerobic conditions.
Table 1 shows compounds and mean con-
centrations found by GC/MS. In general,
the GO/MS data confirmed the decline of
contaminants as determined weekly by GC
analysis, with reductions from initial
concentrations ranging from 3656 to 9B% for
individual compounds. GC/MS identifica-
tions were also made for other chemicals
not monitored by GC which were potentially
miorobial reaction products. The quantity
of these was low but they should not be
ignored as long as incomplete degradation
is a possibility.
2. Surfactant Scrubbing and Photolysis
This research examined the effective-
ness of aqueous surfactant solutions in
removing chlorinated hydrocarbon contami-
nants from Chent-dyne soils, and the use of
photolysis to decontaminate the surfactant
solutions in lab scale tests.
Thirty-eight commercial surfactants
were collected, including anionic,
nonionic, and cationic types and blends.
Fourteen were selected on the basia of
good water solubility, near neutral pH (to
minimize soil adsorption), and generally
low chloride content (except the
cationic).
These 1^ surfactants were tested in
lab scale batch and flow-through column
soil washing experiments. In batch shake
tests 10 gm of contaminated soil was shak-
en for several hours in 250 ml bottles
with 200 ml of 1? and 2% surfactant solu-
tions, allowed to sit overnight, filtered
and decanted. The contaminated surfactant
solution was liquid-liquid extracted,
concentrated and analyzed for chlorinated
hydrocarbons by gas chromotography. Ini-
tial soil duplicates were soxhlet extract-
ed and analyzed so percent removals could
be calculated. The surfactant wash solu-
tion were also subjected to TOX (total
organic halogen) analysis as an indirect
method of monitoring chlorinated hydrocar-
bons in the wash solution.
The surfactants that performed best in
the batch studies were selected for test-
ing in flow-through columns filled with 50
gm of Chem-dyne soil. Two anionic sur-
factants (Suroo 60T and Biosoft N300), two
nonionio surfactants (Makon 10 and Triton
DF165, one cationic surfactant (Emcol CC95
and several blends of these surfactants
(nonionlc-cationic and nonionic-anionio)
were tested in the column studies. The
columns were washed with 500 ml of a 21
surfactant solution at an average rate of
59 ml per hour for a period of one week
and then rinsed with 250 ml of distilled
water. The contaminated surtactant plus
the rinse was liquid-liquid extracted and
analyzed by GC for cholorinated hydrocar-
bons. The washed soils and duplicate
unwashed soils were soxhlet extracted and
analyzed by GC for chlorinated
hydrocarbons.
Two blends of nonionic-cationic sur-
factants (Makon-Emcol) and (Triton-Emcol)
were used to wash similar soil columns for
194
-------
T3
O>
c
3
O
O
en
O
13.00
12.SO -
12.00 -
11.50 -
11.00 -
10.50 -
10.00 -
B.SO -
9.00
O Limited 02/Aerobic
Aerobic/Limited
S-
-a
3
O
O
cn
O
13.00
12.80 -
12.00 -
11.50 -
11.00 -
10.50 -
10.00 -
B.S0 -
9.00
Limited
+ Aerobic/Li mi ted 0
20
60
40
Days
B. Bioaugmented
FIGURE 1, STANDARD PLATE COUNTS UNDER MIXED OXYGEN CONDITIONS
195
-------
7000
6000
n Limited Qg/Aerobic
+ Aerobic/Limited O
60
80
Days
A. Native
7000
6000
p Limited 0,,/Aerobic
•*• Aerobic/Limited 0
60
80
Days
B. Bioaugmented
FIGURE 2, 1,2-DICHLQROBENZENE UNDER MIXED OXYGEN CONDITIONS
196
-------
Table 1. Mean Concentrations By GC/MS (ug/Kg dry soil)
Initial Limited Oxygen then Aerobic Aerobic then Limited Oxygen
native bioag native bioag
Compound 6 wk 12 wk 6 wk 12 wk 6 wk12 wk 6 wk 12 wk
1 , 2-Diehlorobenzene
1 ,3-Diehlorobenaene
1 ,4-Dichlorobenzene
Hexachlorobenzene
Pentaehlorobenzene
27600
12JIOO
15300
20600
6910
5970
7280
5750
5890
815
2960
4270
2900
3380
*
5870
6830
6590
5310
1150
6090
7350
6380
534 0
*
6000
7380
6600
6880
1120
6340
6790
7130
5160
*
5900
6850
6440
5610
2450
6760
9280
6850
5090
*
» Not detected
a period of three weeks. The columns were
filled with 100 gm of soil at the begin-
ning of the experiment. The soil was
washed by 500 ml of 2% surfactant solu-
tions at an average rate of 61 ml/hr. At
the end of each week approximately 25 gm
of soil was removed, soxhlet extracted,
and analyzed for chlorinated hydrocarbons
by GC. The initial soil was also soxhlet
extracted and analyzed by GC. The contam-
inated surfactant solution and 250 ml
distilled water rinse was analyzed by GC
following liquid-liquid extraction. Each
week fresh surfactant solution was used to
wash the soil columns.
To investigate photolysis as a means
of decontaminating surfactant solutions,
samples of contaminated Biosoft (anionie)
and Emcol (cationio) were split into two
aliquots after soil washing. One aliquot
was photoreacted for 24 hours,-liquid-
liquid extracted and analyzed by GC for
chlorinated hydrocarbons. The other
aliquot was extracted and analyzed without
photoreactlon for comparison.
Surfactantand Photolys isHesults. The
results from these experiments showed that
the surfactant solutions removed signifi-
cantly more chlorinated hydrocarbons from
Chem-dyne soil than distilled water alone.
Table 2 shows data from soxhlet extrac-
tions of washed soils in the column
studies, where total removals of all tar-
get compounds ranged from 28 to 5956, with
the maximum found using a nonionic-
cationic blend (Triton-Emcol}. Blends and
Makon 10 alone gave the best removals.
Batch studies showed that 2% surfact-
ant solutions were more effective in re-
moving chlorinated hydrocarbons than
either 1? or 0.555 solutions. The TOX data
showed that 251 solutions removed an aver-
age of 90 rag/kg of chlorinated hydro-
carbons from the soil, while the 11 solu-
tions removed an average of 62 mg/kg, and
the 0.536 solutions removed an average of
30 mg/kg. Gas ohromotography data from
batch studies confirmed these results.
Table 2. One Week Column Study Removals
Soil Soxhlet Extractions
Type
Antonio
Anionic
Nonionio
Nonionic
Cationio
Blend
Blend ' ,
Blend
Blend
Distilled Water
Name
Biosoft N300
Surco 60T
Makon 10
Triton DF16
Emcol CC9
Triton-Biosoft
Triton-Emcol
Makon-Suroo
Makon-Emcol
% lemoval
42
40
55
31
28
50
59
46
44
0.3
Nonionic-cationio or nonionic-anionic
blends were generally slightly more effec-
tive in removing chlorinated hydrocarbon
contaminants than most surfactants used
alone. Column studies show that anionic
and nonionic surfactants were similar in
their removals of contaminants, while
cationics used alone were generally less
effective. Average percent removals for
various types of surfactants are shown in
Table 3.
Table 3. Average Percent Removals
Type
Anionic s
Nonlonios
Cationic
Blends
Batch & Column
(liq.-liq. ex.)
16.2
10
16
22.5
Column
(sox. ex.)
41
43
28
50
Soxhlet extractions of soils in the
column study gave higher percent removal
197
-------
values than liquid-liquid extraction data
from the same soil washings. Problems
were encountered using the liquid-liquid
extraction procedure with the 1)6 and 2%
surfactant solutions, and the liquid-
liquid extraction efficiencies for stand-
ards in surfactant solutions were erratic.
Therefore, data from the soxhlet extrac-
tions were considered to be more reliable.
Preliminary results show that photoly-
sis is effective as a means of decontami-
nating surfactant solutions. Twenty-four
hour photolysis under two 15 watt germaei-
dal lamps destroyed 76% of the 11 analyzed
contaminants in an anionic surfactant
(Biosoft) and 1QOJC of these contaminants
in a oationic surfactant (Emcol).
The data from this study indicate that
in general the following conditions tend
to promote greater removals of chlorinated
hydrocarbons from Chem-dyne soils:
15 A higher ratio of surfactant
solution to soil weight washed.
2) Increased washing time (Up to three
weeks).
3) Higher surfactant concentrations
(Up to 2J6).
Three weeks of soil washing in the columns
showed the greatest removal during the
first week of washing. After the first
week, removal continued but at a slower
rate,
3. Reverse Osmosis
Reverse osmosis was selected for eval-
uation as a potential final polishing
process prior to liquid discharge from a
hazardous waste site. Thus the goal of
this part of the study was to determine
the effectiveness of reverse osmosis
treatment in removing chlorinated hydro-
carbons and pesticides from a hazardous
waste leachate. Two different spiral
wound membranes were used in the study,
and comparisons were made of their removal
capabilities.
The reverse osmosis unit was a pilot
scale OSMO 1919-SB manufactured by
Osmonics. Inc. operated at 190 psi. Both
membranes were manufactured by Osmonics,
each having approximately 19 square feet
of surface area. The first membrane used
was a cellulose acetate (CA) membrane
model SEPA-97E. The system could operate
over a pH of 6 to 8, and at approximately
22% permeate recovery when using this
membrane. The second membrane used was a
polyamide thin film composite membrane
(PA) model SEPA-PA. The system could
operate over a much wider pH range (5-12).
with a permeate recovery of 1)4?.
Batch runs were conducted using both
membranes in single, double, and triple
pass modes of operation. A total of nine
experiments were run, four with the CA
membrane, and 5 with the PA membrane.
Table 4 shows a summary of the experiments
and their feed streams. The column leach-
ate indicated In the table was generated
by passing water through 5 ft columns
containing 60 Kg of soil from the Chem-
dyne site. The leachate contained low but
significant concentrations of various
chlorinated hydrocarbons and pesticides.
Each stream was extracted and analyzed for
ten specific chlorinated hydrocarbons and
pesticides along with chemical oxygen
demand (COD) and total organic carbon
(TOC).
Table
RO Experimental Design Summary
No.
1
2
3
H
5
6
7
8
9
Membrane
CA
CA
CA
CA
PA
PA
PA
PA
PA
Peed Stream
Column Leachate
Concentrate from Exp.
Concentrate from Exp.
Permeate from Exp. 1 I
Column Leachate
Concentrate from Exp.
Column leachate
Concentrate from Exp.
Concentrate from Exp.
1
2
I 2
5
7
6 & 8
Reverse Osmosis Results. Experiments with
the CA membrane showed mass rejection of
individual compounds from 71% to greater
than 9855. Data from experiment 2 are
shown in Table 5. The same relative re-
moval tendencies were found for each pass
through the system. The percent removal
of those compounds having molecular weight
greater than 200 was generally lower than
for those with a molecular weight greater
than 200. The removal of COD and TOC was
high (>98?). These gross parameters
indicate that the membrane was efficient
at removing compounds not directly
analyzed in the study. It is important to
note that even though 71$ to 988 of the
organics were removed, the concentrations
remaining in the permeate from each of the
CA runs were still significant and the
permeate could not be discharged.
The PA membrane showed much better
removal of the compounds studied, as shown
in Table 6 for experiment 6. The concen-
trations of all compounds present in the
feed were reduced to below detectable
198
-------
Table 5. Reverse Osmosis Ixp. 2, CA Membrane
COMPOUND
1,3-DCB
1,2-DCB
1,3,5-TCB
1,2,3-TCB
Hexachlorobutad ie ne
Hexachlorobenzene
Heptaohlor
Heptachlor Epoxide
Dieldrln
Endrin
COD (mg/1)
TOO (mg/1)
Volume (1)
BDL — Below Detectable
Table 6
COMPOUND
1,3-DCB
1,2-DCB
1,3, 5-TCB
1,2,3-TCB
Hexaehloro butadiene
Hexachlorobenzene
Heptaohlor
Heptaohlor Epoxide
Dieldrin
Endrin
COD (mg/1)
TOC (mg/1)
Volume (1)
CONCENflATION
Feed
36.6
50.0
11.3
12.8
227.
11.6
9.77
41.6
2.80
1.45
259.
134.
112.
Limit
Reverse
Cone.
40.0
16.3
8.18
6.36
59.7
9.01
4.75
32.5
3.76
1.75
342.
98.0
86.2
Osmosis
CONCENTRATION
Feed
49.8
12.6
10.2
10.7
161.
15.6
19.6
92.0
3.16
0.88
475.
191.
87.7
Cone,
85.5
16.2
17.7
23.3
323.
16.7
17.0
94.1
6.64
1.44
484.
302.
49.2
Cug/1)
Perm.
30.1
15.8
8.72
8.13
93.7
2.50
1.02
5.98
0.29
BDL
13.2
0.0
26.0
Exp. 6, PA
(ug/1)
Perm.
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
BDL
100.
41.3
38.5
REMOVAL
(«)
81.0
92.7
82.1
85.3
90.4
95.0
97.6
96.7
97.6
>98.4
98.8?
100. %
Membrane
REMOVAL
It)
>98.9
>97.9
>99.1
>98.8
>99.8
>99.4
>99.8
>99.6
>98.6
>95.0
90.9?
90.5*
CONG. RED.
(?)
17.8
68.3
22.8
36.5
58.7
78.5
89.6
85.6
89.6
>93.1
94.9*
100. %
CONC.RED.
(?)
>97.4
95.2
>97.1
>98.1
>99.6
>99.4
>99.5
>99.8
>96.8
>88.6
78.9$
78.4$
limits in the permeate for experiments 5
to 9 (the detectable limit was used to
calculate the percent removal for those
cases). The chromatograms of the feed and
permeate for experiment 6 are shown in
Figure 3. A comparison of these two chro-
matograms illustrates the fact that the PA
membrane effectively removed not only
target compounds, but others as well.
When using the PA membrane, the re-
verse osmosis unit operated at approxi-
mately 44? permeate recovery. After three
passes through the system, the leaohate
was reduced to 17? of its original volume.
The concentration of chlorinated hydro-
carbons in the remaining 83? was below
detectable limits and thus could be
discharged.
The polyamide membrane was found to be
superior to the cellulose acetate membrane
for the removal of chlorinated hydrocar-
bons and pesticides present in this leach-
ate. The PA membrane was found to produce
a higher quality permeate, while reducing
the leaehate to a smaller volume. In
general, the CA membrane does not produce
a. permeate of sufficient quality to war-
rant its use for the decontamination of
this particular leaehate. The PA membrane
is very effective in a multiple pass
system for both volume reduction and
removal of the contaminants from solution.
199
-------
a. Feed
b. Permeate, PA membrane
Figure 3. Chromatograms of Reverse Osmosis Feed and Permeate
SUMMARY OF FINDINGS BIBLIOGRAPHY
These studies show that microbial
degradation is a viable remedial action
technology for this type of soil, although
the process is slow. Microbial degrada-
tion tends to work best in the soil under
conditions siniliar to soil tilling: aero-
bio then limited oxygen and mixing. Sur-
faotant scrubbing is able to remove a high
percentage of pollutants from the soil,
but requires another technique, photolysis
in this case, to destroy the contaminants.
The photolysis experiments were very sue-
cessful at destroying most of the contami-
nants (54 to 98J) in the surfactant solu-
tions. And finallyt reverse osmosis is
the most successful of the alternatives
evaluated at removing contaminants from a
water stream and concentrating them in &
relatively small volume of water, but
requires another technology to destroy the
contaminants in the waste stream. In
conclusion, the results of the studies
tend to indicate that no single technology
could remove or destroy all the contami-
nants, but used together, nearly total
destruction could be attained for the
contaminants on site. Further evaluations
of these and other technologies are in
progress.
Botre, C. et al., "TCDD Solubilization and
Photodecomposition in Aqueous Solutions,"
Environ. Sol. A Tech., 12, #3, pp. 335-
336, 1978.
CH?M Hilli Final RemedialInvestigation
Report, fol.2, Chem-dyne Site, 1984.
Chian, E.S., Fang, H.P.P., and Bruce,
W.H., "Removal of Pesticides by Reverse
Osmosis," Environ. Sci. &Tech., 9, pp.
52-59, 1975.
Huibregste, K.R, et al., "Development of a
Mobile System for Extracting Hazardous
Materials from Soils," in Controlof
Hazardous Materials Spills. Proo. 1980
Conf., Vanderbilt & EPA, 1980.
Kofayassahi, H. and B. Rittmann,
''Microbial Removal of Hazardous Organic
Compounds," Environ. Sol. & Tech.. Vol.
16, No. 3, 1982.
Evaluation of Systems to Accelerate
Stabilization of Waste Piles or Deposits,
JRB Asaoc., Draft Report for EPA contract
No. 68-03-3113, 1983.
200
-------
TREATMENT OF SOILS CONTAMINATED WITH HEAVY METALS
William D. Ellis, Thomas R. Fogg
Science Applications International Corporation
8400 Westpark Drive
McLean, VA 22102
Anthony N. Tafuri
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, NJ 08837
ABSTRACT
The U.S. Environmental Protection Agency's Hazardous Waste Engineering Research Lab-
oratory has initiated a program to evaluate in situ methods for mitigating or eliminating
environmental damage from releases of toxic and other hazardous materials to the soils
around uncontrolled hazardous waste disposal sites. As part of this program, various re-
agents suitable for the in situ washing of heavy metal contaminants from soil were tested
at laboratory scale. The workwas performed on a soil from an actual Superfund site near
Seattle, WA. The soil contained five toxic heavy metals often found In hazardous waste
site soils: cadmium, chromium, copper, lead, and nickel.
The tests demonstrated that sequential treatment of soil with ethylenediaminetetra-
acetic acid (EDTA), hydroxylamine hydrochloride, and citrate buffer was effective in-re-
moving metals from soil, and all were necessary for good cleanup. The EDTA chelated and
solubillzed all of the metals to some degree; the hydroxylamine hydrochloride reduced the
soil iron oxide-manganese oxide matrix, releasing bound metals, and also reduced insoluble
chromates to chromium (II) and (III) forms; and the citrate removed the reduced chromium
and additional acid-labile metals. The best removals observed were: cadmium, 98 percent;
lead, 96 percent; copper, 73 percent; chromium, 52 percent; and nickel, 23 percent.
INTRODUCTION
The U.S. Environmental Protection
Agency's (EPA) Hazardous Waste Engineering
Research Laboratory (HWERL) initiated a
program to develop in situ chemical methods
for mitigating or eTTmfnating environmen-
tal damage from releases of hazardous ma-
terials at chemical spill sites and around
hazardous waste disposal sites. As part
of this program, Science Applications
International Corporation (SAIC), under
EPA Contract No. 68-03-3113, investigated
chemical methods for in situ cleanup of
heavy-metal -contaminated soil.
Toxic heavy metals are frequently
found in soil at uncontrolled hazardous
waste sites, including lead (15 percent of
sites surveyed), chromium (11 percent),
cadmium (8 percent), and copper (7 percent)
(Ellis and Payne, 1983).
201
-------
Conventional remedial methods for
sites containing heavy metals include
excavation followed by land disposal and
groundwater pumping and treatment. The
use of excavation and land disposal is
meeting with Increased opposition not only
because of high cost but also because the
contaminated soil is simply transferred to
another location. Also, pump and treat-
ment methods are costly and are not effec-
tive for removing contaminants sorbed to
the soil. In situ treatment of toxic
metals in soTl and groundwater offers a
potentially cost-effective remedial
alternative. However more research is
needed before the 1n situ methods can be
implemented in theTield.
The objective of this project was to
select the most promising In situ treat-
ment method for metals and evaluate the
method through laboratory studies. The
study was limited to methods suitable for
1n situ treatment of cadmium (Cd), chrOj-
mTum(Cr), copper (Cu), lead (Pb), and
nickel (N1). These metals are found
frequently at hazardous waste sites and
are among the most toxic. Methods that
are effective with these metals might
also be suitable for treating other heavy
metals found at hazardous waste sites.
Potential in situ treatment methods
for metals 1nclif3e methods that immobilize
the metals in soil by means such as preci-
pitation and methods that solubilize and
remove the metals from the soil. Methods
that solubilize and remove the metals
offer an advantage over immobilization
methods because the need for long-term
monitoring 1s eliminated. Immobilization
methods, on the other hand, simply reduce
the concentration of dissolved species.
The potential exists for resolubilization
of the metals through subsequent natural
chemical reactions; therefore, the site
must be continually monitored.
Methods for mobilizing metals in
soils Involve the use of dilute weak
acids, bases, or aqueous solutions of
chelating agents. Considerable research
on a laboratory scale has already been
conducted on the use of chelating and
other complexing agents for selectively
removing metals from soil.
This research demonstrated different
degrees of extractability of any given
heavy metal from soil. The extractability
has been described according to which type
of extraction agent will remove the bound
metal which corresponds to a specific soil-
metal binding mechanism or the chemical
state of the metal. For example, soluble
heavy-metal salts are extractable with
water; metals bound to the soil organic
fraction are extractable with aqueous
alkaline buffers such as tetrasodium
pyrophosphate ("tetrapyrophosphate"); and
metals occluded in the iron and manganese
oxide fraction of the soil are released
by reduction of the oxides with hyroxy-
lamine hydrochloride. These techniques,
if developed further, could be used for
the cleanup of contaminated soil at
hazardous waste sites.
Laboratory Task Description
Laboratory studies were conducted to
determine whether in situ cleanup of heavy-
metal-contaminated soil by treatment with
chelating solutions or acidic buffers was
possible. The soil used in the studies
was collected from the Western Processing,
Inc. Superfund site, near Seattle, WA.
Previous analysis of this soil (Repa, et
al, 1984) had shown high levels of cadmium,
chromium, copper, and lead (>10 ppm).
The laboratory task consisted of:
(1) soil characterization; (2) laboratory
equilibration (shaker table) experiments
designed to evaluate treatment methods
(i.e., single agent treatment vs sequential
treatment with several agents) for metal
removal; and (3) soil column tests to
evaluate cleanup efficiency under gravity
flow conditions.
Based on a review of the literature,
the chelating agent ethylenediaminetetra-
acetic acid (EDTA), the reducing agent
hydroxylamine hydrochloride, and the
acidic citrate buffer were identified as
suitable agents for testing. Shaker table
equilibration studies were conducted in
which various combinations of the above
treatment agents (10:1 w/w agent solution:
soil), either singly or in sequence, were
shaken with the contaminated soil 1n a
closed container on a vibrating platform.
202
-------
Based on these results, an optimum
treatment sequence was designed. Then
column tests of the optimum treatment
sequence were conducted.
The column studies evaluated metal re-
moval under gravity flow conditions, with
analysis of soil and duplicate analysis of
leachate after each treatment. A three-
agent sequential extraction was tested
using five pore volumes of the optimum
concentration and pH fop the EDTA solution
to remove most metals, followed by hydrox-
ylamine hydrochloride to reduce any hexa-
valent chromium to trivalent, and to
reduce any soil iron or manganese oxides
to release any bound metal. Citrate
buffer was then used as a final acidic
leaching agent. The same metal-contamin-
ated soil was used for all tests; all
initial concentrations for each metal were
the same (see Table 1).
Samples were analyzed for trace ele-
ments by atomic adsorption spectrophotom-
etry (AAS) using flame or graphite fur-
nace procedures. Analyses by the method
of standard additions were routinely
performed along with standard calibrations.
When the two calibration curves deviated
significantly, calculations of sample
concentrations were based upon the stan-
dard addition calibration; when they were
the same, a combination of the standard
addition/standard calibration was used.
Sample blanks and National Bureau of Stan-
dards (NBS) standards were analyzed in the
same manner as the samples.
TABLE 1. SINGLE AGENT SHAKER TABLE EXTRACTION EFFICIENCIES
Soil Metals (ppm)
Cd
Pb
2,480
EDTA (0.1 M & pH 6)
% Extracted
114
24
62
14
106
Hydroxylamine hydrochloride
(0.1 M 1n acetic acid)
% Extracted 86
32
43
20
Citrate buffer (0.1 M @ pH 3)
% Extracted 77
24
48
14.5
65
Pyrophosphate (0.1 M)
% Extracted
5.4
9.6
29
2.9
9.7
DPTA (0.005 M in
0.1 M trlethanolamine)
% Extracted
59
48
67
203
-------
RESULTS AND DISCUSSION
Soil Characterization
Soil permeability measured 1n the
laboratory was approximately 5 x 10~5
cm/sec. The grain size distribution was
determined by wet and dry sieve procedures
and plpet analyses on organic-free soil
after a hydrogen peroxide wash. Approxi-
mately 75 percent of the soil was in the
silt and clay range. This probably caused
the rather slow percolation rate. X-ray
diffraction analysis showed alpha-quartz
and feldspar to be the only measurable
constituents of the soil. No measurable
amounts of crystalline aluminum oxide
forms were present. The total carbon
content of the soil averaged 16,400 + 709
ppra by weight (1.64 percent). This Tnter-
medlate level of carbon corresponds to the
phenols and other organic compounds found
in the soil.
The cation exchange capacity (CEC) of
the soil was also determined. The results
were 13 and 8.2 mllliequivalents per 100 g
for bulk and organic-free soil, respec-
tively. These results are quite low and
Indicate an absence of mineraloglc clay 1n
the soil. The pH and Eh measurements
(made in triplicate) yielded an average
soil pH of 7.39 and an Eh of +0.198 v
(electron potential, pe = +7.01), reveal-
Ing a neutral, slightly oxidizing soil.
The iron and manganese oxide mean concen-
trations were 15,000 and 291 ug/g, respec-
tively. The carbonate results yielded an
average value of 1.42 meq/g as bicarbonate.
The results of the determination of
heavy metals of Interest 1n Western Pro-
cessing soil were as follows (in ug/g):
cadmium (47), chromium (349), copper
(219), Iron (30,200), manganese (1,690),
nickel (214), and lead (2,480). These
values were compared with the concentra-
tions of the metals 1n the treatment
solution to assess percent removal of
metals by the treatment.
Shaker Table Studies
In the single shaker table extractions
using EDTA at different concentrations and
pH values, the 0.1 M solution was much
more effective 1n metal removal than the
0.01 M solution. The pH trends, however,
were not so clear cut. A pH of 6 was
chosen as the optimum because it afforded
slightly better chromium removal than that
obtained at pH 7 or 8; EOTA is more ion-
ized at pH 6. This pH and concentration
combination was used in subsequent
studies.
The results of the EDTA, hydroxylamine
hydrochloride, acidic buffer, and dlethyl-
enetriamine pentaacetic acid (DTPA) single-
method shaker table extractions (Table 1)
showed that EDTA was the best single
extraction agent for all metals. However,
hydroxylamine hydrochloride was more
effective at chromium extraction.
Results of the two-agent sequential
extraction (Table 2) indicated that the
EDTA was much more effective in removing
metals than the weaker agents often used
to characterize the mechanism of binding
of metals to soils. Thus, weaker extrac-
tion techniques (magnesium chloride,
potassium fluoride, acetate buffer, tetra-
pyrophosphate) can be eliminated if just
an EDTA solution is used.
The results of the three-agent sequen-
tial extraction studies (Table 3) showed
that, compared to bulk untreated soil,
this extraction scheme removed nearly all
the lead and cadmium, 73 percent of the
copper, almost 52 percent of the chromium,
and only 23 percent of the nickel. Over-
all, this scheme was shown to be better
than three EDTA washes, better than switch-
ing the order of EDTA and hydroxylamine
hydrochloride, and much better than simple
water washes, in subsequent three-agent
tests. However, the EDTA washing alone
might be used with only a slight decrease
in removal efficiency.
204
-------
TABLE 2. TWO-AGENT SEQUENTIAL SHAKER TABLE EXTRACTION EFFICIENCIES
Soil Metals (ppm)
EDTA (0.1 M @ pH 6}
% Extracted
Magnesium chloride (1 M)
% Addnl . Extracted
EDTA (0.1 M § pH 6)
% Extracted
Potassium fluoride (0,5 M)
% Addnl . Extracted
EDTA (0.1 M @ pH 6)
% Extracted
Acetate buffer (1 M @ pH 5)
% Addnl. Extracted
EDTA (0.1 Ml? pH 6)
% Extracted
Tetrapyrophosphate (0.1 M)
% Addnl . Extracted
TABLE
Cd
47
83.6
1.02
95.3
1.17
119
2.36
75.3
23.9
Cr
349
24.4
0.11
28.9
0.37
24.3
2.36
24.2
5.59
Cu
219
77.6
2.22
56.4
1.27
76.3
1.18
59.6
3.11
Ni
214
10.8
1.47
11.6
0.47
10.7
1.89
9.72
0.99
Pb
2,480
84.6
0.29
85.3
0.85
117
1.41
98.2
1.20
3. CUMULATIVE SHAKER TABLE
THREE-AGENT SEQUENTIAL
Soil Metals (ppm)
1) EDTA (0.1 M @ pH 6)
2) Deionized water
3) Hydroxylamine hydro-
chloride (0.1 M in
acetic acid)
4) Deionized water
S) Citrate buffer
(0.1 Mi pH 3)
(=Total % Extracted)
Cd
47
87.2
92.5
96.3
96.6
98.4
EXTRACTION
Cr
349"
24.6
27.5
34.0
34.5
51.9
EFFICIENCIES
Cu
219 !
63.0
67.4
69.8
70.1
73.0
(*)
Ni
£'4— "
13.8
15.4
19.8
20.6
23.0
Pb
2,480
87.1
92.6
94.8
94.9
96.4
205
-------
Column Studies
The results of the metals extraction
achieved during column tests are shown In
Table 4.
The pattern of removal for each metal
was somewhat unique. Lead appeared to
be removed easily by the EDTA; further
removal occurred with citrate. Cadmium
was removed by EDTA and also by hydrox-
ylamlne hydrochloride; removal was slight-
ly Improved with the other treatments.
Copper was removed only by EDTA; the other
treatment methods had little effect on
removal. The data indicated a generally
high extraction efficiency for EDTA. The
analysis of metal remaining in soil versus
pore volume and type of treatment indica-
ted that lead and cadmium concentrations
1n soil decreased steadily from the begin-
ning of treatment to the end. The pattern
for the other metals was similar, but with
slight differences, probably due to random
sampling or analytical errors. Chromium
appeared to exhibit a pattern of migration
from the top to the middle of the column,
followed by rather ineffective removal.
Nickel showed a similar trend. These
latter results suggest that more pore
volumes of each treatment solution (e.g.,
10 rather than 5) would improve the re-
moval , probably to the level of extraction
efficiency achieved in the shaker table
tests.
CONCLUSIONS
The results of the shaker and soil
column studies permit a number of con-
clusions about the potential feasibility
of 1n i1tu cleanup of soil contaminated
with heavy metals.
The Cleanup Efficiency of the Soil Treat-
ment Agents
The various treatment-agent tests
showed that there are definite differences
in efficiency of the agents that vary with
the heavy metal.
TABLE 4. THREE-AGENT SEQUENTIAL EXTRACTION EFFICIENCIES:
SOIL COLUMN TESTS
son
Metals
(ppm)
Cd
Cr_
Cu
Ni
214
Pb
2,480
Water
% Extracted by water
0,2
0.1
EOTA (0.1 M 0 pH 6)
% Extracted by agent
60.5
12.2
47.1
6.8
60.1
Hydroxylamlne hydrochloride
(0.1 M in acetic acid)
% Extracted by agent 23.8
8.9
0.7
8.7
2.3
Citrate Buffer (0.1 M & pH 3)
% Extracted by agent
3.6
12.2
0.2
4.8
8.8
Water Wash
% Extracted by water
0.4
1.1
0.1
0.5
0.5
Total % Extracted:
88.5
34.4
48.1
20.8
71.8
206
-------
The preliminary tests of single heavy-
metal treatment agents provided the opti-
mum concentration and optimum pH for EDTA
treatment. The more concentrated solution,
0.1 M EDTA, 1s clearly more effective. A
pH of 5 is probably as effective as pH 6,
but either Is more effective than pH 7 or
above.
The two-agent tests demonstrated that
weaker agents do not remove any of the
metals of Interest more efficiently than
EDTA alone.
The three-agent tests demonstrated
that EDTA, hydroxylamlne hydrochlorlde, and
citrate buffer are all necessary for good
cleanup of the soil. The EDTA chelates
and solubiHzes all of the metals to some
degree; the hydroxylamine hydrochloride
probably reduces the Iron oxide-manganese
oxide matrix, releasing bound metals, and
also reduces Insoluble chromates to chro-
mium (II) and (III) forms; and the citrate
removes the reduced chromium and addi-
tional acid-labile metals. The chelating
agent/reducing agent/acidic citrate buffer
combination appears to be very effective
in heavy-metal cleanup.
The three-agent test with just EDTA
demonstrated that cleanup of cadmium and
chromium is significantly better with the
sequential EDTA/hydroxylamine/ citrate
than with three treatments of EDTA alone.
However, EDTA alone appears to be suffi-
cient for removing the lead and copper;
although the nickel removal was poor with
EDTA alone, the treatment with all three
agents showed no better removal.
The three-agent test with hydroxyla-
mine hydrochlorlde first, followed by EDTA
and then citrate, demonstrated that the
use of a chelating agent following the re-
duction step does not improve the cleanup.
Effects of the Soil Characteristics on the
Cleanup Efficiency
The efficient cleanup of the heavy-
metal contamination in the soil was prob-
ably facilitated by the low cation ex-
change capacity (CEC) of the soil. How-
ever, the presence of iron and manganese
oxides apparently interferes with heavy
metal removal by EDTA; reducing these
oxides was necessary to remove all the
cadmium.
Feasibility Studies Using Shaker and
ColumnTests
The shaker studies were quick and
effective screening tests for estimating
treatment-agent efficiency. The column
tests, although more difficult and time-
consuming more closely represent the
behaviour that might be expected if the
agents were used for in situ cleanup of an
actual contaminated sTieTThe column tests
model cleanup under gravity flow conditions
through soil with a permeability somewhat
similar to the native soil. If time had
permitted longer soil column tests, extrac-
tion efficiencies would probably have been
similar to the shaker table test results.
Both the shaker and column tests are very
useful for studying the feasibility of
potential soil cleanup agents.
REFERENCES
1. Ellis, W. D., J. R. Payne, and 6. 0.
McNabb. 1985 Treatment of Contamina-
ted Soils with Aqueous Surfactants.
EPA/60Q/S2-85/129U.S.Environmental
Protection Agency.
2. Repa, E. W., E. F. Tokarski, and R. T.
Eades. 1984. Draft Final Report.
Evaluation of the Asphalt Cover at the
Western Processing, Inc., Superfund
SitelEPA Contract #68-03-3ll3. UTS.
Environmental Protection Agency.
207
-------
FIELD EVALUATION OF IN SITU WASHING
OF CONTAMINATED SOILS WITT? WER/SURFACTANTSl
James Nash
Mason & Hanger-Silas Mason Co., Inc.
P.O. Box 117
Leonardo, New Jersey 07737
Richard P. Traver
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, New Jersey 08837
ABSTRACT
Since 1981, the Releases Control Branch of the Hazardous Waste Engineering Research
Laboratory has been developing techniques to wash contaminated soil in place (In situ).
The project includes: design and fabrication of the hardware to carry out the washing,
evaluation of surfactants to do the washing, determination of which geological character-
istics to use to judge the appropriateness of 1n situ washing, development of a monitoring
and reporting system, evaluation of two candidate sites for the field testing of the hard-
ware, and a pilot treatment study at a contaminated site.
This paper summarizes the design and development of the In Situ Contalnment/Treat-
ment Unit (ISCTU) and the evaluation of surfactants for in situ soil washing. The empha-
sis !s on work completed at Volk Air National Guard Base, Camp Douglas, Wisconsin. The
work shows that surfactants will remove otherwise obstinate contaminants from soil even
without mechanical agitation of the soil. However, subsequent treatments of the surfactant
laden leachate is an unresolved problem.
INTRODUCTION
In situ soil washing is the term to
descrfFe washing of contaminated soil with-
This report is a summary of work per-
formed 1n partial fulfillment of Con-
tract Numbers 68-03-3113 and 68-03-3203
under sponsorship of the U.S. Environ-
mental Protection Agency. The U.S. Air
Force through Interagency Agreement
*RW 57931283-01-0 with the U.S. EPA
has also sponsored much of the work
reported here. This paper discusses
out excavating. The washing is accomplished
by applying a liquid at or near the surface
the key activities of four projects:
"Treatment of Contaminated Soils With
Aqueous Surfactants", "Retrofit of the
In Situ Containment and Treatment Unit",
Chemical Countermeasure Application at
Volk Field Site of Opportunity", and
"Site Characterization and Treatment
Studies of Soil and Groundwater at Volk
Field."
208
-------
so that the solution will flow down
through the soil structure. By substitu-
tion, emulsification and/or solubilization,
contaminants are removed from soil parti-
cles and held in the liquid phase. The
liquid will then percolate down to a
perched or unconfined aquifer where it
can be removed through withdrawal wells.
Soil Soils with permeabilities (a measure
of the ability of a fluid to pass through
the soil) of greater than 1Q-4 cm/sec
should be suitable for this technique.
At this time, in situ soil washing
with a surfactant is not a field useable
remedial technique. Although surfactants
do remove obstinate contaminants from
soil, treatment and disposal problems
after removal have not been solved.
Petroleum hydrocarbons and PCBs, which
have low mobility in soil.structures, were
removed from soil in laboratory tests.
Problems still remain regarding separation
of the surfactant from the contaminant and
the water.
Treatment of Contaminated Soils with
Aqueous Surfactants
Soap, a sodium or potassium salt of
fatty acids, was the earliest man-made
surfactant. Credited to the Phoenicians
in 600 BC, there had been no additional
surfactants developed until the twentieth
century. The relatively recent develop-
ment of surfactants has included sulfo-
nates and ethoxylates. In work conducted
at Texas Research Institutel for the
American Petroleum Institute, two commer-
cially available surfactants were identi-
fied for use in lab tests to wash gasoline
from Ottawa sand. Those two surfactants,
used together, were an ethoxylated alkyl-
phenol and a dodecyl benzene sulfonate.
The reasons for their selection at that
time were low interfacial tension and
compatibility with salts found in soil.
The API work as well as studies conducted
under this EPA program revealed that
surfactants in water solutions may
hydrolyze and form floes that block soil
pores or will block the soil pores with
surfactant particles called micelles.
Blockages are also caused by the
surfactant/contaminant emulsion. The use
of two surfactants is required because the
ethoxylated alkylphenol retards the for-
mation of floes and micelles while the
sulfonate is required for cleaning. The
surfactants should be easy to mix with
water and should not cause the fine soil
particles to be suspended in the wash
solution. Mobilized soil fines will block
flow at the narrow passages between soil
particles. A "mat" is formed if enough
passages are blocked along a continuous
front. These mats halt fluid flow and
thereby stop the washing process in that
area. Aging studies of surfactant solu-
tions were performed to observe the forma-
tion of floes. The measurement of turbi-
dity over time was used to demonstrate the
effect of blending. The surfactants
selected for blending in this work were an
ethoxylated fatty acid and an ethoxylated
alkyl phenol. See Figure 1.
'
!W
HD
n
*
K
so
«
a
10-
I—I—r—
I* 22
—I—
U
14
I OBD
Figure 1.
j(me(w)
* « EMNUE
Particle growth as measured
by turbidity increase. For
two surfactants and their
blend.
The crystalline floes formed during
these measurements blocked the pores of a
column of medium to fine sand.
209
-------
The first washing tests were run on
a shaker table and the next test series in
columns. Contaminated soil was compacted
1n 3 in. increments into 3 in. diameter,
5 ft high glass tubes. The tubes were
fitted with nippled glass caps at the
bottom and top. A pressure head of 30 cm
of surfactant solution was applied to the
surface of the contaminated soil. The
soil pores were, therefore, experiencing
saturated flow of the surfactant solution,2
The soil used for the laboratory work
was a Freehold series typic hapludult from
Clarksburg, New Jersey. It was selected
because of its grain size distribution and
similarity to soil at CERCLA candidate
sites in EPA's Region II. Ten percent was
silt or clay, eight percent gravel and 80%
coarse-to-fine sand. Its permeability of
10-4 l;m/sec is at the low end for in situ
washing. Nine to eleven percent of tKi
soil was HC1 soluble. Of the crystalline
structure, 98% was quartz and 2% was
feldspar. Only 0.12% was organic carbon
which is a low value and accounts, in
part, for a low cation exchange capacity.
A topped Murban crude oil 1n methyl-
ene chloride was applied to the soil.
This contaminant was selected because it
contained many organic types including
aromatics, polynuclear aromatics, alipha-
tics, polar and non polar compounds. The
methylene chloride was allowed to evaporate
and the soil was aged prior to being
loaded into the test columns. Other con-
taminants, in separate tests, were chloro-
phenols and a polychlorinated biphenyl.
Gas chromatographic analysis showed
that ten pore volumes of surfactant solu-
tions passed through the columns removed
88% of the topped Murban crude oil and 90%
of the PCB's. Using high performance
liquid chromatograijhy (HPLC), it was shown
that chlorophenols were removed with the
water alone. Surprisingly, removal in the
column studies, where there is a low level
of mechanical washing, was better than re-
moval in the shaker table studies. Start-
Ing at 1000 ppm contamination in the
columns, removal efficiencies as high as
98% were reported.
Control of In Situ Mashing Fluids
Accelerating the natural tendency of
a contaminant to migrate through the vadose
zone into the groundwater is the basic
purpose of in situ soil washing. In order
to do this so there is no adverse Impact
on an aquifer, rigid controls must be
maintained to assure the contaminant is
captured. The EPA's In Situ Containment
and Treatment Unit (ISCTU) was designed
for this purpose. The drawing in Figure 2
represents the parameters (of an hydraulic
budget) that were considered for the
(ISCTU). They are: recharge Ga, discharge
Da, treatment system flow R, evapotrans-
piration E, precipitation P, natural
groundwater flow Uj, and induced ground-
water flow Ug. Variation in these
qualities will change those items in lower
case letters; vadose zone thickness w,
mounding m, drawdown (he-hw), and radius
of influence re (not to be confused with
the radius of capture).
Figure 2. In_ situ parameters
210
-------
Figure 3 Is a simplified drawing of the
ISCTU, which is equipped with recovery and
delivery pumps, batch mixing and propor-
tional-additive metering pumps, flow rate
controls, pressure and flow meters, and a
volatile organic stripping tower. Any
treatment of groundwater requiring more
than air stripping must be done "off-
board." A microcomputer/data logger is
used to monitor environmental conditions
and the effect of pumping and recharge on
the aquifer. To do this, depth gauges,
flow meters, moisture meters, and a weather
station are connected to the data logger.
A. AIR DIAPHRAGM PUMPS
B. PROPORTIONAL CHEMICAL
' AOOITIVE ME TE RING PUMP
C. INPUT MANIFOLD
MAIN ELECTRICAL
BREAKERS
O. PROCESS MONITOR RECORDER
f. WATER PUMP
f. BATCH CHEMICAL METERING PUMP
CHEMICAL MIXING TANK
PULLOUT OPERATOR'S PLATFORM
DIESEL ELECTRICAL
GENERATOR
INJECTION MANIFOLD
Figure 3. In Situ Containment and Treatment Unit
Site Selection for the Field Evaluation
In September 1984 the U.S. Air Force
and the U.S. EPA started in a joint effort
to evaluate in situ washing technology.
The primary objective of the project was
to demonstrate full-scale feasibility.
A secondary objective was to develop a
more comprehensive strategy for the decon-
tamination of fire-training areas of all
Air Force and Department of Defense (OoD)
installations. The following criteria
were used in selecting a site suitable
for full-scale soils washing research.
A site of less than one acre was desired
to reduce soil variability and reduce
sampling costs. Because soil washing is
best suited for permeable soils, a sandy
site was sought. Contaminants at the
site were to be common organic chemicals
found at many other Air Force sites,
i.e., trichloroethane, benzene, toluene,
trichloroethylene. Officials of the
selected installation and responsible
environmental agencies would need to be
cooperative.
Preliminary screening of candidate
sites was accomplished through a review
of Air Force Installation Restoration Pro-
gram (IRP) reports. Over sixty reports and
nearly 800 sites were screened. During
the review, it was apparent that most sites
with organic chemical contamination fell
Into two common categories: sites of fuel
spills and fire training areas.
Fire training areas were especially
suited to this research because of their
limited size and range of contaminants,
which included chlorinated solvents, fuel
components and lubricating oil. ' Fire
training areas are found at almost all Air
Force installations and, because of the
long-term fuel and solvent dumping at these
sites, they have significant off-site pollu-
tion potential.
Following this careful review, a fire
training area at Volk Field, Air National
Guard Base, Wisconsin, was selected as a
research site. Historical records indi-
cate that the Volk fire training area may
211
-------
hav? been established as early as World
War II and has routinely received waste
solvents, lubricating oil, and JP-4 jet
fuel. Although it is impossible to deter-
mine the quantity of chemicals that soaked
Into the ground versus the amount volatil-
ized and burned in fire training exercises,
one estimate is 52,000 gallons. Measure-
ment of volatile organics from groundwater
samples taken in 1980 directly below the
fire pit showed chloroform, trichlorethane,
trichlorethylene, benzene, toluene, and
ethyl benzene totaling above 50 mg/liter.4
Site Studies
Two site studies were made at the fire
pit area during 1985. These studies were
conducted to thoroughly understand the
hydrology and chemistry associated with
the contamination have produced as a
by-product a great deal of data and in-
sight into a chronic oil spill. Initially,
the character of the contamination was
misunderstood. The original concept of
a floating layer of oil that could be
handled easily gave way to the realization
that the contamination had not remained as
a water insoluble oil but had been trans-
formed to soluble organics by biological/
chemical activity. Biological activity .
had been nourished by the firefighting
foams used in the training exercises.
These fire-fighting foams may have also
contributed directly to solubi11 zing the
oils. The groundwater, 25 ft below the
surface (and only 60 ft from the
pit), had up to 50 mg/Hter total organic
carbon (TOC). Infrared spectrophotometric
(IR) scans indicated this contamination
was 1n part esters or organic acids.
Upon emerging from the centrifugal pump
(used for a pumping test), the groundwater
frothed.5 Directly below the pit the
water table was at 12 ft. The hydraulic
conductivity was 5 x 10-2 cm/sec.
Treatment Studies of the Soil
The overall soil contamination had
the physical consistency of a medium
weight lube oil. At a one-foot depth
average oil and grease (determined by
carbon tetrachloride [0614] extraction)
was 13,500 mg/kg. Deeper into the soil,
oil and grease (O&G) values decreased.
At 5 ft, and continuing to the capillary
zone at 10 ft, O&G values were 400-800
mg/kg. Soil samples from the aquifer
taken at 15 ft produced 5000 mg/kg O&G.
The chemical composition of the CCl-4
extract also varied with depth. IR scans
of extracts of soil from 1 ft depth
match scans of parafflnic oil. Esters or
acids of oil become more evident when
approaching the water table. Below the
water table, the oxidized oils although
present, are less prominent. This profile
is apparently a symptom of weathering.
The more soluble oxide forms have been
carried to the groundwater by percolating
rain water.
The volatile contaminants also show
evidence of weathering. In contrast to
O&G, the weathered volatiles are found
closer to the surface than to the water
table and are an order of magnitude less
abundant than O&G extracts. A relatively
high abundance of isoprenoid compounds
(includes many naturally occurring mater-
ials such as terpenes) in relation to
normal alkanes also indicates long term
microbial degradation.6 A terpene-Uke
odor was noticed while taking soil samples
to determine the lateral extent of contam-
ination near the surface. Within 6
in. of approaching the clean soil and
at depths of 6 to 12 in. a
"minty-turpentine" smell was reported by
the field technician.
A part of the fire training area was
prepared so that ten mini soil washings
could be conducted simultaneously. The
first foot of soil was not to be included.
Therefore, ten 1 ft deep holes were dug
and the bottom of each hole was called
the "surface" of the test chamber. Each
"chamber" was a 14-in. depth of soil from
the bottom of the hole down. Surfactants
tested were: an anionic sulfonated alkyl
ester (Pit 17), a polyethylene glycol dio-
leate (Pit #10), ethoxylated alkyl phenol/
ethoxylated fatty acid blend (Pit #8), and
the contaminated groundwater (Pits #2,3,4,
5,9). The dioleate caused soil plugging
immediately. Compared to water, penetra-
tion rates were reduced when any surfac-
tant solutions were used. The groundwater,
212
-------
which* has a low concentration of biologi-
cally produced surfactant, had the least
effect on the penetration rate.
The dominant contamination in the
soil was oil and grease, up to 16,000
mg/kg, where volatiles were less than 100
mg/kg.6 g&G measurements were therefore
used to determine the effectiveness of the
soil washing. To avoid channeling during
the pilot treatment, prewash Q&G measure-
ments were made on samples taken adjacent
to the chambers. Statistically, the O&G
measurements had a coefficient of varia-
tion (CV) throughout the test area of 35%
making it difficult to draw conclusions
of soil washing effectiveness. Figure 4
shows the O&G measurements after the sur-
factant wash process and the blank value.
Pit #8 was washed with the lab-developed
50/50 surfactant blend. It is interesting
to note that the O&G at 12-14 in. has
increased 24% above the blank and the
surface top layer O&G has decreased 50%,
Implying a transport of contaminant down-
ward during the seven days of washing with
14 pore volumes. Keep in mind a CV of 35%
precludes any definitive conclusion. The
expected reduction of contamination at the
12 in. depth to 50% of the original level
was not realized.
Treatment Studies ofthe Groundwater
Bench scale and then pilot treatment
studies of the already contaminated ground-
water were undertaken 1n anticipation of
full-scale soil washing. Bench-scale
studies evaluated addition of: lime,
hydrogen peroxide, alum, ferric chloride,
and various water treating polymers. The
pilot treatment was run using the EPA's
Mobile Independent Chemical/Physical Treat-
ment Unit, a holding lagoon, and an air
stripper made by the Air Force. Figure 5
is a process flow diagram that also indi-
cates sampling points. The three treat-
ments consistently used during the opera-
tion were lime addition, settling, and
24*
22*
18-
16-
W,
12-
10
fi.
f
8
I
*>,
^
'
,
'\
^
t\
>\
'\a
t
n s
^
^
m_
\\
\\
\\
3
g P
1 ?
N /
y 1
\ t
S! 7
S? '
S{ '
bi ^
4
^
S
\
s ™ _,
C ^ ;V'
v ^ / ^ !
S H (*S^
§i_J ^1
S a_, ^ a. A i ,
S 7 8
Mnntar
f
J
f
f
f
$
^
^
/*,
?*(
? ^
A ?
/ S {!
/ U
Aj
»
^
s
^
^
^1
\ t
»
Figure 4, Soil washing data
volatilization. Total organic carbon
(TOC), volatile organic analysis (VOA),
and suspended solids (SS) tests were
used to monitor the effect of these
treatments.
Addition of lime brought about signi-
ficant reductions in TOC. Organics were
removed with an iron hydroxide to form a
floe. (Interestingly, the contaminated
groundwater had up to 56 mg/liter iron
compared to background levels of 0.2 mg/
liter.) Volatiles were 95 to 98% removed
in the lagoon and air stripper. Figure 6
is a bar chart depicting the measured
level of TOC at four points in the process.
Figure 7 is a bar chart showing the mea-
sured levels of four volatiles at three
locations in the process.
213
-------
LIME
Will
VOUTIUS
2
1
|
t
*
—
HASH
•U ,
J6I ~W
Cl MINER
^
11
5 K
_| / 10 «ROUC
UEOOM
Figure 5. Well field effluent treatment process and
sampling points -
m~n
30 it
7 »-X
I"*5
- M'^l
I 4
"*^U
i »-^S'
u^j
KB
1
^
«: I
•^11
w
t|
ra
zh
ll
i
i
I
i
y
^
p
1234
Figure 6. Four data sets showing Level of
TOC at the well field, clarifler
effluent, stripper feed, and
stripper effluent.
I
''/,
I
I
\
I
I
\
\
*m
1
1
1
I
I
I
vTJ
^
^ m
$M K/ta
TGA
BEN
TOL
COMtMD
ETH
BEN
Figure 7. Volatiles at the well field,
stripper feed, and stripper
effluent.
214
-------
In anticipation of conducting a in
situ soil washing of the entire pit,
tests were run to determine control of
the natural groundwater flow beneath the
pit. This was accomplished by a six-
member well field. In total there have
been 13 wells Installed in the study,
7 monitoring wells and 6 withdrawal wells.
Boring logs were' kept during the drilling
operations. Split spoon samples of the
sand and weathered sandstone were used for
chemical analysis and particle size analy-
sis. The fines content of the directly
below the pit is significantly lower than
in the adjacent uncontaminated soils -
2 to 5% versus 10 ,to 15%. Fines content
of soil 8 ft below the water table,
slightly down gradient, and in the plume
is unusually high: 28% versus 10-15%.
The production wells placed in the high-
est contamination zones have the poorest
fluid yield. Paradoxically, according
to equipotential lines constructed from
water table depths, there is a convergence
of flow passing beneath the pit
(see Figure 8).
902.45
AIR
STRIPPER
Figure 8. Treatment site showing water
table equipotential lines
This is directly in line with a pro-
duction well producing water containing
700 rag/ liter TOC at less than 2 gallons
per minute. The average for the rest of
the wells is 260 mg/liter at 6 gallons per
minute. The design pumping rate for each
well was 12 gpm* In spite of well yield
problems the natural gradient of 0.001
(ft/ft) was easily reversed to create a
radius of influence of greater than 100
ft and a radius of capture greater than
the 40 ft training pit radius.
A Follow-up Electromagnetic Survey
An electromagnetic survey was con-
ducted over the ground surface surrounding
the training area to determine the measur-
able extent of the plume. The decision to
do this was based on the low conductivity
of the soil, high conductivity of the
plume (600 micromohs), and the low
conductivity of the background water
(20 micromohs). A study conducted by
the New Jersey Geological Survey7 had been
215
-------
able to map an organic plume from a fire
training area in a sandy aquifer. In the
report of that work, the fire fighting foam
AFFF was felt to be the conductive organic
that made the survey possible. In this
work the high iron content of the plume is
considered the reason for the success of
the survey. The reason for the high iron
content is the reducing conditions that
exist(ed) during biological activity at
the site. Figure 9 is a map of the plume
based on conductivity.
The CCI_4 extract of a soil sample taken
at 12 ft at the point marked "S" in the
figure was Identified as an oxidized oil.
The authors wish to express their
appreciation for the cooperation, encourage-
ment and help given by a number of people
from the Wisconsin Air National Guard and
Department of Natural Resources. But
especially we wish to acknowledge Doug
Downey of the U.S. Air Force for his
gentle persistence in directing the work
done at Volk Field.
CONCLUSION
The mechanical aspects of applying
a surfactant to soil and controlling an
underlying unconfined acquifer to capture
the wash solution have been demonstrated at
a site of opportunity. Issues that remain
to be addressed are treatment, if necessary,
of the used surfactant solutions, isolation
of the containment from the surfactant and
developing a method to recycle the surfac-
tant.
Figure 9. Electromagnetic Survey
216
-------
REFERENCES
1. Texas Research, Institute, Inc.
Underground Movement of Gasoline on
Groundviater and Enhanced Recovery by
Surfactants. September 14, 1979
American Petroleum Institute, 2101 L
Street, NW, Washington, DC.
2. Ellis, W. I)., J. R. Payne, Treatment
ofContaminated Soils with Aqueous
Surfactants (Interim Report)
September 6, 1985 to EPA-HWERL.
Releases Control Branch, Edison, NJ.
3. Waller, M. 0., R. Singh, J. A. Bloom.
Retrofit of a Chemical Delivery Unit
for In Situ Waste Clean-up, EarthTech,
Inc. January 7, 1983.
Releases Control Branch, Edison, NJ.
4. Hazardous Materials Technical Center
Installation Restoration Program
RecordsSearch prepared for 8204th
FieldTraining Site, August 1984
available N.T.I.S.
5. Nash, J. H.» Pilot Scale Soils Mashing
and Treatment at Volk Field ANG,Camp
Douglas Wj, inpreparation.
6. McNabb, G. 0., et. al. Chemical
Countermeasures Application at Volk
Fi el d Site of Opj)ortuni ty,
September 19, 1985 to EPA-HWERL.
Releases Control Branch, Edison, NJ.
7. Andres, K. G. and R. Crance, Use of the
Electrical Resistivity Technique to
Delineate^a Hydrocarbon Spill In the
Coastal Plain Deposits of New Jersey. '-
Proceedings: Petroleuro Hydrocarbons
and Organic;•Chemicals in Ground WateF,
November 5-7, 1984 aval 1 able National
Water Well Association, Dublin, OH.
217
-------
EFFECT OF FREEZING ON THE LEVEL OF CONTAMINANTS IN UNCONTROLLED
HAZABBQUS WASTE SITES - PART II! PRELIMINARY RESULTS
Iskandar K. Iskandar, Lawrence 8. Ferry and Thomas F. Jenkins
U.S. Army Cold Regions Research and Engineering Laboratory
Hanover, NH 03755-1290
and
Janet M. Houthoofd
U.S. Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
Artificial ground freezing has recently been identified as a potential method for
facilitating site decontamination. This study was conducted to evaluate the feasibility
of using artificial ground freezing for dewatering slurries and decontaminating soils at
uncontrolled hazardous waste sites.
Preliminary column studies in the laboratory showed that freezing and thawing
slurries significantly reduced their volume by as much as 40%. Freezing and thawing soils
contaminated with volatile organics such as benzene, chloroform, toluene and tetrachloro-
ethylene decreased substantially the transport of these organics to ground water and
apparently enhanced their loss by volatilization. The effect of freeze/thaw on volatile
organics was dependent on the number of freeze/thaw cycles and the octanol-water partition
coefficient for the specific substance.
For heavy metals, freeze/thaw cycling had a mixed effect. Concentrations of Zn in
the leachate were reduced relative to an unfrozen control while concentrations of Hi
increased* Overall leaching accounted for only a small portion of the metal present.
A larger scale study is underway to investigate the effects of varying freezing
rates, soil types and soil moisture content on the mobility of trace components in the
soil.
INTRODUCTION
Iskandar and Houthoofd (1985) re-
viewed the literature related to the ef-
fect of freezing on the mobility of con-
taminants in uncontrolled hazardous waste
sites. They concluded that artificial
ground freezing could be useful in con-
trolling the migration of contaminants at
these sites and perhaps in treating con-
taminated soils. However, the literature
contained very little information on this
subject. Iskandar and Jenkins (1985) pub-
lished the results of preliminary tests on
the potential use of artificial freezing
for contaminant immobilization.
The purpose of this paper is to sum-
marize the data obtained from a laboratory
column study on potential use of artifi-
cial ground freezing to immobilize con-
taminants in soils and to treat contami-
nated soils.
MATERIALS AND METHODS
Plexiglass columns (described by
Iskandar and Jenkins, 1985) 12.5 cm in
diameter and 80 cm in height were filled
with dredged material from Green Bay,
Wisconsin. The water content of this
dredged material was increased to 176%
218
-------
(W/w) by addition of water containing
spike additions of four heavy metals and
four volatile organlcs. The metals added
were Cd, Za, Cu and Ni in eoneentratioas
ranging from 400 to 800 ug/8 dry soil.
The organics included chloroform, benzene,
toluene and tetrachloroethylene in concen-
trations ranging from 40 to 45 jig/g dry
soil. The slurry was homogenized using a
mechanical stirrer and the columns packed
as quickly as possible to minimize losses
of volatile organics. The pH and oxidiz-
able natter of the dredged material were
determined to be 7.0 and 4.7% respective-
ly.
Six different treatments were com-
pared using one column each. 'Two control
columns were included. In the first, the
slurry was maintained unfrozen at 5°C
throughout. A second column was frozen
from the bottom up and maintained frozen.
The other four columns were subjected to
1, 2, 3 or 5 freeze/thaw cycles, respec-
tively, freezing and thawing from bottom
to top. The columns were instrumented
with thermocouples to measure the soil
temperature at depth, Leachate samples
Were collected during thawing periods by
gravity flow, and subsamples were
analyzed. At the conclusion of the ex-
periments, soils were divided into IQ-cra
sections and subsamples were analyzed.
Metals were analyzed by Inductively
Coupled Plasma (ICP) and organics were
determined by extraction with tetraglyme
according to the method of Gurka et al,
(1984) and analysis on a gas ehromato-
graph-mass-spectrometer (HP 5992 GC-MS)
equipped with a purge and trap sampler (HP
7675A) according to EPA method 624 (Feder-
al Register, 1984). Deuterobenzene (C6D6)
in tetraglyme was added to each sample as
an internal standard just prior to purg-
ing.
During freeze/thaw cycles, the soil
was frpzen gradually from the bottom up.
The soil temperature during freezing
ranged from -1.8° to -16.0°C.The rate of
frost penetration, recorded using a ther-
mocouple array and a data logger, ranged
from 15 to 50 cm/day.
RESULTS AND DISCUSSION
Consolidation •
Table 1 shows the effect of freeze/
thaw cycles on volume reduction. The
Table 1. Effect of freeze/thaw cycles on
consolidation.
Freeze/Thaw
Cycle
0 (unfrozen)
0 (unfrozen)
1
1
2
3
3
5
Time
(days)
50
125
12
70
22
27
41
50
% Volume
Reduction
34
36
28
34
37
36
37
35
volume of the unfrozen treatments was
decreased by 34% over 50 days and 36% over
125 days as a result of natural drainage.
Similar volume reduction was achieved in
only 22 days when the column was subjected
to two freeze/thaw cycles. Increasing the
number of freeze/thaw cycles to five did
not result in additional consolidation.
It should be emphasized that no additional
water was added to the sediments between
the freeze/thaw cycles. Drainage was
permitted during thawing periods.
From the data in Table 1 it was con-
cluded that freezing and thawing saturated
sediments can decrease the time required
to dewater these materials relative to
natural drainage. Also, the rate of water
movement through soils was enhanced by 3-4
times the magnitude by freezing and thaw-
ing (data presented in Iskandar and
Jenkins, 1985). This was explained by the
formation of aggregates and cracks in the
soils as a result of freeze/thaw cycles.
The enhancement of water movement in soils
could be used beneficially in conjunction
with other methods of in-situ treatment
which require several additions of water
or chemicals to flush contaminants from
soil.
Fate of Volatile Organics
Figure 1 shows the concentration of
chloroform in the leachate from the dif-
ferent treatments as a function of time.
The highest levels of chloroform were
found in the leachate from the column that
was maintained unfrozen throughout and
from the column that received one cycle of
freeze/thaw. These concentrations ranged
from 8 to 12 ng/mL. The concentration of
chloroform in the treatments subjected to
219
-------
£O 40
Elapsed Time (days!
60
Figure 1. Concentration of chloro-
form in leachate as a function of
time and freeze-thaw cycles. Un-
frozen (A), one cycle of freeze-thaw
(B), three cycles of freeze-thaw (C),
five cycles of freeze-thaw (D).
20 40
Elapsed Time (days)
60
Figure 3- Concentration of tolu-
ene in leachate as a function of
time and freeze-thaw cycles. Un-
frozen (A), one cycle of freeze-
thaw (B), three cycles of freeze-
thaw (C), five cycles of freeze-
thaw (D).
SO 40
Elapsed Time (days)
60
Figure 2. Concentration of ben-
zene in leachate as a function
of time and freeze-thaw cycles.
Unfrozen (A), one cycle of
freeze-thaw (B), three cycles
of freeze-thaw (C), five cycles
of freeze-thaw (D)•
two or more freeze/thaw cycles ranged from
0 to 5 tig/mL. In general, chloroform con-
centrations in the leachate decreased with
increasing numbers of freeze/thaw cycles.
Figure 2 shows the concentration of
benzene in the leachates as a function of
time and freeze/thaw cycles. With the
exception of the treatment that received
one freeze/thaw cycle, the effect of
freeze/thaw cycles on benzene concentra-
tion in the leachate was similar to that
on the chloroform concentration. The
reason for the higher benzene concentra-
tion in the leachate from the column
undergoing one freeze/thaw cycle as com-
pared with the unfrozen control is
unknown.
Figure 3 shows the variation in
toluene concentration in the leachate from
the different freeze/thaw treatment
cycles. Up to 35 days, the trend was
similar to that of the chloroform and ben-
zene, but the concentrations were much
lower. The concentrations of toluene were
generally less than 1 ug/mL with the ex-
ception of that resulting from the single
freeze/thaw treatment, which ranged from 0
to as high as 2,1 pg toluene/rat.
Tetrachloroethylene concentration in
the leachates was very low (data not pre-
sented). With the exception of the unfro-
zen treatment and the one-cycle treatment,
concentrations of tetrachloroethylene did
220
-------
not exceed 0.03 ug/mL and were usually <
0.01 pg/mL. The leachates from the un-
frozen column and the column receiving one
freeze/thaw cycle varied widely, from 0 Co
0.15 jig/mL. These data confirm that
tetrachloroethylene is not as mobile in
soil as benzene, chloroform and toluene.
In general, it can be concluded that
concentrations of these volatile organics
in the leachate were higher in the unfroz-
en treatment subjected to natural drainage
than in the leachate from the frozen/
thawed treatments. This suggests that
freezing the soil from the bottom up could
decrease groundwater contamination for
these four volatile organics. Also, the
order of concentration of organics in both
the unfrozen and the frozen/thawed treat-
ments was chloroform > benzene > toluene >
tetrachloroethylene. The concentration of
these organics in the leachates is in-
versely correlated to their octanol/water
partition coefficients (kow). The kQW
values for these organics are 93, 135, 490
and 615 for chloroform, benzene, toluene
and tetrachloroethylene, respectively.
Octanol/water partition coefficients have
been successfully correlated with sedi-
ment/water partition coefficients for
many hydrophobic organics (Karickhoff et
al., 1979).
Figures 4, 5 and 6 show the effect of
freeze/thaw cycles on" the concentrations
of chloroform, benzene and toluene, re-
spectively, remaining in the soil at the
Figure 5. Vertical distribution of
benzene in soil subjected to freeze-
thaw. Kept frozen (A), one cycle of
freeze-thaw (B), unfrozen (C), five
cycles of freeze-thaw (D)>
Chloroform (ft-g/g soil)
2O 4O 60 8O
'4 ." I '—I ' 1 I I
20
40
60
sol-
Figure 4. Vertical distribution of chlo-
roform in soil subjected to freeze-thaw.
Kept frozen (A), one cycle of freeze-thaw
(B), unfrozen (C), five cycles of freeze-
thaw (D).
Figure 6. Vertical distribution of
toluene in soil subjected to freeze-
thaw. Kept frozen (A), one cycle of
freeze-thaw (B), unfrozen (C), fiw
cycles of freeze-thaw (D).
conclusion of the experiment. Fte^/.s. ^:«aw
treatments decreased the residual concen-
tration of chloroform in soils relative to
the unfrozen control, and this effect was
correlated with the number of freeze/thaw
cycles. The combination of freezing and
thawing was necessary to achieve the
chloroform removal since levels of chloro-
form, as well as benzene and toluene, were
highest in the column that was maintained
frozen throughout the experiment.
221
-------
Table 2. Mass balance calculation of added and recovered organics as a function of
freeze/thaw cycles.
F/T Cycles, Organic
Added
rag
Recovered
in Soil
rag
%
Recovered in
Leachate
rag
%
Unaccounted For
rag
I
0 (Kept Frozen) (67 days)
Chloroform
Benzene
Toluene
Tetrachloroethylene
0 (Kept Unfrozen) (125
Chloroform
Benzene
Toluene
Tetrachloroethylene
1 Freeze/Thaw (70 days)
Chloroform
Benzene
Toluene
Tetrachloroethylene
2 Freeze/Thaw (22 days)
Chloroform
Benzene
Toluene
Tetrachloroethylene
3 Freeze/Thaw (41 days)
Chloroform
Benzene
Toluene
Tetrachloroethylene
5 Freeze/Thaw (50 days)
Chloroform
Benzene
Toluene
Tetrachloroethylene
848
657
897
739
days)
969
752
1025
844
1073
833
1135
936
1204
934
1273
1049
1126
873
1191
981
1248
968
1319
1087
589
600
690
634
337
388
211
325
547
490
664
759
289
243
TO
499
627
649
837
784
82
29
55
560
69
91
77
86
35
52
21
39
51
59
59
81
24
26
ND
48
56
63
70
80
7
3
4
52
0
0
0
0
11
5
1
0
8
8
1
0
7
2
0
0
7
4
0
ND
4
3
1
0
0
0
0
0
1.1
0.7
0.1
0
0.7
1.0
0.1
0
0.6
0.2
0
0
0.6
0.5
0
ND
0.3
0.3
0.1
0
259
56
207
105
621
359
813
519
518
335
469
176
908
689
ND
550
492
220
354
197
1162
936
1263
527
31
9
23
14
64
48
79
61
48
40
41
19
75
74
ND
52
44
25
30
20
93
97
96
48
* ND, not determined.
Residual concentrations of tetrachlo-
roethylene in the soil profiles from the
various treatments differed less than for
the other three organics studied. This is
probably a result of its reduced mobility
in the soil compared with benzene, toluene
and chloroform, due to stronger binding
with soil organic matter.
Table 2 summarizes the mass balance
data for the four organics in the differ-
ent treatments. The cumulative amounts of
222
-------
organics leached from the soil columns
were low, ranging from 0 to 11 mg. This
did not represent more than 1.1% of the
total added for any of the four organics
for any of the five treatments. The
residual amounts of organics recovered in
soils from the column maintained frozen
throughout were 589, 600, 690 and 634 mg
for chloroform, benzene, toluene and
tetrachloroethylene, respectively, which
constitutes 69 to 91% of the amounts
added. The amounts unaccounted for in
this treatment ranged from 56 mg benzene
to 259 mg chloroform. These amounts
represented from 9 to 31% of the added
benzene and chloroform, respectively.
The amounts of organics recovered
from the control column that was main-i
tained unfrozen were generally less than
those recovered from the frozen treat-
ments. On a percentage basis, the re-
covered organics in the unfrozen treatment
represented 35, 52, 21 and 39% of the
added chloroform, benzene, toluene and
tetrachloroethylene, respectively. The
amounts of organics unaccounted for in the
unfrozen treatment were much higher than
those recovered in the frozen/unthawed
treatment. Thus from a standpoint of
holding these volatile contaminants in
place, maintaining the soil profile frozen
was the most effective.
The effect of freeze/thaw cycles on
the recovery of volatile orgaaics was not
consistent. The one-cycle and three-cycle
treatments showed similar effects on the
amounts of organics recovered in the soil,
which did not differ drastically from the
unfrozen control. This could be due to
the time factor, whereas the unfrozen
treatment continued for 125 days, the one
cycle 70 days, the three cycles was only
41 days. The amounts of organics lost to
the atmosphere is more likely to increase
with time. The two and five-cycle treat-
ments, however, resulted In much less
recovery of these four organics In soils.
For the five-cycle treatment 93, 97, 96
and 48% of the chloroform, benzene,
toluene and tetrachloroethylene added were
"unaccounted for." These volatile organ-
ics which were "unaccounted for" in these
experiments are thought to be lost by
volatilization from the soil surface,
although a change in the analytical
recovery as a result of freeze/thaw
cycling cannot be completely ruled out.
Metals
Figure 7 shows the concentration of
Zn, Hi, Cd and Cu In leachate from the
, unfrozen treatment after one and three
j cycles of freeze/thaw. In the leachate
from the unfrozen treatment, the metal
concentrations were generally In the order
Zn > HI > Cd > Cu. In the three columns
subjected to freeze/thaw cycles, the order
of metal concentration was similar, except
that Ni > Zn. Freezing and thawing sedi-
I6r—
40r—
30 —
10
40 8O
Elapsed Time (days)
120 0
10
ZO
30
Figure 7. Concentration of heavy metals in leachate from unfrozen treatment (A), after
one cycle of freezing (B), and during three cycles of freeze-thaw (C).
223
-------
Table 3. Added metals recovered in soils.
Freeze/Thaw
Cycles
0 (frozen
0 (unfrozen)
1
1
3
Cd
96
81
92
94
90
Cu
100
88
93
98
95
Ni
94
81
91
89
85
Zn
97
86
98
96
93
raents apparently increased the concentra-
tion of Ni and to a lesser extent Zn and
Cd leached out. On a mass basis, though,
leaching accounted for only a small pro-
portion of the added metal and most of the
metals added to the slurry were recovered
in the soil. More metals were recovered
from the frozen-thawed treatments than the
unfrozen treatment. Table 3 shows that
from SIX (unfrozen treatment) to 96%
(frozen treatment) of the added Cd was ^
recovered in the soils. The recovered Cu
in the soils ranged from 88% (unfrozen
treatment) to 100% (kept frozen). The re-
covery of Ni in soils ranged from 81% (un-
frozen) to 94% (kept frozen). The re-
covery of Zn ranged from 86% (unfrozen) to
97% (kept frozen).
It should be restated here that for
the column maintained frozen throughout,
the soil was frozen from the bottom and no
leaching and hence no loss of metals or
volatiles by leaching was possible.
SUMMARY AND CONCLUSIONS
The following conclusions were drawn
from this study:
a. Soil freezing and thawing have a
significant effect on the rate of consoli-
dation of slurries. Two cycles are suffi-
cient to achieve maximum consolidation.
b. The maximum volume reduction
achieved was 37% when the soils contained
176% water on a dry weight of soil basis.
c. The time required to achieve
maximum consolidation was about 22 days
when two freeze/thaw cycles were used
compared with 125 days if no freeze/thaw
was used.
d. Freeze/thaw enhanced the rate of
water percolation through soils due to
aggregate and channel formation.
e. There was no metal or volatile
organic input to ground water when soils
were frozen from bottom to top.
f. Freeze/thaw cycling apparently
enhanced volatilization of organics from
soil. Up to 93, 97, 96 and 48% of chloro-
form, benzene, toluene and tetrachloro-
ethylene, respectively, were removed from
soils by five freeze/thaw cycles.
g. Of the four volatile organics,
tetrachloroethylene was the least volatile
and the least mobile organic in soils.
ACKNOWLEDGMENTS
This work was financially supported
by the U.S. Environmental Protection
Agency (EPA), Containment Branch, Land
Pollution Control Division, Hazardous
Waste Engineering Research Laboratory,
Cincinnati, Ohio, and the U.S. Army Cold
Regions Research and Engineering Labora-
. tory (CRREL), under Interagency Agreement
DW 930180-01-0. The information contained
in this paper represents the authors'
opinions and not necessarily those of EPA
or CRREL. Citation of brand names is not
to be used for promotion or advertising
purposes. The authors wish to thank
Daniel Leggett and Edwin Chamberlain of
CRREL for their technical review of this
manuscript.
LITERATURE CITED
Iskandar, I.K. and J.M. Houthoofd. 1985.
Effect of freezing on the level of
contaminants in uncontrolled hazard-
ous waste sites. Part 1. Literature
review and concepts. Proceedings,
Eleventh Annual Research Symposium,
Cincinnati, Ohio, 29 April - 1 May,
1985.
Iskandar, I.K. and T.F. Jenkins. 1985.
Potential use of artificial ground
freezing for contaminant immobiliza-
tion. Proceedings International Con-
ference on New Frontiers for Hazard-
ous Waste Management, 15-18 September
1985, Pittsburgh, PA, p. 128-137.
224
-------
Gurka, D.F., J.S. Warner, L.E. Silvon, Kariekhoff, S.W., D.S. Brown and T.A.
T.A. Bishop, and M.M. McKown. 1984. Scott. 1979. Sorption of Hydrophob-
Interim Method for Determination of ic Pollutants on Natural Sediments.
Volatile Organic Compounds in Hazard- Water Research 13, p. 241-248. '
ous Wastes. J. Assoc. Off. Anal.
Chem. 67(4) 776-782.
Federal Register. 1984. Method 624 -
Purgeables. Vol. 49, No. 209, Fri-
day, October 26, 1984, Pg, 43373-
43384.
225
-------
FEASIBILITY OF USING MINED SPACE FOR LONG-TERM
CONTROL OF DIOXIN-CONTAMINATED SOILS
M. Pat Esposito, William E. Thompson, and Jack S. Sreber
PEI Associates, Inc.
Cincinnati, Ohio 45246
and
Janet M. Houthoofd
U.S. EPA Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
ABSTRACT
The secondary use of mined space in Missouri for warehousing and other purposes
strongly suggests that such spaces also could be successfully used for the long-term con-
trol of dioxin-containing soils and other hazardous wastes. This paper reviews the types
o'f mines typically converted to secondary uses in the State and the advantages these mines
offer over landfills as an alternative management option for hazardous wastes. Criteria
for siting a mine for such purposes and costs for development and operation of the facili-
ty also are briefly discussed.
INTRODUCTION
In the early 1970's, a waste-hauling
cftntractor removed 2,4,5-trichlorophenol
distillation wastes containing the highly
toxic contaminant 2,3,7,8-tetraehlorodiben-
zo-p-dioxin from a pharmaceutical/chemical
manufacturing plant in Verona, Missouri,
and mixed them with waste oils. Subse-
quently, these dioxin-tainted waste oils
were used as dust suppressants on dozens of
roads, parking lots, and horse arenas
across the State. At the present time, 42
sites have been confirmed to be contaminat-
ed with dioxin. An estimated 200,000 to
300,000 cubic yards (yd3)* or more of con-
taminated soil have been identified at 27
sites in the St. Louis area; also contami-
nated soil and debris (estimated at 50,000
yd3} have been identified in the Spring-
field area. Additional suspect sites are
being investigated, and as more are con-
firmed, the volume of soil requiring reme-
dial action will increase.
To date, cleanup efforts at contami-
nated sites have been severely hampered by
the lack of acceptable options for complete
elimination of the dioxin exposure hazard.
A proposal made in December 1983 to place
some of the contaminated soil (about 50,000
yd3} from the St. Louis area in a concrete
bunker at Times Beach, at a cost of $15.7
million, failed to gain public support.
This was to be a temporary storage measure,
as the proposed bunker itself (and all of
Times Beach) lies in the floodplain of the
Meramec River. Public opposition resulted
in the withdrawal of this plan in late
1984, however, and the contaminated soils
remain in situ. Currently efforts are
underway at two sites (Minker-Stout and
Castlewood) to excavate contaminated soil,
place it in large polypropylene sacks, and
temporarily store it in steel-sided ware-
house-type buildings specially constructed
on site for this purpose. Eventually, at
least the most highly contaminated soils
and those subject to migration via erosion
and runoff must be either detoxified
(through some chemical or physical means
such as incineration, stabilization, or
deehlorination) or permanently isolated in
a manner that will prevent any further
exposures.
Efforts have been underway at EPA's
Hazardous Waste Engineering Research
* Numerical data in this paper are presented in units corresponding to current American
usage. A metric conversion table may be found at the end of this paper.
226
-------
Laboratory in Cincinnati to develop sta-
bilization techniques for preventing dioxin
migration. At the same time, dust suppres-
sion techniques have been evaluated to se-
lect the most effective means of minimizing
fugitive dust losses during remedial soil-
handling operations (such as soil movement
during the stabilization process). In con-
junction with these efforts, EPA contracted
PEI Associates, Inc., to evaluate the feas-
ibility of permanently placing the dioxin-
contaminated soils in underground mined
spaces within the State of Missouri. Pre-
vious research and field experience in both
the United States and Europe had shown that
the underground placement and .management of
hazardous wastes in mines* is, in general, a
technically feasible and environmentally
sound alternative to landfilling. In fact,
the concept offers several distinct advan-
tages over landfill ing:
0 Leachate production in a properly
sited and maintained mine would
be minimal to nonexistent com-
pared with a landfill; as a
result, risks of ground-water
contamination would be effective-
ly eliminated.
0 Wastes would be accessible for
inspection. As a result, con-
tainer integrity could be as-
sured, and leaks, spills, and
releases of the wastes to the
environment could be prevented.
0 With proper control of natural
humidity in the mine, the ex-
pected lifetime of containers
used to hold the wastes could be
lengthened,
0 Incompatible wastes could be
easily grouped, segregated,
monitored, and controlled.
0 Discrete packages of wastes could
be retrieved at any time, if nec-
essary, for treatment, recycle,
etc., leaving other wastes undis-
turbed .
° Valuable land that could be used
for farming, housing, etc., is
not lost to landfilling; alterna-
tively, mined space that would
otherwise become abandoned and a
potential public hazard is main-
tained and utilized.
For this study, literature pertaining
to the geology, hydrology, and mining prac-
tices of the State, and previous secondary
uses of mined space was thoroughly re-
viewed. Also, project staff conducted
site-specific evaluations of various types
of mines in the State in order to fully
appreciate typical mining practices and
conditions at both active and inactive mine
sites.
PROFILE OF UNDERGROUND MINING PRACTICES IN
MISSOURI
Portions of Missouri have been inten-
sively mined since the 1800's, and signifi-
cant operations are currently engaged in
the underground extraction of metallic and
nonmetallic minerals. Many underground
mining operations, especially those con-
nected with the lead and zinc deposits
within the Old Lead District of southwest-
ern and southcentral Missouri, are old and
small. The larger underground mines in the
State consist of shallow (i.e., less than
100 ft deep) limestone, dolostone, and
sandstone mines, and deep (i.e., greater
than 100 ft deep) lead-zinc, iron, and
cobalt-nickel mines. The larger under-.
ground metal and sandstone mines are locat-
ed only in the southcentral to southeastern
portion of Missouri, generally within 50 to
60 miles of St. Louis. Sizable underground
limestone and dolostone mines are located
throughout the State, primarily in the
Kansas City, St. Louis, and Springfield
areas.
SECONDARY USE OF MINED SPACE IN MISSOURI
The secondary use of underground mined
space for commercial purposes is well
established in Missouri, having emerged in
the Kansas City area and grown steadily
over the past 50 years. It is reported1
that Missouri leads the world in secondary
use of underground mined space with roughly
20 million square feet already converted,
providing thousands of jobs. This under-
ground wealth of space is being used to
warehouse food products and perishable ,
consumer goods; some space is also used for
offices, small tool shops, tennis courts,
roller rinks, horticulture, gravel/aggre-
gate storage, and parking (see photos).
Mines dedicated to such uses can be found
throughout the State, principally near the
urban areas of Kansas City, Springfield,
and St. Louis.
227
-------
Mines converted to secondary use are
typically developed at shallow depths (less
than 100 ft) by room and pillar techniques
1n massive, thick-bedded horizontal lime-
stone , dolostone, or sandstone units that
are near the surface and hydrologically
Isolated (I.e., dry, above the saturated
zone). Thirty to sixty feet of overburden
protect such mines from the deteriorating
effects of freeze-thaw weathering. The
shallow subsurface construction allows easy
and direct access by car, truck, or rail
for moving inventories, personnel, and
customers in and out of the mine. Interior
concrete block partition walls, finished
concrete floors, permanent lighting, and
humidity controls combine to create famil-
iar and pleasant working environments that
are safe, dry, and secure. Properly sited,
constructed, inspected, and maintained, a
converted mine can be expected to remain
very dry and structurally sound for at
least as long as any conventionally built
aboveground building.
MINE STORAGE—AN ALTERNATIVE SOLUTION TO
MISSOURI'S DIOKIN AND OTHER HAZARDOUS WASTE
PROBLEMS
Given the well-established commercial
use of underground space in the State, and
speaking strictly from a technical point of
view, the placement of dioxin-contaminated
soils in underground mines in Missouri
appears to be a viable and safe concept for
protecting the population and environment
from further exposure to this hazardous
waste. The soils 1n question have been
extensively studied by EPA and others and
have been found to be chemically and physi-
cally stable (i.e., unreactive and immo-
bile) in the ground/soil environment. The
dloxin contaminant adheres tightly to the
soil particles, and is of no apparent
environmental hazard except when directly
contacted by the skin or inhaled/ingested
through windborn releases of contaminated
fugitive dust. Isolation of these wastes
1n a protective environment such as that
found in an underground mine should be a
safe and effective way of managing the
soils to prevent further exposures.
It is entirely possible that a suita-
ble existing underground mine could be
identified for rehabilitation and develop-
ment into a repository for these wastes.
Approximately 80 to 100 such mines ranging
in size from 1 to 100+ acres are in exis-
tence today in the State. Some are already
dedicated to commercial use; others are
abandoned and structurally unsound, or sub-
ject to flooding by ground water or nearby
rivers and streams. Still, there are
others that may be suitable for development
and use as waste repositories; a statewide
survey of all available mines would be
needed to identify viable existing candi-
date sites. Alternatively, it may be more
desirable to site, design, and develop a
new mine expressly for this purpose. The
development of a new mine may be cost-com-
petitive with the rehabilitation of an
existing mine and would overcome design,
structure, and location problems associated
with many existing mines.
In either case, the mine should be
situated along a natural drainage divide in
a massive low-permeability host rock such
as limestone/dolostone, well removed from
seismically active and populated areas, and
with direct access to a primary state
highway. The mine should be developed by
typical room and pillar methods to provide
structural stability as well as regular
room dimensions and pillar spacings.
External access should be limited to only a
few front portals, which could be converted
to secure loading/unloading dock areas. At
a minimum, the bare mine should be finished
through 1) scaling and removal of all loose
rock; 2) installation of rockbolts as
needed; 3) grading of the subbase limestone
floor; and 4) installation of drip pans to
collect infiltration. Additionally, it may
be desirable to construct a reinforced
concrete floor with drainage trenches and
sumps, apply an impervious sealant/coating
to the concrete floor, shotcrete the walls
and ceiling, and install electrical, water
treatment, and docking systems.
Wastes should be stabilized and/or
packaged at the point of origin, transport-
ed to the mine by truck or rail, and ware-
housed in the prepared mine according to a
predetermined plan to allow adequate aisle
space for inspections and maintenance. Two
options seem most viable: 1) the wastes
could be mixed with stabilizing agents,
cast as blocks, and then shrink-wrapped in
plastic, or 2) the wastes could be placed
directly (unstabilized) in containers such
as supersacks for transport and storage.
From a regulatory point of view, the
site would have to be initially permitted
under the Resource Conservation and Recov-
ery Act (RCRA) as either a demonstration
228
-------
test facility or as a hazardous waste stor-
age facility. At this time, the facility
could not be permitted as a hazardous waste
disposal site because EPA has not estab-
lished regulatory operating/design stan-
dards for disposing of hazardous wastes in
mines. Under a RCRA storage permit, the
facility would be required to provide a
containment system for collecting and re-
moving liquids resulting from precipitation
infiltration. Water seepage into the mine
from precipitation infiltration through the
overburden is expected to be of little con-
sequence, and can be conveniently diverted
from the mine interior and its contents by
passive collection systems consisting of
drip pans suspended from the ceiling di-
rectly under the seep, and diversion piping
leading to a central sump and discharge
point. Surface grading above the mine can
also be used to minimize moisture infiltra-
tion into the mine. The facility would be
required to have an NPDES permit for dis-
charge of any seepage from the mine, and
must meet all applicable OSHA standards.
During active mining periods (such as
during construction of a new mine or during
expansion of an existing mine), compliance
with the Mine Safety and Health Act (MSHA)
would be required.
Implementation of the concept will
require the following sequence of steps:
1) an accurate determination of the quanti-
ty of contaminated soil to be placed in the
mine; 2) selection of the best soil han-
dling and placement scheme based on the
stabilization decision and further study of
available packaging/packing options; 3) de-
termination of the mine size needed to
accommodate 1) and 2) above; 4) identifica-
tion and purchase of a specific parcel of
land or mine for development of the facili-
ty; 5) development of a facility plan;
ty;
5)
6) construction of the facility; 7) trans-
fer of the contaminated soils to the facil-
ity; and 8) long-term monitoring/mainte-
nance of the site.
Preliminary capital costs have been
calculated for development of a range of
facility sizes capable of handling 10,000
to 100,000 yd3 of contaminated soil in a
new or existing limestone mine. Mine
preparation and outfitting costs range from
between $0.8 and $4.22 million for existing
mines to between $3.8 and $20.6 million for
a new mine. These costs do not include
those handling costs associated with excava-
tion, packaging, transport, and placement
of the soil. The handling of soils in
supersacks is expected to cost between
about $150 and $225 per yd3; stabilized
blocks will cost about $125 to $300 per
yd3. Estimates of direct annual operation
and maintenance costs (i.e., utilities,
equipment, labor, maintenance, and over-
head) ranged from $139,000 to $180,000 per
year. Annual costs calculated do not
include depreciation, taxes, insurance, and
capital charges.
CONCLUSIONS
Mine storage is a technically feasible
and environmentally sound alternative to
landfill ing for the long-term management of
dioxin-contaminated soils and possibly
other types of hazardous wastes in Mis-
souri. , The concept offers distinct techni-
cal advantages over landfill ing such as
little or no leachate generation, the
ability to inspect and retrieve individual
waste packages/containers at any time,
potentially longer container/packaging
life, total isolation of the waste, and
land preservation. Based on preliminary
cost estimates, mine storage appears to be
less costly than the previously proposed
bunker method of land disposal.
In Missouri, shallow (less than 100 ft
deep) underground limestone/dolostone mines
possess the best overall set of character-
istics for development and long-term opera-
tion and maintenance of such a facility
because of their hydraulic setting, size
and structural stability, ease of access,
monitorability, and placement in massive
low-permeability, chemically compatible
rock. Further onsite engineering and tech-.
nical studies of existing limestone/dolo-
stone mines in the State are needed to
identify the best candidate (or candidates)
for development as a hazardous waste repos-
itory. These onsite studies should in-
clude, for each site, a specific analysis
of the host rock's condition, stability,
lateral extent, and vertical extent, rock
mechanics, mine design, local and area
hydrology and hydrogeology, surrounding
land use, presence of natural dissolution
cavities or caves in the neighboring vicin-
ity, availability of road and rail access,
and ownership of the property and its
mineral rights. If no existing mines are
found to be totally satisfactory, alterna-
tive efforts should be directed toward the
identification of a suitable land parcel
for development of a new mine. Onsite
229
-------
evaluations of candidate land parcels for
such development should include the same
study considerations as those outlined
above for an existing mine. Once these
studies are complete and the most promising
candidate site has been identified, further
study of the capital and O&M costs associ-
ated with the final design and development
of the facility at the site of choice will
be required.
SELECTED BIBLIOGRAPHY
1. Stone, R.B., et a!., Using Mined Space
for Long-Term Retention ofNqnradio-
active Hazardous Haste , Volume 1,
EPA-6QQ/2-85-021.O7 EPA Hazardous
Waste Engineering Research Laboratory,
Cincinnati, Ohio, 1985, PB85-177111/
AS.
2. Stone, R.B., et a1.» Evaluation of
Hazardous Hastes Emplacement in Mined
Openings. EPA-600/2--75-040.U.S. EPA,
Hazardous Waste Engineering Research
Laboratory, Cincinnati, Ohio, 1975.
3. Stauffer, T., Jr., "Grain, Seed, Food
Storage and Farm Machinery Manfuactur-
,ing—Kansas City's Use of Underground
Space in the U.S. Agricultural Mid-
west," Proc. of International Symposi-
um on Subsurface Space (Rockstore 80),
Stockholm, June 1980.Pergamon Press,
Inc., New York, New York, I, 1981.
4. Stauffer, T.P., Sr., "Kansas City's
Use of Limestone Mines for Business,
Industry, and Storage," Proc. of the
First International Symposium on
Storage in Excavated Rock Caverns
JRockstore 77), Stockholm, September
1977.Pergamon Press, Inc., New York,
New York, I, 1978.
5. Oohnsson, 6., "Undertage-Deponie
Herfa-Neurode. Hazardous Waste Em-
placement in Mined Openings—4 Years
Experience in the Disposal of Special
Industrial Wastes in a Mined-out
Section of a Potash Mine in the Werra
Basin {Summary)," Proc. of the First
International Symposium on Storage_in
jpccayatetf Rock Caverns (Rockstore 77),
Stockholm, September 1977.Pergamon
Press, Inc., New York, New York, I,
1978.
6. Oohnsson, 6., "Underground Disposal at
Herfa-Neurode," Proc. of the NATO CCMS
Symposiumon Hazardous Maste Disposal,
Washington, D.C., October 1981.
Plenum Press, New York, New York,
1983.
7. U.S. Department of Energy, TheMoni-
tored Retrievable Storage Concept—A
Reviewof Its Status and Analysis ~of
Its Impacton the Waste Management
SystemV DOE/NG-Q019.Office of
Nuclear Waste Management and Fuel
Cycle Programs, Washington, D.C.,
1982, 3-1 to 3-21.
8. Whitfield, J.W., Underground Space
Resources in Missiouri, Missouri
Department of Natural Resources,
Division of Geology and Land Survey,
Rolla, Missouri, 1981.
9. "Economics of Mined Repositories for
Acutely Hazardous Wastes," The Hazard-
pus Waste Consultant, September/Octo-
ber 1983, 32-36.
10. Forsberg, C.W., "Disposal of Hazardous
Elemental Wastes," Environmental
Science...and Technology, February 18,
1984, 56A-62A.
11. Hallowell, J.B., DeNiro, D.P., and
Lawson, J.S., Jr., Comparative Assess-
ment of Alternatives for Waste Dispos-
al and Storage, Batte'TTe 'Columbus
Laboratories, Columbus, Ohio, 1984.
METRIC CONVERSION TABLE
1 mile = 1.609 kilometers
1 cubic yard = 0.7646 cubic utter
1 acre = 4047 square meters or 0.4047
hectare
1 foot = 0.3048 meter
ACKNOWLEDGMENT
This study was conducted under EPA
Contract No. 68-02-3693, Work Assignment
No. 13. The authors are especially thank-
ful to Dr. James Hadley Willams and his
staff of geologists at the Missouri Depart-
ment of Natural Resources for their assist-
ance in scheduling and conducting mine site
visits, and to Jim Vandike, also of the
Missouri DNR, for the use of his photos.
230
-------
Tennis court developed in
underground limestone mine.
Tire warehouse inside
sandstone tunnel.
Natural interior of inactive
limestone mine near St. Louis.
231
-------
Roller rink developed 1n
underground sandstone mine.
Office area Inside limestone
mine warehouse.
Active underground limestone
mining operation In which
portions are used for gravel/
aggregate storage.
232
-------
Docking area for rail freight
inside limestone mine.
Truck dock receiving area inside
limestone mine.
233
-------
Rail entrance to underground
limestone mine warehouse.
Product storage area 1n underground
limestone mine warehouse.
234
-------
REVIEW OF ALTERNATIVE TREATMENT PROCESSES FOR NON-HALOGENATED SOLVENT WASTES
Benjamin L. Blaney
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
There are both economic and regulatory incentives to treatment of wastes whose or-
ganic component makes them hazardous. This paper discusses management alternatives for
the first set of non-halogenated wastes streams to be regulated under EPA's land disposal
restriction program, the three spent solvent waste categories: F003, F004 and F005.
Three management alternatives for such wastes are addressed: recycle/reuse; destruction;
and treatment followed by land disposal. The influence of physical waste characteristics
and organic content upon the selection of waste treatment options is described.
INTRODUCTION
In November 1984, Congress amended the
Resource Conservation and Recovery Act
(RCRA) to restrict the land disposal of a
number of hazardous wastes. Various types
of waste will be banned from land disposal*
unless the EPA determines that: 1) there
are land disposal options which do not
threaten human health and the environment,
or 2) there are waste treatment options
which will alter the waste and thereby
minimize the risk of its land disposal (1).
The first two wastes are scheduled to be
banned in November 1986. They are dioxin
and solvent wastes.
This article characterizes the non-
halogenated solvent wastes which Congress
listed for banning and describes techniques
for treating these wastes, either to make
them amenable to land disposal, or as al-
ternatives to land disposal. Other sym-
posium papers describe treatment processes
for dioxins (2) and for halogenated or-
ganics including halogenated spent solvents
(3). For a more detailed discussion of
alternative techniques for managing both
halogenated and nonhalogenated solvent
wastes, see Reference 4.
SOLVENT WASTES
Types/Characteristics
The 1984 RCRA Amendments specify five
categories of solvent wastes which are to
be banned from land disposal. They have
EPA Waste Codes F001, F002, F003, F004 and
F005. These wastes are comprised of
certain spent halogenated (F001 and F002)
and nonhalogenated solvents (F003 - F005),
along with sludges or still bottoms from
the recovery of these solvents and certain
spent solvent mixtures or blends.
Table 1 lists those non-halogenated
solvents which are included in categories
F003 to F005. EPA Code Number F003 wastes
were listed because of their ignita-
bility.** Code Numbers F004 and F005
*The Amendments excluded deep well in-
jection from the ban until after November
1988.
** F003 wastes have already been banned
from landfills because of their ignit-
ability, but will be considered here be-
cause they have not been banned from
land disposal in general.
235
-------
TABLE 1. REGULATED NON-HALOGENATED WASTE SOLVENTSl
EPA Waste Code
Solvents^
F003
F004
F005
Spent non-halogenated solvents: xylene,
acetone, ethyl acetate, ethyl benzene, ethyl
ether, methyl isobutyl ketone, n-butyl
alcohol, cyclohexane and methanol.
Spent non-halogenated solvents: cresols and
cresylic acid, and nitrobenzene.
Spent non-halogenated solvents: toluene,
methyl ethyl ketone, carbon disulfide,
isobutanol, and pyridine.
1. Source: Reference 5.
2. Besides spent solvents containing these compounds, still bottoms
from the recovery of these spent solvents are also included under
each of the three waste codes. Also included are all spent solvent
mixtures/blends containing, before use, a total of 10 percent or more
(by volume) of one or more of the compounds listed under a given waste
code.
wastes were banned because of the toxicity
of the of the non-halogenated solvents
which they contain.
Characteristics
Large volumes of non-halogenated and
halogenated solvent wastes are generated
annually in the United States. It is es-
timated that in 1981, 3.1 billion gallons
of both types of solvents wastes were gen-
erated. Of these, approximately 1.2
billion gallons would be covered by pro-
posed land disposal restrictions. Solvent
wastes are usually organic liquids or
sludges (e.g., still bottoms) at the point
of generation. However, in some instances
they become mixed with water between the
point of generation and disposal, and the
resultant aqueous waste streams are in-
cluded in the solvent waste category.
The EPA estimates that approximately
214 million gallons of both halogenated and
non-halogenited solvent wastes which is
currently 1and disposed will require
treatment. This is comprised of 185
million gsilons of solvent-water mixtures,
14.7 million gallons of organic liquids,
7.5 million gallons of organic sludges or
solids and 6.7 million gallons of inor-
ganic sludges or solids (5). The fraction
of these wastes which are non-halogenated
are not currently available.
The treatment of the solvent-water
mixtures, organic liquids and organic
sludges and solids which fall under waste
codes F003, F004 and F005 will be the focus
of this paper.
Degree of Required Treatment
Land disposal of these wastes may not
eliminate the health effects and the flam-
mability of F003 solvents for which they
were originally listed. The low molecular
weight organic constituents of solvent
wastes may react with synthetic liners
used in landfills and surface impoundments,
thereby reducing liner integrity. Their
low molecular weight also makes the organic
constitutents mobile and some have been
shown to readily pass through liners. In
addition, these compounds are all volatile
to some degree and will be emitted to the
air at the disposal site. The EPA is con-
sidering all these factors in determining
236
-------
to what degree wastes must be treated in
order to minimize the risks of land
disposal. The EPA has proposed concen-
tration "screening levels" of 2 parts per
million (ppm) or lower for solvent constit-
uents in land disposed wastes. The 2 ppm
level is a "safety cap" to solvent constit-
uent concentrations based on liner integ-
rity studies. Lower screening levels exist
for several compounds inside to protect
against the health hazard which they could
pose if released from disposal facilities.
The screening levels would set the maximum
concentrations allowed for liquid wastes
with under 1% solids which are land dispos-
ed (e.g. of wastewaters in surface im-
poundments). For other wastes, an extrac-
tion procedure is followed and the extract
analyzed for solvents content.
WASTE TREATMENT ALTERNATIVES
There are three principal options to
direct land disposal of solvent wastes.
1. Destructive treatment, including
thermal oxidation;
2. Removal of toxic or flammable
constituents prior to land
disposal;
3. Recycle/Reuse, including use as a
fuel substitute; and
A fourth option to land disposal is reduc-
tion or elimination of waste generated.
Such waste minimization is being pursued by
many generators, since it is often prefer-
able to the increasing costs of waste
management, but will not be considered here
because it is generally very specific to
each manufacturing process. EPA has on-
going studies to address these waste min-
imization options.
The choice of treatment technique will
be dependent upon several factors, in-
cluding: waste composition, waste volume
and treatment cost. This paper addresses
the influence of waste composition on
choice of treatment technique. First, we
discuss general procedures for deciding how
to treat three broad categories of solvent-
bearing wastes: 1) aqueous and mixed
aqueous/organic liquids, 2) organic liquids
and 3) sludges. Then we describe a number
of specific treatment techniques which are
applicable to solvent wastes.
For the purposes of this paper, we
distinguish between streams of very low
solids content and those with greater than
one or two percent solids content, which we
define as sludges. We define streams of
low (<1%) solids content as aqueous, mixed
aqueous/organic or organic streams.
"Aqueous streams" are those which have a
water content of 99% or higher, while or-
ganic streams are defined as containing 50%
or more organic liquids. Mixed aqueous/
organic streams fall in between.
There will always be some ambiguity
regarding the choice of one of these three
classifications for a given waste, and the
exact composition of the waste stream and
economic considerations will dictate what
treatment steps are most important. How-
ever, categorization of solvent wastes into
these three groups helps to structure the
following discussions of available the
treatment options.
In the following discussion, the waste
treatment techniques are broken out by the
concentration of organics in the waste
which they are typically used to treat.
Although there is not a clean cut catego-
rization of treatment techniques based on
influent waste stream organic concentra-
tion, three general categories of con-
centration are used: high (10 to 100
percent), low-medium (0.1 to 10 percent),
and low (below 1000 parts per million).
The discussion of each of these tech-
niques will include a brief description of
its operating principles and its current
uses for-solvent waste treatment. A de-
scription of the limitations associated
with using these techniques to treat
solvent wastes, if any, is also provided.
Streams of High Solvent Concentration
Incineration
Incineration is the thermal oxidation
of wastes at high temperatures for the
purpose of destroying the material. There
are approximately 200 hazardous waste
incinerators operating in the United States
under permits which require them to destroy
waste stream principal organic hazardous
constituents (POHCs) with an efficiency of
at least 99.99%.
There are two principal types of haz-
ardous waste incinerators in use in this
237
-------
country; liquid injection and rotary kiln.
A liquid injection system consists of one
or more refractory lined combustion
chambers into which waste and any required
fuel are injected through atomizing
nozzles. The combustion gases are cooled
and pass through a series of pollution
control devices prior to release to the
atmosphere. Liquid injection incinerators
can only handle atomizable wastes and are
not generally designed for wastes which
will leave large amounts of ash residue
(6, 7).
A rotary kiln incinerator is a re-
fractory-lined cylinder mounted at a slight
incline and followed by an afterburner.
The kiln mixes waste and combustion gases
through its rotary motion, converting waste
to gases and an ash residue through partial
burning and volatilization. The gases then
pass to the high temperature afterburner
where the combustion reactions are com-
pleted. Rotary kilns are capable of han-
dling solids, sludges, liquids or gaseous
wastes (6» 7).
Table 2 summarizes the applicability
of incinerators to non-halogenated solvent
wastes. Essentially all solvent wastes of
high organic content should be amenable to
incineration in either liquid injection or
rotary kilns. In fact, it is estimated
that 60% of the wastes being incinerated
in 1981 were solvent derived wastes (8).
While the majority of these incinerated
solvents were nonhalogenated liquids,
solvent sludges were also disposed of
through rotary kiln incineration.
The EPA uses the heat of combustion of
Appendix VIII compounds as an indicator of
their incinerability (9). Solvents listed
under F003 are obviously readily in-
cinerable, since they are classified as
hazardous due to their ignitability.
Compounds under waste codes F004 and F005
have heats of combustion of at least 5.50
kcal/grams (for nitrobenzene). Compounds
with such heats of combustion have been
found to be ignitable (10).
While the heating value of wastes de-
creases with increased content of water and
other inorganics, aqueous/organic mixtures
and even aqueous wastes can be incinerated.
Supplementary fuel can be used to produce
the required combustion chamber temperature
or the waste stream can be blended with
high Btu wastes to provide an adequate
heating value of the Incinerator feed. In
some cases, operators mix wastes of high
water content with other highly flammable
wastes in order to limit the combustion
chamber temperatures to within design
specifications, while maintaining high feed
rates. Alternatively, water can serve as a
source of hydrogen to promote more complete
combustion of high molecular weight hydro-
carbons (8).
The ashes from waste incineration must
be analyzed to determine whether or not
they are hazardous. Non-hazardous ashes
can be disposed of In landfills following
standard procedures for such wastes, while
hazardous ashes will have to be stabilized
prior to disposal at a RCRA-permitted site.
In addition, particulate control of
equipment must be adequate to meet existing
emissions regulations.
Agitated Thin-Film Evaporation
Evaporation is the removal of a sol-
vent (including water) as a vapor from a
solution or from a sludge of low viscosity.
Agitated thin-film evaporation relies on
the exposure of a large surface area of
heated waste to atmospheric or vacuum
conditions to enhance the separation of
more volatile constituents from the waste
by evaporation. Waste is spread on the
walls of a cylindeMcal or tapered tube by
an assembly of wiper blades. A heated
outer shell elevates the temperature of the
waste.
Table 3 summarizes the applicability
of agitated thin-film evaporators to non-
halogenated solvent wastes.
Agitated thin-film evaporators are the
"work horse" of many solvent waste recycl-
ing firms. They provide the first step in
the processing of those solvent wastes
which are suitable for recycling, i.e.,
those with sufficiently high percentage of
recoverable organics (typically over 60-
70%). Wastes are treated by agitated
thin-film evaporation to separate the low
boiling solvents from solids and high
boiling organics. The overhead fraction is
often processed further by distillation,
water extraction, etc., prior to reuse.
Thin-film evaporators can be used to
treat liquid wastes and sludges of of
low solids content. Waste viscosities
238
-------
TABLE 2. INCINERATION OF NONHAL06ENATED SOLVENT WASTES
Liquid Injection Rotary Kilns
Treatment Objective
Thermal Oxidation of Hydrocarbons to H20 and C02
Principal Application
Destruction of Organic
Bearing Liquids
Destruction of Organic
Bearing Liquids and
Solids
Pri nci pal Waste Categorles
Organic Liquids Sludges
Aqueous/Organic Mixtures Organic Liquids
Aqueous/Organic Mixtures
Restrictions on Waste
Characteristics
1. Low-medium viscosity None
2. Atomizable
Compound-Specifi c
Restrictions
None
None
TABLE 3. APPLICABILITY OF AGITATED THIN-FILM EVAPORATION
TO NON-HALOGENATED SOLVENT WASTES
Treatment Objective
Removal of solvent as a vapor from
a solution or a sludge
Pr1nci pal Appl1cation
Gross separation of low boiling solvents
from sludges and from high boilers in
sol vent recovery
Principal Haste Categories
Sludges
Organic Liquids
Restrictions of Waste
Characteristics
1. Viscosities below 10,000 centepoise
2. Solid's size less than blade clearance
3. No reactive constituents
Compound-Speci fie Restrictions
Organic liquids and sludges: Low solvent
boiling point
239
-------
generally cannot exceed 100 poise and the
size of solids 1n a sludge is limited by
the clearance of the agitator blades (typi-
cally less than 2.5 mm). Reactive wastes,
such as those which will polymerize, cannot
generally be treated using thin-film evapo-
rators because of the elevated tempera-
tures. However, some agitated thin-film
evaporators are designed to minimize the
waste residence time in order to reduce
this problem (11).
Agitated thin-film evaporators are
best suited for separation of the organic
solvent from a sludge, although they have
been used to separate low boiling organics
from oils in fuel oil clean up (12). If
the waste stream contains water, this
compound will generally be evaporated, too.
For this reason, thin film evaporation is
probably not readily applicable to the sep-
aration of organic solvents from mix aque-
ous/organic streams which contain little
solid materials.
Fractional Distillation
In the most general sense, the term
distillation describes the separation of
two or more components of a liquid mixture
by the vaporization and recovery of the
more volatile compounds. Distillation in-
volves application of heat to a liquid
mixture, vaporization of part of the mix-
ture, and removal of heat from the vapor-
ized portion to condense it. The result-
ant distillate is richer in the more vola-
tile components, and the residual bottoms
are richer in the less volatile mate-
rials.
Fractional distillation utilizes a
series of liquid-vapor equilibria inside a
vertical tower to provide a vapor which has
progressively higher concentrations of the
most volatile component. This produces an
overhead product which is more enriched in
low boiling components than is possible
with just a single evaporation and conden-
sation operation.
Table 4 summarizes the applicability
of fractional distillation to nonhalogen-
ated solvent streams. The solids content
of a distillation stream is limited, since
solids may clog the column or the coils in
the reboiler. For this reason, many waste
recycling firms use thin-film evaporators
or settling tanks to separate out solids
prior to fractional distillation.
Fractional distillation is frequently
found at solvent recycling firms. Such
units provide separation of solvent compon-
ents from liquid waste streams for solvent
recycling. They are particularly used to
produce high purity distillates, such as
are required for electronic component manu-
facturing.
Since the volatile components in the
overhead of a fractional distillation
column will be more concentrated than in a
single stage evaporator, their concentra-
tions in the process bottoms will be even
lower. Fractional distillation is used by
recycling firms to separate different or-
ganic components. It is also used to
remove miscible organics from aqueous/or-
ganic mixtures in order to reduce the
concentration of organics in their waste-
water discharge to levels acceptable to
municipal treatment facilities. Organic
concentrations below 1000 ppm have been
obtained in such instances.
For example, recent field tests by the
U.S. Environmental Protection Agency at one
recycling firm studied the batch fractional
distillation of two mixed aqueous/organic
streams. One contained approximately 95
percent water; 3 percent methyl ethyl ke-
tone; and 2 percent alcohol, chlorinated
hydrocarbons and other organics. The
second stream contained 77 percent water,
21 percent acetone, and 2 percent other
organics. Fractional distillation of these
two streams resulted in 99 percent removal
of the organics in the aqueous bottoms
(13).
Streams of Low to Medium Solvent
Concentration
Steam Stripping
Steam stripping is a form of simple
distillation in which steam is injected
Into a waste in order to vaporize the more
volatile components. Steam stripping is
not generally used for streams of high or-
ganic content, since the processor usually
wants to avoid further water contamination.
Steam stripping can be performed 1n
the batch or continuous mode. In the batch
process, waste is charged to a boiler and
steam is injected directly into the waste.
The injection of live steam both heats the
waste to volatilize low boiling components
and creates turbulence in the waste, thus
240
-------
TABLE 4. APPLICABILITY OF FRACTIONAL DISTILLATION
TO NON-HALOGENATED SOLVENT WASTES
Treatment Objectives
Separation of constituents in liquid
stream
Prlnci pal Appl 1 cations_
1. Separation of organic constituent
from each other
2. Removal of organics from water
Principal Haste Categories
Organic Liquids
Aqueous/Organic Mixtures
Restrictions on Waste
Low dissolved or suspended solids content
Compound-Spec1f1c Restr1ct1ons
Aqueous/Organic Mixtures; Organics are
water soluble
increasing the rate of volatilization. The
gases which are condensed from a steam
stripper will contain water along with the
more volatile organic components of the
waste in the form of a two phase mixture.
This mixture is decanted and the organic
component is drawn off for reuse or dis-
posal. The aqueous layer is fed back to
the stripper for further treatment.
In continuous steam stripping, waste
flows down a column while steam flows up.
The column is designed to promote heat
transfer from the steam to the waste, to
cause turbulence in the waste and to create
a large waste surface area. All of these
properties promote transfer of volatile
components from the waste to the gas phase.
Different liquid-vapor equilibria exist in
the column, with the highest relative
concentration of the most volatile compon-
ents found at the top.
Table 5 summarizes the applicability
of steam stripping to non-halogenated
solvent waste streams.
Both steam strippers and batch distill-
ation units can be used to remove volatile
organics from aqueous or mixed organic
waste streams. Batch distillation can be
used to separate the various organics from
each other as part of the stripping proc-
ess. Such a separation will not occur with
steam stripping, but if separation of the
organics in the overhead is unimportant,
steam stripping is the preferred mode of
operation since it has lower capital costs
and is generally cheaper and easier to
operate. In addition, batch steam strip-
pers are not as easily fouled as batch
distillation reboilers because there are no
coils involved and the steam maintains the
waste in a turbulent state.
One major drawback to steam stripping
is the wastewater residual that is produced
after decanting. If the solvent to be re-
moved from the waste does not readily di s-
solve in water, then the organic concentra-
tion of the aqueous decantent is low enough
that further treatment of this water is
either not needed, or can be accomplished
by simply mixing the material with the
feed. However, as the soluability of the
organic increases, more steam must be in-
jected into the system to accomplish strip-
ping and the rates of contaminated water to
separated organic in the condensate in-
creases, reducing the efficiency of the
stripping process. Therefore, steam strip-
ping for removal of non-halogenated sol-
vents from aqueous streams is most suited
for the following compounds: ethyl ben-
zene, toluene and xylene.
Liquid-phase carbon adsorption removes
organics from dilute aqueous streams by
adsorbing them onto an activated carbon
matrix of high surface area. This matrix
can then be regenerated using thermal oxi-
dation, which destroys the organics, or
241
-------
TABLE 5. STEAH STRIPPING
Treatment Objective
Separation of organics from water
Principal Appl1cations
1. Solvent recovery
2. Wastewater treatment
Principal Haste Categories
Aqueous
Restrictions on Haste
Characteristics
1. Continuous operation:
content
2. >lppm organics
low solids
Compound-Spedf1c Restri ctions
Water solubility
using a non-destructive process (e.g. steam
stripping), which allows for reclamation of
the organics. Regeneration has found in-
creased use as disposal of spent carbon
becomes more expensive.
Table 6 shows the applicability of
carbon adsorption to the treatment of
non-halogenated solvent streams. Carbon
adsorption has been used extensively to
remove organic pollutants from drinking
water and wastewater streams. It is has
been applied to aqueous streams with or-
ganic solute concentrations up to one
percent, but is generally used for streams
of less than 1000 parts per million (ppm)
organics.
Some pretreatment of the waste stream
may be required prior to processing through
carbon. Suspended solids in the waste in-
fluent generally are reduced to less than
50 ppm by settling and filtering, although
levels of up to 2,000 ppm have been handled
1n specially designed units.
Oil and grease in the waste stream
should be less than 10 ppm to avoid carbon
fouling. High levels of dissolved inor-
ganics may cause scaling or loss of carbon
activity during thermal carbon reactiva-
tion, but in many cases such problems can
be minimized through pH control and soften-
ing of the waste influent or by acid wash-
ing of the carbon before reactivation.
The degree to which carbon adsorption
can be used to remove small concentrations
of solvents from wastewater stream will be
dependent on a number of factors, including
the compound(s) to be removed, concentra-
tions of other organics in the stream and
choice of carbon material. Most of the
compounds listed as solvents are Priority
Pollutants and studies of these pollutants
show that in distilled water their affinity
for carbon varies over two orders of magni-
tude (15).
Biological Treatment
A number of biological degradation
processes have been used to destroy organic
pollutants in industrial and municipal
wastewaters. These include: activated
sludge, aerated lagoons, trickling filters,
and stabilization ponds.
There are several factors which must
be considered when comparing the biological
degradation techniques applicable to aque-
ous waste streams, including: efficiency of
degradation, required holding time and sus-
ceptability to destruction of the microbial
environment. Aerated lagoons, stabiliza-
tion ponds and trickling filters have been
found to be resistant to microorganism de-
struction due to variation in influent
stream composition. However, they each
have a drawback when compared to activated
sludge. Trickling filters are generally
not efficient enough to be used alone for
biodegradation; they are usually followed
by some other biological treatment process.
Trickling filters are often used upstream
of activated sludge systems to provide a
a more uniform influent composition.
Aerated lagoons also have lower removal
242
-------
TABLE 6. CARBON ADSORPTION
Treatment Objective
Separation of organlcs from water
Principal Application
1. Drinking water purification
2. Wastewater polishing
Pri nci pal Waste Categorl es
Aqueous
Restrictions on Waste
Characteristics~
Compound-Spec if1c Rest ri ct i ons
1. Low suspended solids
2. < 1,000 ppm organics
3. Little oil and grease
4. Other organics in waste
may interfere
None
efficiencies than activated sludge systems
and are less flexible in maintaining efflu-
ent limitations under varied influent
loading. Stabilization ponds are more
sensitive to inorganics and solids than are
the other three processes and require sub-
stantial land acreage and suitable climatic
conditions (16).
Activated sludge treatment is exten-
sively used in industry and is probably
the most cost-effective method of destroy-
ing organics present 1n aqueous waste
streams. It consists of a.suspension of
aerobic and facultative microorganisms
maintained in a relatively homogenous state
by mixing or by the turbulence induced by
aeration. These microorganisms oxidize
soluble organics and agglomerate colloidal
and particulate solids in the presence of
dissolved molecular oxygen.
Table 7 indicates the general applica-
tion of activated sludge systems to non-
halogenated solvent stream treatment.
Activated sludge systems have been
used extensively by industry for waste-
water treatment, including those which pro-
duce solvent-bearing wastes: organic
chemicals manufacturing, petroleum refin-
ing, paint and ink formulation, and gum
and wood chemicals. In general, the sys-
tems should be able to readily degrade
alkanes, alkenes, and aromatics. Potenti-
ally significant amounts of highly volatile
solvents may be released from the system
through volatilization, especially if the
system is not properly operated (e.g., well
mixed). In such instances, covered aera-
tion basins with air pollution control may
be required. If there are metals in the
influent stream, these will often be con-
centrated in the waste sludge, posing
problems for proper documentation or land
disposal of this by-product.
The activated sludge system treats
aqueous organic streams having less than
one percent suspended solids. The appro-
ximate upper limit to biological oxygen
demand (BOD) of the influent stream which
can readily be handled is 10,000 mg/1. The
biodegradation process produces C02, water,
organic products of partial biodegradation
and microorganism cellular material (17).
Activated sludge systems cannot toler-
ate high fluctuations in influent concen-
trations, but a number of steps can be
taken to reduce this problem. Neutraliza-
tion and equalization of the waste stream,
as well as suspended solids removal, usual-
ly proceed the activated sludge system.
Trickling filters may also be employed up-
stream to reduce variations in waste stream
composition. In addition, two operating
characteristics of the process itself
promote tolerance to such variations.
First, activated sludge systems can be
designed to thoroughly mix pollutants in
the aeration basin. Secondly, recircula-
tion of the activated sludge at 25 to 100
percent of the influent flow rate greatly
helps to acclimate the biomass in the
lagoon to the influent pollutants (17, 18).
243
-------
TABLE 7. APPLICABILITY OF ACTIVATED SLUDGE
SYSTEMS TO NQN-HALOSENATED SOLVENT WASTES
Treatment ObJect i ves
Separation of organlcs from water
Princi pal Appllcation
Wastewater Purification
Principal Haste Categories
Aqueous
Restriction on Haste
Characteristics
1. Low Suspended Solids
2. < 10,000 ppm organics
3. Metals and other toxics
not present
4. Uniform waste stream composition
Compound-Specific Restrictions
Volatile solvents may be stripped
from system in significant amounts
Solvent concentrations less than
those which are toxic to acclimated
microorganisms
SUMMARY
The toxicity, mobility and/or flamm-
abllity of non-halogenated solvent wastes
makes it difficult to land dispose of them
without significantly endangering the
environment. Increased costs for this ap-
proach to disposing of these wastes and
regulations which severly restrict such
disposal will force waste generators to
either reduce solvent generation rates or
to utilize waste treatment or recycling as
solvent waste management alternatives.
Alternatives to direct land disposal
appear to exist for all non-halogenated
solvent wastes, although the choice will
be dependent upon waste composition, produc-
tion rate and economics. Some techniques,
such as incineration, are from a technical
standpoint applicable to most solvent
wastes, regardless of their composition.
Recycling of solvent wastes which contain
sufficiently high amounts of liquid or-
ganics to make recovery economical is
presently practiced widely in the United
States. Similarly, wastes with sufficient
BTU value are being used in a number of
locations as fuel substitutes.
Mixed aqueous/organic liquid wastes
may also be recycled if their organic
component is sufficiently large to make re-
cycling economical. Solvent constituents
can be removed from aqueous streams by sim-
ilar recycling techniques, but the cost
will generally exceed the value of the re-
covered organics. Under anticipated regul-
atory incentives, such physical separation
techniques may soon be the most economic
means of disposing of aqueous and mixed
aqueous/organic solvent wastes. Alterna-
tively, these low concentration organics
can be destroyed through chemical or bio-
logical treatment processes, or by inciner-
ation.
Sludges of low solids content can be
disposed of in the same way as liquid
wastes, but some sort of solids separation
may have to proceed the primary treatment
step. Little data is available on the
treatment of solvent sludges of high solid
content other than by incineration. Tech-
niques exist, however, which should be
applicable to separating some fraction of
solvent constituents from these waste.
REFERENCES
1. 0. C. White, "EPA Program for Treatment
Alternatives for Hazardous Wastes,"
JAPCA 35 (4), 369 (1985).
2. H. Freeman, "A Review of Alternative
Treatment Processes for Dioxin Waste
244
-------
Streams," Incinerationand Treatment of
Hazardous Waste; Proceedings.of the
Twelfth Annual Symposium, UStPA (1986).
3. R. Turner, "A Review of Alternative
Treatment Processes for Halogenated
Organic Waste Streams," Incineration
and Treatment of HazardousWaste:
"Proceedings of the Twelfth Annual
Symposium. USEPA (1986).
4. B. Blaney, "Alternative Techniques for
Managing Solvent Wastes," Journal of
the Air Pollution Control Association
36 (3), 275 (1986).
5. Hazardous Waste Management System;
Land Disposal Restrictions; Proposed
Rule, Federal Register, 51 (9) 1602-
1766 (1/14/86).
6. W. E. Sweet, R. 0. Ross and 6. V.
Verde, "Hazardous Waste Incineration:
A Progress Report," JAPCA, 35 (2), 138
(1985). ~~
7, H. M. Freeman, et al., "Thermal De-
struction of Hazardous Waste: A State
of the Art Review, Journal of Hazard-
ous Materials, (198fJI
8. D. Oberacker and R. Mournighan, Haz-
ardous Waste Engineering Research
Laboratory, USEPA, private communica-
tions.
9. Guidance Manual forHazardous Waste
Incineration Permits, Office of Solid
Waste, USEPA, No. SW-966 (1983).
10. B. Dellinger, et al, "Examination of
Fundamental Incinerability Indices for
Hazardous Waste Destruction," Inciner-
ation and Treatment of Hazardous
TOstes: Proceedings of the Twelfth
Annual Symposium, USEPA (1985).
11, R. H. Penny and C. H. Chilton, Chemi-
cal Engineer's Handbook, 5th Edition,
McGraw-HillBook Co., New York (1973).
12. C. Allen, et al, Field Evaluation of
Hazardous Waste Pretreatment as an Air
Pollution Control Technique, HWERLjj
USEPA, Contract 68-02-3992 (In Public-
ation).
13. C. Allen, "An Evaluation of Batch
Fractional Distillation for Removing
Organics from Mixed Aqueous/Organic
Wastes," Incineration and Treatment of
Hazardous Wastes: Proceedings of the
Twel ft h Annua 1 Syinpos i urn, I)SEPA,
(1986).
14. J. Spivey et al, Pr e 11minary Assess-
ment of Hazardous WastePretreatment
as an Air Pollution Control Device,
HWlRL, USEPA, Contract No. 68-03-3149
(In Publication).
15. R. A. Dobbs and J. M. Cohen, Carbon
Adsorption Isothermsfor Toxic Or-
ganics, EPA-600/8-80-Ogg, MfRL.irSEPA,
Cincinnati, OH, 1980.
16. Physical, Chemical and Biological
Treatment Techniques for Industrial
Wastes. Office of Solid Waste, USEPA,
No. SW-148 (1977).
17. Y. H. Kiang and A. A. Metry, Hazardous
Waste Process1ng Techno!ogy, Ann Arbor
Sclence Publishers, Inc., Ann Arbor,
MI (1982).
18. U.S. Environmental Protection Agency -
Treatability Manual.Volume III, IERL,
USEPA, Cincinnati, OH, EPA-60Q/2-82-
0016 (1983).
245
-------
PRELIMINARY ASSESSMENT OF AERATED WASTE TREATMENT SYSTEMS
AT TSDFs
David A. Green
C. Clark Allen
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
Aerated wastewater treatment unit operations are used for the removal of organic
compounds from hazardous waste and industrial wastewater streams. In some operations,
aeration is required to supply oxygen for aerobic decomposition of organics; in other
operations, incidental air/water contact occurs to varying degrees. Methods for esti-
mating emissions resulting from air stripping of volatile organic compounds that can
accompany aerated treatment are reviewed and applied to full scale and pilot plant
wastewater treatment plants.
The basic problem in evaluating air emissions from aerated waste treatment pro-
cesses is to determine the importance of competing mechanisms. In an activated sludge
biological treatment process, for example, the major mechanisms for removal of dis-
solved contaminants from the aqueous waste are biological oxidation, adsorption on
biomass, and mass transfer into the air. Where one or more of the removal mechanisms
is destructive of the contaminant, e.g., chemical or biological oxidation, and others,
e.g., air stripping or aerosol dispersal are non-destructive, measurements of the rela-
tive importance of the competing mechanisms become complicated.
The presence of competing removal mechanisms, most importantly, biooxidation, re-.
duces the accuracy of emissions predictions based on plant influent concentration data.
If bulk wastewater concentrations are known for indivdual treatment units, then reason-
ably accurate predictions can be made. However, when only the concentration of the
influent to the entire plant is known, accurate rate data for biooxidation becomes more
important and the absence of these data has a greater influence on the accuracy of
emissions estimation.
Estimates of the fate of volatile organic compounds in waste treatment systems were
made using available theoretical and semi-emperical mathematical models. The predic-
tions were compared with experimental measurements made at full-scale and pilot-scale
treatment systems. It was concluded that reliable emissions estimates (i.e., estimates
within the accuracy which is expected to result from variations in sampling and chemi-
cal analysis) can be made using selected mathematical models and correlations. Where
plant operating conditions including residence times, aeration rates, dimensions, etc.,
are available, the limitations in applying the mathematical models result from lack of
accurate biooxidation and partition (gas/liquid and liquid/solid) coefficient data.
246
-------
INTRODUCTION
The U.S. EPA Office of Air Quality
Planning and Standards (OAQPS) is devel-
oping regulations under the 1976 Resource
Conservation and Recovery Act (RCRA) to
control air emissions from hazardous
waste treatment, storage, and disposal
facilities (TSDFs). The purpose of the
air emissions regulations is to protect
human health and the environment from
emissions of volatile organic compounds
(VOCs), and particulates.
The sources of TSDF emissions include
storage tanks, treatment processes, sur-
face impoundments, lagoons, landfills,
land treatment, and drum storage and
handling facilities. Those processes
involving the use of air or oxygen (bio-
logical treatment and cooling), subject
to the introduction of air (stirred
equalization and neutralization), or
exposed to the air (clariflers and open
tanks), may emit VOC as a consequence of
air stripping.
In order to further understand and
better estimate the source and extent of
VOC emissions from TSDFs, model aerated
treatment facilities were investigated.
Various mathematical models can predict
the mechanism and extent of VOC emissions
during the different process conditions
that are encountered. The mathematical
models include a methodology for esti-
mating the relative importance of com-
peting removal pathways (e.g., adsorption
and biological oxidation).
The basic problem in evaluating air
emissions from aerated waste treatment
processes is to determine the importance
of competing mechanisms. In an activated
sludge biological treatment process, for
example, the major mechanisms for removal
of dissolved contaminants from the aque-
ous waste are biological oxidation, ad-
sorption on biomass, and mass transfer
into the air. In a rough analysis of
"removal efficiency" concentrations of
the contaminant in the aqueous influent
and effluent are measured and the frac-
tion of the contaminant which disappears
in the process is reported as an effi-
ciency. This procedure gives no informa-
tion about the relative importance of
competing removal mechanisms.
Where one or more of the removal
mechanisms is destructive of the contami-
nant, e.g., chemical or.biological oxida-
tion, and others, e.g., air stripping or
aerosol dispersal are non-destructive,
measurements of the relative importance
of the competing mechanisms become com-
plicated. In this case the contaminant
removed by the non-destructive mechanism
must be recovered and the balance assumed
to have been destroyed. Errors In chemi-
cal analyses, and particularly, errors
involved in sampling will have a great
impact on calculations of the fraction of
the contaminant which has been destroyed.
If adsorption, aerosol dispersion,
and chemical reactions are unimportant
(expected to be true with VOCs), then any
difference between influent and effluent
contaminant levels may be attributed to
air stripping and biological oxidation.
Thus, if a rate expression describing
mass transfer through air stripping is
available, it can be used to calculate
contaminant levels which are susceptible
to destructive removal mechanisms. Con-
versely, if rate data for the destructive
removal mechanisms (e.g., biological
oxidation kinetics) are available, these
can be used to estimate the relative
importance of air stripping (through mass
balance calculations).
One approach to the problem is to
measure the total disappearance in the
field and then apply mathematical models
obtained from laboratory data to simul-
taneously estimate the rate of stripping
and biological oxidation. Mien these
models account for significantly more or
significantly less contaminant removal
than the observed total removal, several
steps can be taken. The simplest is to
assume that the mass transfer model is
more accurate than the biological oxida-
tion model and account for biological
oxidation (and other destructive removal
mechanisms) by difference. The mass
transfer models are in fact more accurate
predictors than the biological oxidation
models because they are less affected by
changes in operating conditions and in-
fluent wastewater composition. In some
cases, however, air sampling programs
must be conducted to verify the predic-
tion of stripping models. The interpre-
tation of field sampling data is, how-
ever, subject to considerable error.
247
-------
The objectives of this study were to
present and evaluate the various predic-
tive fate Models for VOC removal pathways
that are applicable to aerated wastewater
treatment systems. Several different
aerobic biological waste treatment sys-
tems were considered, including activated
sludge, aerated lagoons, and spray ponds.
The wastewater treatment systems typical-
ly contain a number of associated process
units such as bar screens, grit chambers,
equalization/neutralization tanks and
primary and secondary clarifiers. The
plant configuration is variable and de-
pends on the composition and flow rate of
the waste stream as well as other consid-
erations ,
The approach taken was to investigate
mathematical models and correlations
available in the literature for mass
transfer in systems similar to wastewater
treatment units. Classical engineering
approaches to strippers and absorbers,
packed bed design, and stream and lake
reaeration were considered and adapted as
necessary to wastewater treatment sys-
tems. Experimental data were compared
with the predictions of the mathematical
nodeIs,
PROCEDURE
Pretreatment units such as grit cham-
bers were modeled as plug flow systems
with the gas phase mass tranfer coeffi-
cient calculated from a "wind speed"
correlation (MacKay and Yeun, 1983) and
the liquid phase mass transfer coeffi-
cient calculated from a "stream reaera-
tion" correlation (Owens, et al,, 1964).
Equalization basins are assumed to be
completely mixed and modeled with the
same gas and liquid phase mass transfer
coefficient correlations as the pretreat-
ment units. The same correlations for
liquid and gas mass transfer coefficients
are used for clarifiers (primary and
secondary) but plug flow is assumed. The
flow over clarifier weirs is modeled
separately with an arbitrary reduction in
liquid phase mass transfer coefficient
due to reduced turbulence in the freely
falling stream. It was assumed that
negligible biooxidation takes place in
these units.
Aeration basins were assumed to be
well mixed. Surface aeration systems
were considered to be partially agitated
and partially quiescent. The quiescent
portion of the surface was modeled in the
same way as equalization basins. Mass
transfer from the agitated portion of the
surface was calculated using a liquid
phase mass transfer coefficient from
Thibodeaux (1978) and a gas phase mass
transfer coefficient from Reinhart
(1977). Both correlations are based on
agitator power. Mass transfer from dif-
fused aeration basins was based on air
bubbles reaching equilibrium during the
travel time from diffuser to the surface.
Additional diffusion from the surface was
calculated as for clarifiers. Biological
decay is incorporated in the models in
the same manner regardless of the means
of aeration, with a first order depend-
ence on biomass concentration, and a zero
order dependence on dissolved oxygen
concentration.
As a step in verifying the accuracy
of the suggested mathematical models,
data from an extensive sampling program
at a well characterized (with respect to
residence time, dimensions and influent
and effluent concentrations) full-scale
industrial wastewater treatment system
were used to test the models. In addi-
tion, pilot-scale data from controlled
experiments conducted by EPA (Petrasek,
1981) and some less extensive industrial
wastewater treatment system data from EPA
plant surveys were also used.
A list of input specifications which
describe the system to be modeled is
given in Table 1. Values given are those
describing a wastewater treatment system
at a Union Carbide chemical manufacturing
plant (Alsop, et al., 1984). For cases
where units of a particular type are not
used in a particular system (e.g. dis-
solved air flotation in the example
given) a specification of zero for area,
number, or air flow eliminates the inap-
propriate part of the model. Dimensions
and flow rates are used to calculate
residence times and surface areas. Sur-
face aerator characteristics are used to
calculate mass transfer coefficients.
Henry"s Law coefficients are taken from
an on-line data base for the specific
compounds of interest.
248
-------
RESULTS AND DISCUSSION
The results of the calculations on
the sample plant specification are given
Table 2.
The rate of loss of volatiles from
the Union Carbide Plant as predicted by
the mathematical models agreed well with
measured results, with the exception of
the UNOX biological treatment unit. The
rate of removal in the ONOX system was
overpredicted by the model, particularly
in the case of tetralin (73 percent pre-
dicted vs 5 percent observed). It is
unknown whether analytical procedures or
the assumptions of the model were respon-
sible for the variation. Though the loss
in the aeration basin was overpredicted
by the model for tetralin and naphthalene
(89% was the observed loss in both
cases), the absolute errors were still
only 5-10*. Therefore, it nay be con-
cluded that the models were successful in
simulating the field conditions for the
plant, and that most of the VOCs were
lost to the atmosphere.
In an EPA pilot investigation
(Petrasek, 1981), a number of VOCs were
processed in a clarifier followed by an
aerated biological oxidation unit.
Enough data were provided to specify the
input parameters of the model. For both
the clarifier and the aerator, the model
predicted VOC losses in close agreement
with the pilot data. The model predicted
that most of the VOCs were volatilized in
the aeration basin.
Limited data were available from
several other industrial wastewater
treatment systems. For these cases model
predictions varied considerably from
reported data, but extensive assumptions
were made regarding process configura-
tions and steam characteristics which nay
have caused the discrepancies. In the
two wastewater treatment units for which
source descriptions as well as dependable
field measurements were available, the
agreement of the model with the reported
VOC losses was acceptable for estimation
of emissions (usually ± 20%). In situa-
tions where the source characteristics
were not known, the predictions of the
model were less accurate, however the
accuracy of the reported compositions has
not been established.
CONCLUSIONS AND RECOMMENDATIONS
The available theoretical and semi-
empirical methods for predicting the
rates of mass transfer of volatile organ-
ic compounds from dilute solution into
air can be used to predict emissions from
aerated wastewater treatment processes in
the absence of competing removal mecha-
nisms. Where liquid phase concentration
data and system operating conditions are
well characterized, it is anticipated
that VOC emissions from aerated lagoons,
activated sludge processes, and clarifi-
ers associated with wastewater treatment
systems can be estimated within the accu-
racy of sampling and chemical analysis
results.
The presence of competing removal
mechanisms, most importantly, biooxida-
tion, reduces the accuracy of emissions
predictions based on plant influent con-
centration data. If bulk wastewater
concentrations are known for individual
treatment units, then reasonably accurate
predictions can be made. However, when
only the concentration of the influent to
the entire plant is known, accurate rate
data for biooxidation become more impor-
tant and their absence has a greater
influence on the accuracy of emissions
estimation.
Biooxidation rate data are available
in the literature for various organic
compounds which may be present in indus-
trial wastewater. The data are highly
dependent on the conditions under which
they were obtained and are not easily
adapted to real systems which may have
different initial concentrations, biomass
loadings, nutrient and inhibitor concen-
trations, etc. A table of biooxidation
rate data compiled from the wastewater
treatment literature (Allen, et al.,
1985) can be used as a starting point for
rough estimation purposes when no site
specific data are available. Where time
and resources are limited, the biooxida-
tion rate in the aeration basin is the
most important measurement to be obtained
because these data are the most difficult
to estimate. In an ideal system where
only one biodegradable organic compound
is present, this rate can be determined
from respirometry studies of aeration
249
-------
basin mixed liquor. Often, however, one
or more specific VOCs of concern are
present In a wastewater containing a
larger quantity of biodegradable materi-
al. In such cases, the specific biode-
gradatlon rate must be determined through
a sequence of chemical analyses in a
system where alternate removal pathways
are controlled (laboratory respiroaetry
studies or covered pilot plants).
Industrial waste treatment systems
vary widely In design reflecting the wide
variation In the waste streams to be
treated. The recommended predictive
mathematical models are adaptable to
widely varying volumetric flow rates.
Residence time, which in actual system
designs accounts for the concentration
and difficulty of removal of organic
compounds and suspended solids, must be
specified fro* actual system designs or
estimated (less accurately) from removal
rate data. The best mathematical models,
as determined in this study, have been
compared with experimental measurements
of full-scale and pilot-scale wastewater
treatment systems.
In the absence of site-specific bio-
oxidation rate data, the models cannot be
verified. However, the assumptions of
very low biodegradation rates in the
example cited led to accurate (± 25*)
predictions of total disappearance of
specific VOCs in the aeration basin.
Where plant operating conditions includ-
ing residence times, aeration rates,
dimensions, etc., are available, the
limitations in applying the mathematical
models result from lack of accurate bio-
oxidation and partition coefficient data.
When descriptions of specific plant oper-
ating parameters are not available, it is
anticipated that rough estimates can be
•ade based on Model plant designs, ad-
justing for wastewater flow rates.
Alsop, G. M., R. L. Berglund, T. W.
Siegrist, G. M. Whipple, and B.
E. Wilkes. "Fate of Specific
Organics in an Industrial Biolo-
gical Wastewater Treatment
Plant," Draft Report. Prepared
for the U.S. Environmental Pro-
tection Agency, IERL. June 29,
1984.
MacKay, D. and A. T. K. Yeun. "Mass
Transfer Coefficient Correlations
for Volatilization of Organic
Solutes from Water." Environ.
Sci, Technol. 17(4):211. 1983.
Owens, M., R. w. Edwards, and J. W.
Glbbs. "Some Reaeratlon Studies
in Streams." Inter. J. Air &
Water Poll.. 8:496. 1964.
Petrasek, A. C., Jr. "Removal and
Partitioning of the Volatile
Priority Pollutants in Conven-
tional Wastewater Treatment
Plant." Capsule Report. Pre-
pared for the U.S. Environmental
Protection Agency, HERL. Octo-
ber, 1981.
Relnhart, J. R. "Gas-Side Mass
Transfer Coefficient and Inter-
facial Phenomena of flat Bladed
• Surface Agitators." University
of Arkansas. Ph.D. Thesis.
1977.
Thibodeaux, L. J. "Air Stripping of
Organics from Wastewater: A
Compendium." Proceedings of the
Second National Conference on
Complete Water Use. Chicago, IL.
May 4-8, 1978.
REFERENCES
Allen, C. C., D. A. Green, J. B.
White, and J. B. Coburn. "Pre-
liminary Assessment of Aerated
Waste Treatment Systems at
fSDPs," Draft Report, Prepared
for the U.S. Environmental Pro-
tection Agency, HWERL. November
1985.
250
-------
TABLE 1. MODEL INPUT DATA FOR UNION CARBIDE PLANT
(Alsop, et al., 1984)
(M)
Water Temperature (Deg C)
Wind Velocity (cm/S)
Concentration of Benzene (Mole Fract)
Concentration of Naphthalene (Mole Fract)
Concentration of Ethylbengene (Mole Fract)
Concentration of Methylcellosolve (Mole Fract)
Concentration of Tetralin (Mole Fract)
Biorate of Benzene (Hours)
Bi orate of Naphthalene (Hours)
Biorate of Ethylbenzene (Hours)
Biorate of Methylcellosolve (Hours)
Biorate of Tetralin (Hours)
Waste Flow Rate (M3/Sec)
Area of Pretreatment Basin
Depth of Pretreatment Basin
Diameter of Clarifier (M)
Depth of Clarifier (M)
Number of Clarifiers
Length of Aeration Basin (M)
Width of Aeration Basin (M)
Depth of Aeration Basin (M)
Area of Agitators (Each)
Number of Agitators
Power of Agitation (Each) (HP)
Impeller Diameter (cm)
Impeller Rotation (rpm)
Number of Secondary Clarifiers
Diameter of Secondary Clarifiers (M)
Depth of Secondary Clarifiers (M)
Area of Equalization Basin
Depth of Equalization Basin
Air Flow in DAF (Pretreat) (SM3/S)
Flow of Submerged Air (M3/Sec)
Power of Sub-Agitation, Each (HP)
Length of Sub. Aeration Basin (M)
Width of Sub. Aeration Basin (M)
Depth of Sub. Aeration Basin (M)
Area of Sub. Agitators (Each) (M2)
Number of Sub. Units in Series
Impeller Rotation (rpm)
Impeller Diameter (cm)
Weir Height (cm)
Thickness of Weir Flow (cm)
Enter {1} for Covered Subm. Agit.
25
200
.001
.005
.001
.001
.0002
400
400
400
800
200
50
3
19
2,
1
170
170
3.
96
30
75
30
2000
1
37
3
5185
3
0
50
33.
8,
8,
70,
3
70
243
30
1
1
.07
251
-------
TABLE 2. A COMPARISON OF THE EXPERIMENTAL LOSS OF VOLATILES FROM
THE UNION CARBIDE PLANT FIELD TEST TO THE PREDICTED
LOSS FROM MATHEMATICAL MODELS
Unit/Volatile
Fractional Loss
(experimental)
Fractional Loss to Air
(predicted)
Clarifier Surface
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
Clarifier Overflow
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
Total Clarifier
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
Equalization Basin
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
Aeration Basin
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
UNOX
Benzene
Ethyl Benzene
Toluene
Dichloroethane
Naphthalene
Tetralin
.058
.103
.13
.143
.045
.225
.391
,35
.281
.389
.269
.115
.9984
.993
.993
.987
.892
.905
.26
.44
.19
- . 08
.31
.05
.16
.15
.15
.13
.12
.15
.01
.01
.01
.01
.006
.01
.17
.16
.16
.14
.116
.16
.31
.29
.29
.31
.27
.29
.998*
.997*
.998*
.994*
.994* •
.998*
.58*
.55*
.62*
.23*
.26*
.73*
*Includes biological loss (approximately 0.01)
252
-------
AN EVALUATION OF BATCH FRACTIONAL DISTILLATION FOR REMOVING
ORGANICS FROM MIXED AQUEOUS/ORGANIC WASTES
C. Clark Allen
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
Batch fractional distillation was evaluated as a technique for removing volatile
organics from mixtures of water and solvents. The tests were performed at a hazardous
waste recycling firm. Two waste streams were chosen for study, both with high water
content. One stream contained approximately 95 percent water; 3 percent methyl ethyl
ketone; and 2 percent alcohols, chlorinated hydrocarbons, and other organics. The
second stream contained 77 percent water, 21 percent acetone, and 2 percent other or-
ganics. Fractional distillation of both of these streams resulted in 99 percent re-
moval of the organics from the aqueous bottoms, with resultant total organic concentra-
tions in water of under 1,000 milligrams per liter.
INTRODUCTION
Wastes that contain volatile organic
compounds (VOCs) are a potential source
of air emissions when disposed of in
landfills, in surface impoundments, in
land application, or in other ways where
the waste comes into contact with the
environment.
Wastes with a high organic content
can be recovered (with waste volume re-
duction) by thin film evaporators or they
can be incinerated if the fuel content is
high. Dilute aqueous wastes can be air
stripped or chemically reacted, either
biologically or with oxidants. Direct
steam stripping can remove organics which
are immiscible with water, but the re-
moval of water soluble organics cannot be
cost effectively handled by simple steam
stripping. Fractional batch distillation
can remove water soluble organics from
wastes present in concentrations 3 to 50
weight percent.
Distillation is most often carried
out in large packed or tray columns in
separation processes associated with
chemical manufacturing plants. When
multiple components are to be separated
in continuous distillation systems, mul-
tiple columns are required, since typ-
ically only one group of compounds are
separated per column. In contrast, batch
fractional distillation permits the se-
quential removal of a series of different
compounds from a mixture, using only one
column. The purity of the components
depends on the column design and opera-
tion; a series of different distillation
fractions can be removed from the batch,
with the low boiling point compounds
removed first.
Water is less volatile than many
organic solvents commonly present in
wastes. The water in the vapors is se-
lectively removed in the distillation
column as the vapors passing up the col-
umn contact liquid flowing down the col-
umn. The concentration of the most vola-
tile component increases from the bottom
to the top of the distillation column.
253
-------
TREATMENT PROCESS
TEST BATCHES
The use of distillation to treat
hazardous waste was investigated since
low concentrations of VOCs in water can
be removed and recovered as an organic
stream. In batch distillation, the batch
can be treated until the concentrations
of VOCs are below specifications. Thus,
organic materials which have significant
water solubilities can be recovered with
a distillation column. With steam strip-
ping, on the other hand, a mixture of
water and organics would be obtained from
the process which would require further
processing. Plant B was selected for
field testing because it contains eight
fractional distillation systems offering
a variety of treatment capabilities.
The fractional distillation systems
at Plant B each consist of a reboiler, a
tray column and condenser, an accumula-
tor, and associated pumps, valves, and
piping (Figure 1). The system selected
for any particular separation Is depend-
ent upon a number of factors such as
throughput, relative vapor pressures, and
required purity of the process streams.
The reboiler contains a steam coil for
heating. The steam supply header pres-
sure is controlled at 125 pslg, but there
is some fluctuation in pressure due to
the nature of the controller. Vapor from
the reboiler enters the bottom of the
tray distillation column, and the flow
from the bottom of the column recycles
back to the reboiler by gravity. Reflux
is provided to the top of the column from
the accumulator by a small centrifugal
pump. The overhead condenser is a verti-
cal, shell-and-tube, water-cooled conden-
ser. The distillate accumulator (approx-
imately 50 L) is a small tank from which
the column reflux is pumped, with the
product overflowing to any of a number of
product storage tanks.
The steam flow rate to the reboiler
and the reflux rate to the column are
controlled by an operator through manual
adjustment of hand valves. The column
head and reboiler temperatures and the
distillate rate and appearance are the
primary factors used in control of the
process.
Two different waste streams were
selected for the field evaluation at
Plant B and one batch of each was moni-
tored and sampled during processing for
the evaluation. The waste streams were
both aqueous organic and consisted pri-
marily of methyl ethyl ketone and ace-
tone, respectively. The individual com-
ponents in these waste streams are char-
acterized in Tables 1 and 2. The facili-
ty objectives in processing these wastes
were to reclaim solvent and reduce the
VOC content to a level acceptable for
disposal in a muncipal wastewater treat-
ment facility. However, due to some
light sludge, oil, or heavy organic con-
tamination in the wastes as received, it
was unlikely that the residue from this
batch could be made acceptable for dis-
posal to the sewer. The residue was sent
to a TSDF which handles wastes with low
concentrations of VOCs.
Process data collected during pro-
cessing including reboiler temperature,
column head temperature, reflux rate, and
in one case, product rate. There was no
means to measure the steam flow to the
reboiler during distillation, but
approximate rates could be obtained by
enthalpy calculations. The quantity of
distillate recovered could not be mea-
sured since there were no liquid level
instruments on the product storage tanks.
These tanks were very large, on the order
of 37,850 liters (10,000 gallons) or
greater, and in both cases contained
product from previous batches. The batch
volume, VOC content, and process data for
both Batch 1 and Batch 2 are given in
Table 3.
Batch: 1 Aqueous Methyl Ethyl Ketone
Batch
A 30,000 liter (8,000 gallon! batch
of the methyl ethyl ketone (MEK) waste
stream was .charged to the 42" system
(designated by column diameter) reboiler.
Figure 1 gives a schematic of the actual
distillation system used to process the
aqueous MEK batch. This system has a
42,000 liter (11,000 gallon) reboiler,
and the column in this sytem is a 106.7
cm (42 in.) diameter, 30 tray column.
The 30,000 liter charge was determined by
an operator through visual observation of
the level in the reboiler.
254
-------
Heatup of the batch was Initiated by
opening a hand valve in the steam line to
the reboiler. The column was operated
under total reflux during the heatup and
until the reboiler and column head tem-
peratures remained constant. The system
was held at this condition for about 1
hour before distillation was started to
insure steady-state operation. The re-
boiler and column head temperatures and
the rate of distillate flow were moni-
tored to determine the progress of the
distillation. The distillation was es-
sentially completed when the reboiler
temperature had reached 100 °C and a
rapid rise in the column head tempera-
tures occurred, indicating depletion of
the MEK. However, stripping was con-
tinued until a VOC level of less than 0.1
percent was achieved as evidenced by
laboratory analyses.
The overhead product was put into a
tank in which organic distillate was
being accumulated for further refining
prior to being returned to a specific
client.
Batch 2: Aqueous Acetone Batch
The 32" distillation system (desig-
nated by column diameter) was used for
processing this batch. This system in-
cludes a 13,000 liter (3,500 gallon)
reboiler and an 81 cm (32 in.) diameter,
30 tray column.
An 11,000 liter (3,000 gallon) batch
of the acetone waste stream was charged
to the reboiler. Heatup of the batch was
initiated by opening a hand valve in the
steam line for the reboiler. The column
was operated under total reflux during
the period of heatup and until the re-
boiler and column head temperatures re-
mained constant. The system was held at
this condition for about 3 1/2 hours
prior to the start of distillation solely
for plant process scheduling purposes.
The reboiler and column head temperatures
and the distillate rate were monitored
during processing to determine the pro-
gress of the distillation.
This waste stream contained some
contamination of heavy organics or light
sludge as received. The aqueous residue
would normally have been stripped to a
level acceptable for disposal to the
sewer, but the unexpected contamination
would preclude this option. Therefore
the distillation was completed after the
batch temperature had reached 100 °C and
the column head temperature began to rise
rapidly indicating depletion of the or-
ganic solvent. The final batch sample
was taken at that time.
Lower Level of VOCs Obtained
The results of analyzing the waste
material in the reboiler during stripping
is presented in Figures 2 and 3.
In the methyl ethyl ketone (MEK)
batch, 2,2-dimethyl oxirane (DM0) and MEK
were present in concentrations greater
than 0.5*. MEK and DM0 were removed from
the process with less than 10 ppm
remaining after 615 minutes. The tri-
chloroethane was somewhat more slowly
removed with 530 ppm or 0.05 percent
residual trichlorpethane in the water
material after 720 minutes. The methy-
lene chloride was rapidly removed from
the batch with greater than 99.7 percent
removal (less than 10 ppm remained) after
215 minutes. All other major components
were removed to below the detectability
limits of the analytical procedures.
In the acetone batch, there was re-
moval of all the VOCs to less than 700
ppm in the final treated waste. There
was 99.7 percent removal of acetone (re-
moved to 690 ppm). All other components
were removed to below their detectability
limits.
Removal of VOCs With Time—Process
Kinetics
To simulate the removal of the VOCs
from the batch, a multicomponent distil-
lation computer program was modified for
batch operation. The relative volatili-
ties of the components are used with the
PROCESS LIMITATIONS
The waste in the reboiler should
contain relatively little dissolved sol-
ids: the solids could foul the heating
coils. Reactive wastes and wastes that
255
-------
are highly corrosive to the reboiler
should not be separated In this process
without stabilization,
PROCESS RESIDUALS
Air Emissions
In addition to the removal of VOCs
from the waste material, the absence of
significant treatment emissions is also
of interest. The treatment would be of
little use if substantial quantities of
VOCs are released to the atmosphere dur-
ing processing.
The air emission sources from the two
different distillation processes were
evaluated at Plant B. The vent locations
are indicated in a typical process dia-
gram of the equipment which is presented
In Figure 1. In both columns, the con-
denser was vented to the atmosphere on
the downstream side of the product flow
through a vertical atmospheric vent with
a tee near the top and a vertical pipe
with a discharge several meters lower
than the tee. The upper condenser dis-
charge was beneath metal plates in the
scaffolding support. The upper condenser
vent on the 110 cm (42-inch) column was
accessible for air sampling and velocity
determinations as planned, but the upper
vent on the condenser from the 81-cm (32-
inch) column was inaccessible for the
measurement of velocities due to person-
nel safety consideration. Air sampling
was obtained on the condenser vent and
the receiver vent of the 81-cm (32-inch)
column at approximately mid-cycle. With
a cross-sectional area of 13 cm2 (0.014
ft<*), the average flow during a 10-minute
period was estimated to be less than 1.3
L/sec (0.05 fts/sec).
The product vent of the aqueous MEK
batch was sampled during filling, and
visible fumes were observed with a rela-
tively low velocity. The volumetric flow
rate from the product vent was assumed to
be equal to the volume displaced by the
product. The temperature In the gas vent
was measured with a mercury thermometer
where accessible. The temperature of gas
emitted from the condenser vent on the
110-cm (42-inch) column was typically
slightly warmer than the atmosphere,
indicating that some of the flow from the
vent could have originated with the dis-
tillation process.
The concentrations of the volatiles
in the headspace over the distillate
obtained midway through the process cor-
related within a factor of 33 percent
with the concentrations obtained from the
material in the vapor phase of the MEK
process. The concentrations in the air
emission decreased as the batch was pro-
cessed, and the concentrations of the
volatiles in the laboratory analysis of
headspace over both the distillate and
the reboiler contents tended to decrease
as the batch was processed. The average
emissions are expected to be a combina-
tion of both the rate of flow and the
composition drift of the various com-
ponents. Table 4 estimates the air emis-
sions on the basis of the average of the
results of the test on three air samples
taken from the aqueous MEK batch. The
volumes emitted from the vents were esti-
mated on the basis of the data obtained
from working losses and vent velocity
measurements. The same procedure was
used with the acetone batch and presented
in Table 5. In both cases, the emissions
were only a small fraction of the VOCs
recovered from the process.
Liquid Residuals
The treated waste would have low
enough concentration (less than 0.1 per-
cent) so that the waste could be dis-
charged as waste water to a publicly-
owned treatment facility. Because 3,2-
dimethyloxirane was present in the MEK
waste, and because a heavy organic con-
taminant was present in the acetone
waste, the waste was shipped to a hazard-
ous waste treatment facility which
handled low concentration wastes.
Solid Residuals
There were no solids from waste
treatment, but any suspended solids pre-
sent in the waste would be removed in a
decanter and landfilled.
256
-------
PROCESS COST
CONCLUSIONS
Specific, detailed operating and 1.
capital costs were not discussed pri-
marily because of the relatively large
number of, and differences in, process
systems at this facility. In addition,
the complexity of the process operations
vary over a wide range, depending on the
particular feed stream being processed 2.
and the intended use of the product.
Plant B stated that it would not be
economical to fractionate waste streams
with less than 6 to 8 percent reclaimable
organics. Some order of magnitude costs
provided were:
$0.26 per liter (1.00 per gallon)
operating cost when organic is 3.
stripped as the overhead product.
$0.40 per liter (1.50 per gallon)
operating cost where water is
stripped overhead with the organic
being the bottoms product. 4.
The individual VOC components can be
removed from the waste material by
batch distillation. Removal effi-
ciencies of 99 percent and greater
were observed, with resultant VOC
concentrations below 0.1 percent.
The removal rates of the components
are a function of the waste matrix
and the ratio of the rate of steam to
the batch size and are generally
proportional to the residual concen-
tration in the waste. In the distil-
lation process with reflux to the
column, the more volatile materials
are removed first from the waste.
In the two processes tested, air
emissions from the process vents
represented only a relatively small
fraction (less than 0.2 percent) of
the VOC present.
Costs of treatment were typically in
the range of SO.20 to $0.?0/L of VOC
recovered, but were estimated to be
as high as S1.18/L for streams of low
(<10 percent) VOC concentrations.
-To
Product
Storage
Tank
Figure 1. Schematic of a typical fractional distillation system si Plant B.
257
-------
10.000 -
1,000 -
• MEK
A Methanol
V Methylene chloride
• Carbon tetrachloride
O 2.2-DMO
liopropanol
Computar simulation!
400
Time (minutes)
600
800
Figure 2. Concentrations of VOC in the MEK waste during distillation.
A Acetone
O TrichloroothQna
Itopropanol
• Toluana
O 1.1.VTrichloroeth
• MEK
— Computer simulation!
200
Tima (minutes)
400
Figure 3. Concentrations of VOC in the aqueous acetone waste during distillation.
258
-------
TABLE 1. WASTE CHARACTERIZATION OF BATCH 1
(Aqueous Methyl Ethyl Ketone)
Number of phases 1
Total solids (mg/L) SAO
pH 5.5
Water (weight percent) 95
Oil (weight percent) Negligible
VOC (weight percent) 5
Density (g/cm ) 1
VOC analysis (Aqueous phase)
Compound Concentration
(mg/L)
Methyl ethyl ketone 30,000
2,2-Dimethyl oxirane 6,400
Isopropanol 1,900
Methylene chloride 3,100
Methanol 3,500
Carbon tetrachloride 1,700
1,1,1-Trichloroethane 710
Other VOCs 2,130
TABLE 2. WASTE CHARACTERIZATION OF BATCH 2
(Aqueous Acetone)
Number of phases 1
Total solids (mg/L) 7,500
pH 13
Water (weight percent) 77
Oil (weight percent) Negligible
VOC (weight percent) 23
Density (g/cm ) 1
VOC analysis (Aqueous phase)
Compound Concentration
(mg/L)
Acetone 212,000
Trichloroethene 9,500
1,1,1-Trichloroethane . 2,800
Toluene 2,700
Methyl ethyl ketone 2,300
Isopropanol 440
Aromatics 291
259
-------
TABLE 3. DISTILLATION WASTE CHARACTERIZATION AMD PROCESS DATA
Initial volume (L)
VOC (mass percent)
Water content
Process Data
Distillation time3.
Steam rate (kg/hr)
Stripper residue volume (L)
Bottoms VOC content (weight %)
VOC recovered (kg)
% VOC recovered3
neasured
b
Estimated.
Batch 1
30,000
4.7
95
12.0 hr
860
28,500
0.05
1,400
99
Batch 2
11,400
23
77
8,0 hr
615
8,800
0.07
2,614
99.8
Estimated by material balance.
TABLE 1 . AIR EMISSIONS: AQUEOUS METHYL ETHYL KETONE PROCESS
Source
Condenser vent 1
Condenser vent 2
Condenser vent 3
Condenser vent average
Accumulator vent
Time
10s42A
1:15P
2: SOP
2:05P
VOC
(«g/L)
762
0.26
8.1
7.1
Flow
(L/s)
0.72
0.41
0.8
0.4a
Emissions
(g/8)
0.55
0.00011
0.0064
0.186
0.0028a
rteximum from working losses: not representative of process
260
-------
TABLE 5 . AIR EMISSIONS; AQUEOUS ACETONE PROCESS
Source
Condenser vent 1
Condenser vent 2
Condenser vent average
Receiver vent
Receiver vent
Time
10s31A
2s36P
10:42A
10:42A
VOC Flow
(mg/L) (L/s)
1.1 <1.3
1.3 <1.3
1.5 <0.6f
1.5 1.0b
Emissions
(g/s)
<0.0014
<0.0017
<0.0016
<0.0009a
0.0015
Estimated as less than half of the flow from the condenser vent.
Estimated from average product reflux rate.
261
-------
REVIEW OF ALTERNATIVE TREATMENT PROCESSES FOR
HALOGENATED ORGANIC WASTE STREAMS
Ronald J. Turner
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Many of the solvent and nonsolvent halogenated organic wastes exhibit high toxicity,
mobility, persistence and bioaccumulation. The EPA's Office of Research and Development
Is conducting a research program to develop information on the applicability, effective-
ness, capacity, cost and environmental impact of existing alternative (to land disposal)
halogenated waste treatment technologies. Processes evaluated at commercial offsite
treatment facilities include evaporation, distillation, steam stripping and fuel blending.
INTRODUCTION
Large quantities of organic solvent
and nonsolvent halogenated wastes are
generated and land-disposed annually in the
United States. The United States Environ-
mental Protection Agency has estimated that
60 million gallons of organic solvent
wastes and 1.44 million gallons of nonhalo-
genated organic wastes were landfilled
In 1981. Much larger quantities of these
wastes were land-disposed (400 million
gallons solvent wastes and 3.1 million gal-
lons of halogenated nonsolvent wastes).
The data base used to estimate the volumes
of these wastes is the "National Survey of
Hazardous Waste Generators and Treatment,
Storage, and Disposal Facilities Regulated
under RCRA" (1981).
As many of these wastes exhibit high
toxicity, mobility, persistence and bio-
accumulation, the current land disposal
practices may not sufficiently protect
human health and the environment. The EPA
previously promulgated regulations under
the RCRA which restrict the land disposal
of ignltable, reactive, incompatible, and
liquid wastes, including ignitable sol-
vents, explosive wastes, and reactive cyna-
nides. The EPA is now reviewing existing
controls to determine if further restric-
tions on land disposal of hazardous wastes
are warranted and whether alternative waste
management methods exist which are techni-
cally, environmentally, and economically
practical.
WASTE CHARACTERIZATION
An analysis was made of the hazardous
wastes listed in the following four RCRA
Sections (1):
0 261.24 - Characteristics of EPA
Toxicity (DOXX)
° 261.31 - Non-Specific Process Wastes
(FOXX)
0 261.32 - Specific Process Wastes
(KXXX)
0 261.33 - Discarded Commercial Chemical
Products, Off-Specification
Products, etc. (UXXX and PXXX)
This analysis identifies a total of 142 ha-
zardous halogenated wastes (Table 1).
There are 109 halogenated organic compounds
represented by the 142 waste codes. Two of
the more frequently generated halogenated
wastes are F001 and F002:
F001 - Spent halogenated solvents used in
degreasing: tetrachloroethylene,
trichloroethylene, methylene
chloride, 1,1,1-trichloroethane,
262
-------
TABLE 1. RCRA-LISTED WASTES CONTAINING HALOGENATED ORGANIC COMPOUNDS (HOCs) (2)
Total Total
Number Number
Waste Listed Containing
Category in Part 261 HOCs (%)
Listing of Specific Hazardous Waste Codes
Containing One or Hore HOCs
rsa
e>
tc
DOXX
FOXX
KXXX
PXXX
uxxx
17 6 (35)
13 2 (IB)
76 27 (36)
107
233
21 (20)
86 (37)
D012, 0013, 0014, 0015, D016, D017
F001, F002
K001, K009, K010, K015, K016, K017, K018, K019, K020
K021, K028, K029, K030, K032, K033, K034, K041, K042
K043, K073, K085, K095, K096, K097, K098, K099, K1Q5
P004, P016, P017, P023, P024, P026, P027, P028, P033
P036, P037, P043, P050, P051, P057, P058, P059, P060
P095, P118, P123
U006,
U030,
U042,
U061,
U074,
U083,
U132,
U192,
U226,
U240,
U017,
U033,
U043,
U062,
U075,
U084,
U138,
U207,
U227,
U242,
U020,
U034,
U044,
U066,
U076,
U097,
U142,
U208,
U228,
U243,
U023,
U035,
U045,
U067,
U077,
U121,
U150,
U209,
U230,
U246,
U024,
U036,
U046,
U068,
U078,
U127,
U156,
U210,
U231,
U247
U025,
U037,
U047,
U070,
U079,
U128,
U158,
U211,
U232,
U026,
U038,
U048,
U071,
U080,
U129,
U183,
U212,
U233,
U027,
U039,
U049,
U072,
U081,
U130,
U184,
U222,
U235,
U029
U041
U060
U073
U082
U131
U185
U225
U237
Totals
446
142 (32)
-------
carbon tetrachloride, and chlorin-
ated fluorocarbons; and sludges
from the recovery of these sol-
vents in degreasing operations.
F002 - Spent halogenated solvents: tet-
rachloroethylene, methylene chlo-
ride, trichloroethylene, 1,1,1-
trichloroethane, chlorobenzene,
1,1,2-trichloro- 1,1,2-trifluoro-
ethane, ortho-dichlorobenzene, and
trichlorofluoromethane; and 1,1,2-
trichloroethane; the still bottoms
from the recovery of these sol-
vents, and certain spent mixtures
/blends.
An example of a halogenated, nonsol-
vent waste is K001, a wastewater treatment
sludge from wood preserving operations us-
ing creosote and/or pentachlorophenol.
Large quantities of KOQ1 wastes presently
are stored in impoundments. Cleaning up
these contaminated impoundments is re-
quired. Recovery of the creosote and oily
components of the K001 sludges could parti-
ally offset closure costs. Quantities of
other halogenated process wastes which were
reported in the Part A data base are given
in Table 2. The Part A data base was EPA's
initial effort at registering facilities
that generated and/or managed hazardous
wastes.
TREATMENT ALTERNATIVES TO LAND DISPOSAL
Alternatives such as waste reduction,
recycling, treatment, and incineration are
usually preferable to land disposal. In a
recent survey, waste management practices
for eight halogenated "K" wastes (Table 3)
were obtained from manfacturers of chlorin-
ated aromatics, A total of sixty-five
facilities reported their management tech-
niques. Incineration is indicated to be
the most prevalent form of treatment with
the exception of K099 - untreated waste-
water. Storage in tanks, containers and
piles, recycle/reuse, and wastewater treat-
ment were the other major management op-
tions reported. Alternative treatment
technologies applicable to waste solvents
and other halogenated organic wastes are
discussed in the following sections.
ALTERNATIVE SOLVENTWASTE TREATMENT
TECHNOLOGIES'
The principal alternative treatment
techniques with potential application to
TABLE 2 - WASTE QUANTITY DATA FOR HALOGENATED WASTES
Haste Code
F001
F002
K001
K009
K010
K015
K016
K017
K018
K019
K020
K021
K028
K029
K030
K032
K033
K034
EPA's Part A
Data Base (MT)
1
1
1
,419
392
34
122
,028
144
,061
99
169
226
162
43
57
102
259
55
81
36
,310
,282
,473
,933
,806
,429
,518
,625
,741
,043
,546
,468
,704
,920
,815
,920
,611
,083
Waste Code
K035
K041
K042
K043
K073
K085
K095
K096
K097
K098
KQ99
K105
DO 12
D013
0014
D015
DO 16
D017
EPA's Part A
Data Base (HT)
119,385
68,649
74,518
79,125
25,349
27,383
23,597
9,082
917
14,525
14,979
918
120,988
72,906
77,898
69,546
61,219
79,617
264
-------
both halogenated and nonhalogenated wastes,
depending upon the physical form, are:
Air Stripping
Steam Stripping
Wet Ai r Oxidation
Incineration
Di stillation
Solvent Extraction
Evaporation
Filtration
Drying
Carbon Adsorption
Resin Adsorption
Biological
Several of these technologies are used to
separate and concentrate dilute solvent
mixtures before further processing. The
preferred treatment alternatives for waste
solvents include reduction or elimination
of the wastes generated at the point of
use, recycle/reuse, and thermal oxidation.
Solvent reclamation techniques e.g.
distillation, frequently require prelimi-
nary treatment of incoming waste materials.
Technologies for solvent waste "pretreat-
ment" include simple settling and flota-
tion, screening, and filtration to remove
solids. In some treatment facilities,
thin-film evaporation is employed to sepa-
rate solids from the recoverable organics.
The separated residues may then be blended
into a chlorine substitute suspension for
use in manufacturing low alkali cement, a
fuel for use at kilns and furnaces per-
mitted to burn hazardous wastes, or stabi-
lized for disposal at a secure landfill
site. In addition to distillation, other
general types of solvent reclamation tech-
nologies which are considered to have
application to halogenated solvent waste
streams are evaporation, fractionation, and
steam stripping. A brief discussion of
these technologies follows.
DISTILLATION (NON-FRACTIONATING)
Steam distillation is likely the most
commonly used technique for recovery of
halogenated solvents. Steam distillation
is a process in which separation of mate-
rials is achieved by using their differ-
ences in boiling points. The two general
types of simple steam distillation proc-
esses differ in the method of steam use,
through coils or direct injection. Stills
with steam coils (Figure 1) are widely used
onsite by operators which do not have high
solvent demands. Direct injection steam
stills (Figure 2) are most effectively used
on low boiling solvents that are not mis-
cible in water. In tests conducted for the
HWERL, a commercial solvent reclaimer chs-..
tilled two batches of halogenated and two
batches of nonhalogenated solvents. Over
90 percent of the initial solvents were
recovered for reuse in less critical
application (Table 4).
TABLE 4. FATE OF WASTE SOLVENT
Volume (Percent]
Initial
Recovered as Solvents
Recovered as Fuel
Still Bottoms
Wastewater
1540 gallons
1400 gallons
60 gallons
gallons
60
20 gallons
(91)
(4)
(4)
(1)
About 60 gallons of still bottoms required
solidification due to high chloride con-
tent. This facility usually disposes of
low chloride distillation residues in fuel
blends.
EVAPORATION
The two types of evaporators or sepa-
rators usually employed to recover spent
solvents are the scraped surface evaporator
and thin-film evaporator. A scraped
surface evaporator (Figure 3) is designed
to facilitate density separation of spent
solvent materials. This type of separator
is suited for solvent streams with a high
concentration of suspended solids and
sludges. The thin-film separator (Figure
4) operates on the same principle as the
scraped surface separator except that a
film of liquid material is spread against
the vessel wall where it is exposed to
heat. The thin-film evaporator may be best
suited for reclaiming low boiling point
solvents. The five largest recycling
facilities in California use thin-film
evaporation to reclaim solvents directly
or as preliminary treatment (3).
FRACTIONAL DISTILLATION
Two types of fractionation processes
are commonly employed by industry to re-
claim spent solvents: bubble tray or
packed column. The bubble tray column
(Figure 5) is used when a high purity prod-
uct is desired. Pretreatment is usually
required since solid materials carried over
tend to build up on the trays and reduce
the contact areas between the vapor and
liquid phases, decreasing the overall
efficiency. Settling in a tank followed by
thin-film separation can remove solids to
improve the purity of the liquid feed to
the fractionation column. The packed
265
-------
TABLE 3. CURRENT MANAGEMENT PRACTICES FOR SPECIFIC HALQGENATED PROCESS HASTES
ACCORDING TO EPA's INDUSTRY STUDIES PROGRAMS
Ol
en
Waste
Code
K016
K017
K018
K019
K020
K028
K085
K099
Waste Stream Description SIR
Heavy ends or distillation residues from 2
the production of carbon tetrachloride
Heavy ends (still bottoms) from the 1
purification column in the production of
epichlorohydrin
Heavy ends from the fractionation column in 1
ethyl chloride production
Heavy ends from the distillation of ethyl ene 8
dichloride in ethylene dichloride production
Heavy ends from the distillation of vinyl 3
chlorine in vinyl chloride monomer production
Spent catalyst from the hydrochlorinator
reactor in the production of 1,1,1-trichloro-
ethane
Distillation of fractionation column bottoms
from the production of chlorobenzene
Untreated wastewater from the production of 2,4-D 1
TOTALS 16
Management
IW LF" IN
1 3
2
2
9
5
1 1
2-2
_
2 2 24
Technique!3
RR
1
1
1
5
1
2
_
-
11
SI BB
-
1
1
-
-
1
1
6
10 0
Total
7
5
5
22
9
5
5
7
65
a Source: EPA's Industry Studies Data Base
b STR: Storage in tank, container, or pile
LF: Landfill
RR: Recycle/reuse
BB: Burned in boiler
IW: Injection well
IN: Incinerator
SI: Wastewater treatment in surface impoundment
NOTE: Numbers in column refer to the number of facilities surveyed which reported the given management technique.
-------
column fractionation system (Figure 6)
differs from the bubble-tray system in that
the column is packed with rings to increase
the surface contact area and flow residence
time. This modification can result in a
higher quality reclaimed solvent. Sepa-
ration or removal efficiencies of 99 per-
cent and greater are possible for individ-
ual volatile components via fractional
distillation.
STEAM STRIPPING
Steam stripping by direct injection of
live steam can be used to treat aqueous
wastes (less than 10 percent organics in
water) and wastes containing over 10 per-
cent organics. This'process is used by
some waste recyclers for recovery of vola-
tile organics. However, a disadvantage is
that additional treatment steps may be re-
quired to further reduce volatiles in the
subsequently increased aqueous waste stream
(4).
ALTERNATIVE NONSOLVENT HALOGENATED ORGANIC
WASTE TREATMENT
As with the solvents, other halogena-
ted organic wastes usually require prelimi-
nary treatment, depending on the physical
matrix. Examples of some processes which
could pretreat aqueous or non-aqueous
wastes (5):
Sedimentation/skimming
Filtration
Neutralization
Dissolved air flotation
Heavy metal removal
Fluid extraction
Vacuum filtration
Gri ndi ng
Blending
In general, settling, flotation or
filtration will accomplish solids removal
from liquid wastes. Liquids and solids
destined for incineration may require
grinding and blending to produce a homo-
geneous injectable mixture with specific
moisture, halide and heat contents. At
one treatment facility tested by a HWERL
Contractor, organic materials including
paint, paint sludges, varnish and spent
solvents were ground to reduce the solids
particle size to a few microns. The final
blend was suitable for burning in an incin-
erator or cement kiln (total organic halide
less than 1 percent, heat value over 10,000
Btu/lb).
Other treatment methods for non-
aqueous halogenated wastes include chemical
dechlorination and chlorinolysis.
INCINERATION
For many halogenated organic wastes,
incineration is the desirable disposal
method because essentially complete de-
struction is possible. Incineration can
be accomplished using a variety of methods
depending on the physical characteristics
of the waste. For example, the type of
incineration selected may depend upon the
viscosity of the liquid and the amount and
size of solids. A rotary kiln or fluid-
ized bed may be chosen for a viscous, high
solids waste. Liquid injection would
probably be chosen for a waste with low
viscosity and low solids because of the
lower cost of this process (5).
AQUEOUS WASTE TREATMENT
Generic treatment options for aqueous
halogenated organic wastes include biolog-
ical, chemical, and physical processes.
Generally, only dilute aqueous wastes with
an organic content of less than I percent
are amenable to biological treatment and/or
carbon adsorption (physical process). Wet
air oxidation (chemical process) is usually
applicable to wastes which are too dilute
to incinerate economically and yet too
toxic to treat biologically. Supercritical
water oxidation is a newer chemical process
where liquid phase oxidation destroys
aqueous wastes containing high organic
concentrations (1-20 percent). UV/ozone is
another chemical process in which an aque-
ous waste is subjected to ultraviolet
radiation and ozone. This method is also
usually restricted to a 1 percent or lower
concentration of organics.
CONCLUSION
There are alternative management meth-
ods available for treating many waste
streams containing halogenated organic
materials subject to the land disposal
ban legislation. Commercial hazardous
waste treatment facilities employ practical
and economically viable processes for waste
halogenated solvents and other organic
wastes. Because of the Congressional
267
-------
mandate to protect human health and envi-
ronment, some of these processes now em-
ployed In commercial and onsite treatment
of wastes may need improved equi pment
design and instrumentation together with
effective operating and maintenance proce-
dures 1n order to comply with this pro-
vision of the RCRA.
REFERENCES
1. Hazardous and Solid Waste Amendments
of 1984. Public Law 98-616,
November 8, 1984.
2. D. Roeck et al. Assessment of Wastes
Containing Halogenated Organic Com-
pounds and Current Disposal Practices,
draft report prepared for the OSW,
August 1984 (68-01-6871-2).
3. J. Radimsky et al. Recycling and/or
Treatment Capacity for Hazardous
Wastes Containing Halogenated Organic
Compounds, California Department of
Health Services, November 1984.
4. C. C. Allen, B. L. Blaney, Techniques
for Treating Hazardous Wastes to
Remove Volatile Organic Constituents,
JAPCA, 35:841 (1985).
5. H. Arienti et al. Technical Assess-
ment of Treatment Alternatives for
Wastes Containing Halogenated
Organics, draft report prepared for
the OSW, October 1984 (68-01-6871-9).
268
-------
TO VAPOR
RECOVERY
TO VAPOR RECOVERY
COLD WATER -7 I .-MIST PAD
•M MIST ELIMINATOR/
I J PHASE SEPARATOR
• *- TO RECYCLED
SOLVENT TANK
COLD WATER
SOLDS&
SLUDGE TO TANK
IND.
MAKE-UP WATER
STiLL
TO VAPOR
RECOVERY
I (3> - TO RECYCLED
r SOLVENT TANK
HEAT-. COND.
SETTLINGl JACKET
TANKr"
SOLIDS & SLUDGE
TO TANK
MAKE-UP WATER
FIGURE 1. TYPICAL STEAM-COIL STILL SYSTEM
FIGURE 4. TYPICAL THIN-FILM SEPARATOR
SYSTEM
TO VAPOR
RECOVERY
TO VAPOR RECOVERY
COLD WATER-7 t /-MIST PAD
MIST ELIMINATOR/
HLETHEAT I /' ,/HASE SEPARATOR
EXCHANGER r~\ COND. X—TO RECYCLED
SOLVENT TANK
I STEAM
SPARGE
MAKE-UP WATER
SOUDS & SLUDGE
TO TANK RECYCLED WATER
TO TREATMENT
MIST ELIMINATOR/ PHASE SEPARATOR,
TO RECYCLED
SOLVENT TANK
COND.
SOUOS & SLUDGE TO TANK
FIGURE 2. TYPICAL STEAM-STRIPPER STILL
SYSTEM
FIGURE 5. TYPICAL BUBBLE-TRAY COLUMN
FRACTIONATION SYSTEM
TO VAPOR
RECOVERY
HEAT-
SETTUNGL JACKET
TANKf"
MIST ELIMINATOR/PHASE SEPARATOR |
-TO RECYCLED
SOLVENT TANK
TO VAPOR
RECOVERY
[REFLUX 3-PHASE
(SEPARATION
TO RECYCLED
SOLVENT TANK
SOLIDS & SLUDGE
TO TANK
MAKE-UP WATER
COND.
SOLIDS & SLUDGE TO TANK
FIGURE 3. TYPICAL SCRAPED-SURFACE
SEPARATOR SYSTEM
FIGURE 6. TYPICAL PACKED-COLUMN FRACTION-
ATION SYSTEM
269
-------
MANAGEMENT OF HAZARDOUS WASTES CONTAINING
HALOGENATED ORGANIGS
Douglas R. Roeck and Peter H. Anderson
GCA Technology Division, Inc.
Bedford, MA 01730
ABSTRACT
The 1984 RCRA amendments direct EPA to study available treatment
technologies for waste streams containing halogenated organic compounds. If
it is determined that existing technology and capacity is sufficient for
management of these wastes, then effective July 8, 1987, wastes containing
halogenated organic compounds will be prohibited from land disposal. This
paper presents estimates for volumes of halogenated wastes generated and
managed in the U.S. and a brief overview of treatment technologies that have
been demonstrated to be effective in handling these streams. Also presented
are descriptions of and preliminary sampling results from two treatment
facilities evaluated by GCA during our continuing performance evaluation
program for EPA's Hazardous Waste Engineering Research Laboratory (HWERL).
INTRODUCTION
In the 1984 RCRA amendments,
halogenated wastes are defined as any
hazardous waste containing halogenated
organic compounds (HOCs) in total
concentration greater than or equal to
1,000 mg/kg.1 Unless ongoing studies
show that there is insufficient
technology and/or capacity for managing
these wastes, they will be prohibited
from land disposal on July 8, 1987.
Similar land disposal restrictions are
already in effect in California and
New York State.^ In California,
hazardous wastes containing HOCs in total
concentration greater than or equal to
1,000 mg/kg were restricted from land
disposal effective January 22, 1983. In
Mew York State, hazardous waste
containing more than 5 percent by weight
(50,000 mg/kg) of halogenated chemicals.
were prohibited from land disposal after
March 31, 1985.
WASTE GENERATION
National estimates of quantities of
halogenated organic wastes generated and
land disposed are given in Table 1. As
ohown, relatively small quantities of
solvents and nonsolvents are land
disposed, compared to total generation.
Most of the land disposed quantity for
'solvents is handled by deepwell
injection, while for nonsolvents, about
half is landfilled and half is deepwell
injected.
WASTE TREATMENT ALTERNATIVES
EPA has proposed regulations for
implementation of the land disposal
prohibitions, including treatment
standards and effective dates for waste
solvents (FR, Vol.51 ,No. 9, January 14,
1986, pp 1602-1766). The Agency
concluded that biological degradation,
steam stripping, air stripping, carbon
270
-------
TABLE 1. ESTIMATES OF VOLUMES (106
gal/yr) GENERATED AND LAND
DISPOSED FOR HALOGENATED WASTES
Category
Solvents
Nonsolvents
Total Quantity ft,200 24.2
Generated
Total Quantity 400 3.1
Land Disposed
Source: References 3 and 4.
adsorption, distillation, incineration,
and use as a fuel substitute are
demonstrated technologies for treatment
of (halogenated) solvent wastes and that
resin adsorption, chemical oxidation, wet
air oxidation, chemical reduction,
encapsulation, and chemical fixation/
solidification have potential
applicability but are not yet fully
demonstrated. EPA has defined best
demonstrated achievable technology (BOAT)
as steam stripping, carbon adsorption,
biological treatment, or some combination
for waste (halogenated) solvents amenable
to separation/removal techniques and
either incineration or use as a fuel
substitute for wastes not amenable to
separation/removal methods. Final
treatment processes with potential
application to halogenated organic wastes
are similar to those available for
solvents and include nonaqueous and
aqueous technologies:
1. Nonaqueous waste treatment—
* Liquid injection
ine iner at ion
• Rotary kiln incineration
• Fluidized-bed incineration
• Molten salt incineration
• Plasma arc incineration
(pyrolysis)
* High temperature fluid
wall reactor
• Lime/cement kiln
coincineration
• Chemical dechlorination
* Chlorinolysis
2. Aqueous waste treatment and
wastewater treatment—
Wet air oxidation
Activated carbon adsorption
Steam stripping
Biological treatment
UV/ozonation
Supercritical water
oxidation
Table 2 presents an overview of each
of these technologies as applied to
various physical matrices. Detailed
discussions of each process are available
in the literature.*
FIELD STUDIES
GCA has conducted preliminary
investigations of commercial hazardous
waste treatment, storage, and disposal
facilities (TSDFs) as part of a
continuing program for HWERL. The
purpose of this work is to gather
background and performance data on
various treatment processes for use by
HWERL (and ultimately EPA's Office of
Solid Waste) in the development of
regulations pertaining to waste-specific
land disposal restrictions.
To date, fifteen TSDFs have been
visited under this program and two
facilities handling halogenated organic
streams have been evaluated through field
measurement activities. A brief
discussion of these sampling programs
follows.
Plant A
This facility handles both
halogenated and nonhalogenated solvent
wastes designated as F001-F005. Still
bottoms containing volatile halogenated
organic compounds are processed in a
nonagitated thin film evaporator (TFE) to
remove residual solvent. In addition,
the company treats other spent
halogenated solvents using a combination
of direct steam injection (roughing
still) and indirect, convection heated
(polishing still) distillation.
271
-------
TABLE 2. TREATABILITY MATRIX FOR HALOGENATEB NONSOLVENT WASTES
Treatment alternative
Aqueous Inorganic Organic Organic
streams streams liquids solids
Established Technologies
Rotary kiln incineration
Fluidized-bed incineration
Liquid injection incineration
Lime/cement kiln coincineration
Carbon adsorption
Steam stripping
Biological treatment
X
X
X
Emerging Technologies
Molten salt incineration
Plasma arc incineration (pyrolysis)
High temp, fluid wall destruction
Supercritical water oxidation X
UV/ozone treatment X
Wet air oxidation X
Chemical dechlorination
Chlorinolysis
X
X
X
X
X
X
X
X
X.
X
X
X
X
X
272
-------
Sampling was conducted on the TFE on
three different days: one while
processing a waste oil containing
1,1,1-trichloroethane (1,1,1-TCE) and
trichloroethylene (TCE), and the other
two during processing of a
perchloroethylene (PCE) — oil mixture at
two different feed rates. Operating
conditions are noted in Table 3 and
preliminary results are provided in
Tables 4-7.
The TFE operated at atmospheric
pressure, temperatures of 250-300°F, and
a feed rate of 1.0-2.2 gpm. Runs 1 and 3
involved processing of a waste stream
much higher in solvent than normally run
through the unit. Lab analyses performed
at Plant A prior to waste treatment
showed a PCE content of only 3 percent,
whereas the generator had claimed that
the PCE content was much greater than
20 percent. Thus, Plant A did not do any
preliminary treatment on this stream in
one of their batch stills. Subsequent
analyses (GC/FID and GO/MS) at the CCA
laboratory showed much higher PCE levels
of around 35 percent. The reduction in
solvent from waste feed to separated oil
for Runs 1 and 3 was, therefore, much
less than expected (20-30 percent) and
this stream required additional treatment.
Run 2, which involved a waste stream of
only 2 percent total solvent, provides a
better indication of the TFE's
capabilities as solvent reduction from
the feed to the bottom stream ranged from
90 percent for PCE to 98 percent for
1,1,1-TCE. Run 2 also showed a
significant increase in heat content (due
to both solvent and water removal) from
the feed to the oil stream as noted in
Table 5. Metal analyses as shown in
Tables 6 and 7 indicate significant
increases in concentration from the feed
to the separated oil stream. The highest
levels were recorded for copper, zinc,
lead, and chromium. This may warrant
further attention to these types of
facilities since these bottom streams are
typically used offsite as supplemental
fuels in boilers, blast furnaces, and
cement kilns.
One day of sampling was devoted to
the steam distillation unit, during which
time a waste solvent consisting primarily
of 1,1,1-TCE was processed. The feed
rate for this unit was variable, but
during the 264 minute test period, 10
drums were treated for an average feed
rate of 2.1 gpm. The operating
temperature ranged from 112 to 212 °F.
Samples collected included the roughing
still bottoms, the water separated in a
coalescer, and solvent product obtained
downstream from a second coalescer after
the polishing still.
The waste feed to this process was a
two-phase liquid (upper phase about 30
percent of the volume and lower phase
about 70 percent) wherein the upper phase
was found to contain approximately 5
percent 1,1,1-TCE and the lower phase
about 100 percent 1,1,1-TCE. Other
solvents found to be present in the waste
feed included methylene chloride, methyl
ethyl ketone, dimethyl ketone, TCE, and
PCE. Following water separation, the
product stream from the polishing still
was 89 percent 1,1,1-TCE. The roughing
still bottoms, which were also two-phase,
showed 2.8 percent and 4.4 percent
1,1,1-TCE in the upper and lower phases,
respectively.
Air monitoring was also conducted at
this site using a Century Systems Organic
Vapor Analyzer (OVA). These measurements
were taken to provide gross estimates of
fugitive solvent emissions at the site.
Measured values ranged from about 10 ppm
(as methane) at a background location to
about 700 ppm (methane) at an induction
fan located on the roof above the
processing area. Several high readings
recorded (500-700 ppm) were attributed to
minor spills that occurred during the
test period.
Plant B
GCA also sampled a LUWA agitated TFE
at a second TSDF during processing of a
halogenated waste stream containing mostly
TCE (F001). (Other nonhalogenated streams
273
-------
TABLE 3. TFE OPERATING CONDITIONS, PLANT A
Run No.
Waste
description
RCRA Haste Code
No. of drums
processed
Test duration, miti
Feed rate, gal/min
Avg. temp., °F
Pressure
1
PCE/oil mixture
from decreasing
F001
6
212*
2.2
300
atmospheric
2
1,1,1-TCE
and TCE mixture
from degreasing
F001
6
237
1.4
250
atmospheric
3
Same as
run No. 1
F001
6
318
1.0
300
atmospheric
*The unit was shut down several times during the test period.
Run No.
TABLE 4. ORGANIC ANALYSES, PLANT A (TFE)
Waste Feed
PCE
1,1,1-TCE
TCE
Ethyl benzene
Separator Oil
PCE
1,1,1-TCE
TCE
Ethyl benzene
Toluene
Solvent Product
PCE
Methylene chloride
37%
3,100 ppm
6,500 ppm
257.
87%
6,500 ppm
1,050 ppm
2,400 ppm
1.73%
210 ppm
108 ppm
57 ppm
860 ppm
26 ppm
3 ppm
NA
NA
33%
26%
841
274
-------
TABLE 5. TOTAL CHLORIDES AND HEAT CONTENT, PLANT A (TFE)
Waste Separated Solvent
feed oil product
chloride concentration, 10^ y g/g
Run no. (heat content, Btu/lb)
239 170 437
(13,633) (15,605)
(10,287) (18,520)
236 160 548
(13,580) (15,230)
275
-------
TABLE 6. METAL ANALYSES, PLANT A (TFE RUNS 1 AND 3)
Metal concentrations, Wg/g
Run 1 (Run 3)
Metal waste feed separated oil
Cr
Cu
Pb
Zn
1.1
(0.36)
5.9
(3.6)
4.8
(3.3)
9.4
(8.8)
1.0
(0.56)
8.2
(4.1)
8.6
(1.5)
13.0
(12.5)
Note: All solvent products streams <0.06 ug/g
TABLE 7. METAL ANALYSES, PLANT A (TFE RUN 2)
Metal concentrations (ug/g)
Metal
Ag
Cd
Cr
Cu
Ni
Pb
Zn
waste feed
7-1 ;
9.1
15.6
570
12.2
46.3
320 \_
separated oil
17.7
18.3
38.7
1,358
27.3
57,3
734
Note: All solvent product streams <0.6 Mg/g,
except: Cu (5.1) and Zn (6.5).
276
-------
were also sampled at Plant B.) The
halogensted stream treated was a
combination of degreasing wastes from two
paper mills, a transformer manufacturer,
and a cannery. Sampling lasted four
hours, during which time 342 gallons were
processed resulting in an average
throughput of 86 gph. This feed rate was
roughly one-third of normal (300 gph) due
to problems experienced with the still
bottoms pump. The percent recovery
achieved for this waste stream was 70
percent. The average hot oil (used to
heat the TFE) inlet temperature during the
run was 311 °F and the average vacuum was
5.6 in Hg. Preliminary analytical
results from this facility are not yet
available.
CONCLUSIONS
Both theoretical and field
investigations conducted to date indicate
that there appear to be an adequate
variety of demonstrated treatment
technologies available for management of
halogenated organic waste streams. The
relatively small quantities of these
wastes that are believed to be land
disposed may mean that treatment capacity
is also available, although this has not
been determined as yet.
Evaluation of two waste treatment
facilities handling halogenated organic
streams revealed two particularly
important observations. First, strict
attention was given to the characteristics
of still bottom streams so as to avoid
subsequent land disposal. Residual
solvents are intentionally left in these
bottom streams so that they remain
pumpable and have sufficient heat content
to suit prospective customers. This may
be especially important since certain
metals were shown to be enriched in the
bottom stream (separated oil) at one
facility where the ultimate (offsite) use
of the bottom stream is as a supplemental
fuel in a cement kiln.
REFERENCES
Hazardous and Solid Waste Amendments
of 1984. Public Law 98-616
[H.R. 2867]; November 8, 1984.
D. Roeck et al. Assessment of
Wastes Containing Halogenated
Organic Compounds and Current
Disposal Practices. Prepared by
GCA/Technology Division for
U.S. EPA, Office of Solid Waste,
Land Disposal Branch, October 1984.
Engineering-Science. Supplemental
Report on the Technical Assessment
of Treatment Alternatives for Waste
Solvents. Prepared for U.S. EPA,
September 1984.
M. Arienti et al. Technical
Assessment of Treatment Alternatives
for Wastes Containing Halogenated
Organics. Prepared by
6CA/Technology Division for
U.S. EPA, Office of Solid Waste,
Waste Treatment Branch, October 1984.
277
-------
WASTE MINIMIZATION CASE STUDIES FOR SOLVENTS AND METALS WASTE STREAMS
Thomas J, Nunno, Mark Arienti
GCA/Technology Division
Bedford, MA 01730
ABSTRACT
In response to the Hazardous and Solid Waste Ammendments of 1984 (HSWA), the EPA
Office of Research and Development (ORD) Hazardous Waste Engineering Research Laboratory
(HWERL) initiated a program to develop case studies demonstrating waste minimization and
recycling as options for improving hazardous waste management practices in the United
States. The case study program is focused on solvents and metals waste streams from the
metal plating and printed circuit board industries. Specific waste streams being
studied in this program include: 1) waste solvents from resist stripping and developing
operations; and 2) plating baths and waste sludges from plating operations. The paper
presents the progress of the facilities testing program in the early stages of this
program. Technologies being tested range from small batch distillation units to
continuous rinsewater treatment systems. Facilities studied are in the medium to large
size range. The case study assessments being conducted under this program include a
discussion of the process operations, waste streams, mass throughputs, economics of the
technology, and process residuals.
BACKGROUND
With the enactment of the Hazardous
and Solid Waste Ammendments (HSWA) in
November 1984, Congress set forth a
schedule for evaluating the restriction of
various classes of hazardous wastes
including; 1) 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 has also asked EPA to
evaluate the potential for onsite waste
minimization to reduce the quantity or
toxicity of hazardous 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 a Joint research
effort 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 1 iterative 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. The following paper
presents the program approach, summaries
of technologies, and preliminary findings
of HWERL's case study program.
APPROACH
The case study development work was
divided into two phases with Phase I
involving: (1) waste category assessments;
(2) the identification of the data
requirements and organization of the case
studies; and (3) the selection of specific
sites/streams for use in the case
studies. The second phase of work was
278
-------
devoted to testing of the waste
minimization processes and development of
the case study reports.
The waste category assessments were a
series of five reports aimed at
identifying key industries that generate
wastes which are being considered for
restriction from land disposal. The five
waste categories included: 1) solvent
wastes; 2) metals-containing wastes; 3)
cyanide and reactive wastes; 4)
halogenated organic non-solvent wastes;
and 5) corrosive wastes. The results of
the findings of these reports were used in
conjunction with the findings of other
aspects of the case study selection
approach to help direct the final
selections.
As part of the case study
identification/selection process the
project team contacted trade associations
and state agency representatives to
solicit ideas and advice. As a result of
these meetings it was determined that case
study selection should focus on a single
industry or waste stream. The electronics
industry was initially judged as a good
choice because it is a growth-oriented
industry and ranks in the top 20
industries generating solvent wastes.
A primary solvent generation practice
in the electronics component manufacturing
industry is photoresist stripping and
developing operations employed in negative
(organic solvent based) photoresist
printed circuit manufacturing processes.
While initial reactions by facility
representatives was encouraging, it became
apparent that the scope of the case
studies should be expanded to include
other waste minimization steps being
undertaken within the industry. It was
determined that metal plating bath waste
reduction techniques could be assessed
under this program as well. Thus, the
program has focused upon waste
minimization technologies for the
following two waste streams: 1) waste
resist stripping and developing solvents
from the electronic components
manufacturing industry, usually classified
as RCRA hazardous waste codes F001 and
F002 (specific spent halogenated
solvents); 2) metal plating bath wastes
from electroplating and printed circuit
board facilities usually classified as
RCRA hazardous waste code F006.
CASE STUDY SELECTION
During the case study selection
process over 50 facilities were contacted
by mailings or telephone to explain the
case study program and determine their
interest and anticipated level of
cooperation. Based on the initial
screening, 15 metals waste case studies
and 12 resist strip solvent case studies
were identified. Ten (10) facilities were
visited for pretest site visits to assess
the facility's suitability for testing and
further explain the intent and scope of
the case study program. In the final
selection, six facilities were determined
to be suitable to the scope of the program
and willing to cooperate. Table 1 briefly
summarizes salient characteristics of the
waste minimization technologies tested
under this program.
METAL PLATING BATH WASTE MINIMIZATION CASE
STUDIES
Metals plating wastes generated from
plating bath dumps, rinses, etching
machines and scrubbing operations generate
Copper, Nickle, 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
ethant baths and Copper and Tin/Lead
rinsewaters.
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
type of measurements data collected; and
the results obtained or anticipated.
Facility A
Plant A is basically an off-site TSD
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 metals concentration in the
incoming waste is approximately 12 g/L
(12,000 ppm). These waste streams are
279
-------
TABLE 1. SUMMARY OF FACILITIES TESTED UNDER WASTE MINIMIZATION CASE STUDY PROGRAM
facility KIM
facmty A
fictUty B
Facility C
Description
Treatment Storage Disposal
facility handling electroplating
bithi* waste etchar.tt, spills etc,
dpi city; IGOOspN (JA.OCOgpd)
Contract PCS Bamjficturlng Shop
E»p1oy*e$: 77
Production: SOO.OOQ sf/yr.
Sales! $7KH/yr.
towputtr Kamif*etur«r <
employees; 10,000
Technology
Sodium Hydroxide Precipitation
Sodlirra Borohydride Reduction
Alkaline Chlorlnatfon
GOHM Equlpnant dht, by ET1CAH-RI
Sodlua BorohydHdc Reduction
Hemtek Ultrafntratton System
i Solvent Batch Distillation
*
*
*
*
*
*
•
«
Hastes Treated/Reduced
Sfckle Plating Baths j
Copper Plating Baths ! -Sludge
Cyanide )
tuprtc Chloride Exchant j
Electroless Plating Rinses | Sludge
Electroplating Rinses )
Product
Product
ficlltty D
Electronic Equipment Manufacturer
PC losrd Manufacturing using the
*ubtr«ctlve technique In the
KaeBerpld process.
Evplo^ees: 2SQ
Computer Manufacturer
PC Board Manufacturing using
iddltive techniques
tuployccs: 600
Production: 600,000 sf/yr,
PC Board Manufacturer
Z silled Single liycr circuit boards
Production: 480,000 sf/yr.
' 2 Stage Solvent Distillation
(I) Bypont RISTOH SRSt 120
solvent recovery still
(2) Zerpa Industries
> Activated Carbon Regeneration of
spent plating baths
Agmet Equipment Corp.
electrolytic recovery units
1,1,1 Tricnloroethsne resist developer
1,1,1 Trlchloroethane still bottom
Acid Copper Plating Bath
Add Copper Plating R1nscw«ters
Iln/Lead Plating Rlnscwaters
classified into the following four
categories: 1) acidic metals solutions;
2) alkaline metals etchant solutions; 3)
cyanides; and 4) chelated metals
solutions. The case study for this
facility focuses on the metals and
cyanides wastes.
The unit processes employed 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 (SBH) reduction (with sodium
metabisulfite stabilization),
sedimentation, plate and frame filter
press, rapid sand filtration, and ion
exchange columns for effluent polishing.
Initially the facility was designed to
operate using lime and ferrous sulfate
precipitation of metals as the primary
means of waste treatment. When the high
cost of land disposal of the lime sludges
was considered an alternate means of
treating and disposing of the wastes was
selected.
Plant A determined that sodium
borohydride (SBH) reduction was effective
in reducing the elemental metals and
creating a very high metals content sludge
which could be marketed to precious metals
smelters in Europe.1 While sodium
borohydride treatment creates an ideal
sludge by-product for smelting it is not
completely effective in removing all the
metals found in Plant A's widely varied
influent. For example nickle is not
reduced effectively by SBH reduction.
Consequently, Plant A employs a sodium
hydroxide precipitation pretreatment step
to remove nickle as well as a cyanide
oxidation step to destroy cyanides freeing
up the complexed metals for removal by
borohydride reduction.
FacilityB
Plant B is a captive printed circuit
board manufacturing facility employing 77
people in Santa Ana California. Gross
sales are approximately $7 million
annually on production of 500,000 sf of
board. Production at Plant B uses a
special hybrid process, employing elements
of both additive and serai-additive printed
circuit production techniques. Process
wastes of interest to this study include
rinsewaters from the electroplating and
etchant baths. The principle components
of the acid copper electroplating baths
are copper sulfate and sulfuric acid.
Plant B uses a slower acting etchant
280
-------
(sodium chloride, sodium chlorate and
muriactic acid) which etches copper from
the board producing cupric chloride.
Plant B uses a rather 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
sulfurfc acid. Sodium borohydide is added
to obtain an oxidation reduction potential
(ORP) of approximately -250 or less. The
reacted waste then feeds from the
concentration tank to a Memtek
ultrafiltration unit from which the
permeate is discharged to the sewer and
the concentrate is returned to the
concentration tank. A small plate and
frame sludge filter press is used to
dewater the sludge which is drawn from the
bottom of the concentration tank.
The use of the sodium borohydride and
ultrafiltration treatment at Plant B is
favored by the use of the chloride etch
process in 1ieu of the more commonly
preferred ammonium peroxide etch. The
ammonium based etchants create borohydride
sludges stability problems which require
tighter treatment process control and the
use of stabilizers- such as sodium
metabisulfite. However, SBH treatment has
reportedly been used effectively with
ammonia based systems.2
Facility E
Facility E began operations in
January 1982 as a manufacturer of
customized, fine-line multi-layer printed
circuit boards. Facility E initiated an
ambitious waste minimization program in
mid-1984, Since that time, production has
roughly doubled but liquid discharge to
the wastewater treatment plant has
remained constant and wastewater sludge
generation has dropped roughly 30
percent. Waste minimization efforts
continue to center around in-process
modifications to use non-hazardous or
reclaimable solutions, to reduce water
consumption and bath dump frequency, and
to optimize wastewater treatment
operations. A brief description of Plant
E's pattern plating production and plating
bath reclamation are provided below.
At Facility E, boards are pattern
plated with eight acid copper and one
aqueous lead/tin plating baths in a 48
tank plating line. The line begins with a
nitric acid HN03 rack strip tank. After
the racks are stripped, boards are loaded
and then undergo rinsing, cleaning with
phosphate solutions (H3P04, Electroclean
PC2000), and more rinsing before being
plated. Acid copper baths contain CuS04,
organic brighters, and chlorides with
copper concentrations of 24 ounces per
gallon. The general processing procedure
is to activate the board surface (HC1),
plate, clean/rinse and re-plate. Copper
and solder plating baths are treated with
activated carbon once every three months
and every month, respectively. The
frequency of cleaning is determined by
organic contaminant build-up.
Electroplating baths never have to be
dumped with this arrangement under normal
processing conditions.
Activated carbon treatment is
performed in a batch mode for acid copper,
solder and nickel microplating baths in
three separate.systems, the systems
consist of a holding tank, mixing tank and
MEFIAG paper-assisted filter. For acid
copper treatment, 2400 gallons of
contaminated solution is pumped into a
3000 gallon mixing tank. Hydrogen
peroxide is added to oxidize volatile
organic species and the temperature of the
bath is maintained at 120-130°F for one
hour. Powdered activated carbon (80
pounds) is added and the contents are
mixed for 3 to 4 hours. The solution is
recirculated through a paper lined MEFIAG
filter several times to remove activated
carbon. The filter solids and paper are
removed as needed when a predetermined
pressure drop across the filter is
reached, when the bulk of the activated
carbon has been removed (generally after 3
passes of the solution through the
filter), the filter is loosely packed with
5 gallons of diatomaceous earth. The
solution is again recirculated through the
filter until a particulate test indicates
sufficient solids removal (no residue
detected on visual examination of
laboratory filter paper). Total spent
solids from plating bath purification is
1-1/2 drums every three months which is
disposed in a sanitary landfill.
Facility F
Plant F, employing approximatley 300
281
-------
people, is a job shop producing an average
of 40,000 sq. ft. of double-sided, single
layer printed circuit boards per month.
One of the primary waste streams
associated with the manufacturing
operations at this facility is plating
process rinsewater contaminated with
dissolved metals. At many facilities the
rinsewaters from electroplating, etching
and acid or alkaline eleanning baths are
combined and treated in an end-of-pipe
system which removes dissolved metals by
precipitation with lime. When this type
of treatment is used, a hazardous metal
hydroxide sludge (F006) is normally
produced. In lieu of installing a
precipitation treatment system, Facility F
installed an electrolytic recovery system
to remove metals from the rinsewater at
the source. As a result, the composite
waste stream which is collected in the
wastewater sump is released to the sewer
without treatment to remove dissolved
metals.
Four electrolytic reactors are used
to recover copper from the dragout bath
foilowing copper electroplating, and three
are used to recover tin and lead from the
dragout bath following tin/lead plating.
These units are operated 24 hours per day,
seven days per week except when shut down
for periodic maintenance. Maintenance
primarily involves scraping or peeling off
the metal which has been plated onto the
stainless steel cathode. This is a simple
procedure and requires no more than an
half hour of time for each unit.
Prior to installing the electrolytic
recovery units, this facility utilized a
standard two stage rinse system. With
this system, rinsewater containing an
average of 3000 ppm of dissolved metals
was released to the wastewater sump. With
the new system, the concentration of
metals in this rinsewater stream has been
lowered to an average of 25 ppm.
Approximately 20 pounds of copper, and 10
pounds of tin/lead are recovered each
week. This has been accomplished
utilizing electrolytic plate-out units at
a cost of 3500 dollars per unit. The
units which are used for the tin/lead
recovery cost $1000 extra because they
require a columbium anode as opposed to
the standard titanium anode. This
requirement is due to the extreme
corrosivity of fluorboric acid in the
tin/lead plating bath.
RESIST DEVELOPING SOLVENT RECOVERY CASE
STUDIES
Two case studies under this program
focused upon the minimization of developer
solvent wastes and sludges which might
require land disposal or alternatively
incineration. In general, the recovery of
resist stripping and developer solvents is
not unique within the PC board
manufacturing industry. However, the
recovery systems evalvated at the two
facilities discussed below represent
state-of-the-art technology applications.
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 performs to the
further recovery of the solvent bottoms
product of the initial recovery unit.
Facility C
Facility C manufactures computing
equipment including logic, memory and
semiconducter devices, multilayer
ceramics, circuit packaging, intermediate
processors and printers. One of the major
hazardous waste streams that is generated
is spent halogenated organic solvents
(RCRA code F002). The solvents and their
uses are; 1) Methylene Chloride used in
resist stripping of Electronic Panels; 2)
Methyl Chloroform (1,1,1 trichloroelhane)
used in resist developing of Electronic
Panels and Substrate Chips; 3} Freon used
in surface cleaning and developing of
substrate chips; and 4) Perch!oroethylene
used in surface cleaning of electronic
panels.
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 box distillation units to
recover methylene chloride and
perch!oroethylene, flash evaporators to
recover methyl chloroform, and a
distillation column to recover freon. The
two processes being evaluated under this
program include the flash evaporters and
282
-------
the distillation column.
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
100-110°F. The units are operated one to
two shifts per 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 gallons per
hour. MCF condenses on the packing and
falls back into the reboiler. The
distillation bottoms are removed when the
concentration of methyl chlorform reaches
80 percent (approximately 1-2 weeks).
Facility D
Facility D manufactures mobile
communications equipment components in
their Florence, S.C. facility. The
operation consists of a small metal
forming shop, prepaint and painting lines,
electroplating, printed circuit board
manufacutre, and a 3000 6PD on-site
wastewater treatment plant. Printed
circuit boards are produced using the
subtract!ve technique and solvent based
photoresists. Methylene chloride resist
stripper and 1,1,1-trichloroethane (TCE)
developer are continuously recycled in
closed-loop stills. Until recently all
still bottoms were drummed and shipped
off-site for reclamation at a solvent
recycling faciltiy. Facility D purchased
a Zerpa Recyclene RX-35 solvent recovery
system in October, 1985 to recover TCE
still bottoms on-site.
The sampling plan focussed on
evaluating the effectiveness of the
two-stage distillation system in
recoverying TCE solvent. The system
permits full recovery of solvent
(excluding process losses) and reduces
hazardous waste handling to disposal of a
dry solid photoresist cake. The
Recyclene's stainless steel boiler is
immersion heated in silicone oil bath and
lined with disposable nylon and teflon
bags. This unique design permits complete
separation of volatile and non-volatile
compounds with easy removal of the
distillation bottoms.
PRELIMINARY FINDIN6S
Preliminary findings of the waste
minimization case studies being tested
under this program are given in Table 2.
The data given in Table 2 are based
primarily on preliminary data collected by
the facilities and have not been verified
by laboratory results in many cases.
The results summarized in Table 2 are
intended to show that all of the case
studies provide reductions in hazardous
wastes. In addition, the preliminary
economic analyses show that these waste
reductions were accomplished with a net
cost savings to the facility, usually due
to reductions in disposal costs that would
have been incurred under other
alternatives.
While the waste reduction results are
reasonably valid for all of the case
studies the reader is cautioned against
siting the economics data presented. The
economics data cited in Table 2 are based
on assumptions favoring the technology and
in addition are very plant specific. Note
that the economics of SBH reduction are
particularly plant specific and that the
cost savings shown have not been fully
verified at this writing.
For a more complete discussion of
these case studies the reader is asked to
contact either the authors of this paper
or the EPA Project Officer Mr. Harry
Freeman. The completed case study reports
should be available in May 1986.
REFERENCES
1. Rosenbaum, Wayne, ETICAM-RI Inc.
Warwick, RI. Personal communication
with Thomas J. Nunno, GCA/Technology
Division, 12 December 1985.
2. Hackman, Mathew, Morton Thiokol
Ventron Division Personal Communication
with Thomas J. Nunno, GCA/Technology
Division, 20 March 1986.
283
-------
TABLE 2. SIW4ARY OF PRELIMINARY FINDINGS OF WASTE MINIMIZATION CASE STUDIES
Facility Technology
Facility A Sodium Borohydride
Reduction
Facility 8 Sodimi Borohydride
Reduction
Facility C Solvent Batch
Distillation
ro
00
Facility 0 2-St«ge Solvent
Distillation
Facility E Activated Carbon
Regeneration of
Spent Plating Baths
Facility F Agnet Electrolytic
Recovery Unit
Haste Reduced
Hetals Sludge
Metals Sludge
Methyl chloroform
Freon
1,1,1 Trichloroethane
Resist Developer
Still Bottoms
Plating Bath
Treatment Sludge
Betals Sludge
(Cu, Sb, n}
Haste Production
Under Conventional
Mode {tons/yr}
34,250*
700a
19,850
1,293
42.9
2,085a
160°
Projected Haste
Production Under Conventional
Tested Process Disposal Costs
(tons/yr) (Credit)
500 $6.85 Krtb
22 $17,500-285,000
1,140 ($957,000)
138 ($54,600)
4.3 $2,880
H.2 $417,000
1.4 $32,000
Additional Projected Annual
Hew QSM Costs Annual Cost
Disposal Costs (Credit) Savings
0 $2.1 HH $4.75 MM
$5,000-12.500 $8,250-20,750 $31,700-250,000
0 ($7.0 HH) $6.04 MH
0 ($1.4 HH)
$1,050 ($31,270) $32,900
$440 ($240,000) $656,560
$300 $1,500 $33,200
Estimated or projected
-------
A REVIEW OF TREATMENT ALTERNATIVES FOR DIOXIN WASTES
Harry M. Freeman and Robert A. Olexsey
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
INTRODUCTION
The 1984 Amendments to the Resource
Conservation and Recovery Act (RCRA),
called the Hazardous and Solid Waste Act
(HSWA) amendments, direct the EPA to deter-
mine if dioxin bearing wastes should be
prohibited from landfills. This determina-
tion is to be made by November 1986.
Barring a determination by the Agency that
the land disposal of dioxin wastes does not
represent a threat to the environment, or
that appropriate alternative treatment
technology is not available, a land dis-
posal prohibition would go into effect at
that time (1). The purpose of this paper
is to discuss the state of the art for
several treatment processes for dioxin
wastes that have been proposed to be used
instead of land disposal for disposing of
solid and liquid wastes containing dioxins.
CURRENT EPA REGULATIONS FOR DIOXIN WASTES
Under existing EPA regulations, dioxin
bearing wastes may be stored in tanks,
placed in surface impoundments and waste
piles, and placed in land treatment units
and landfills. However, in addition to
meeting RCRA requirements for these storage
and disposal processes (9), the operators
of these processes must operate in accord
with a management plan for those wastes
that is approved by the EPA Regional Ad-
ministrator. Factors to be considered by
the Regional Administrator include:
1. Volume, physical, and chemical
characteristics of the wastes,
including their potential to
migrate through soil or to
volatilize or escape into the
atmosphere.
2. The alternative properties of
underlying and surrounding soils
or other materials.
3. The mobilizing properties of other
materials co-disposed with these
wastes.
4. The effectiveness of additional
treatment, design, or monitoring
techniques.
The Regional Adminstrator may deter-
mine that additional design, operating, and
monitoring requirements are necessary for
facilities managing dioxin wastes in order
to reduce the possibility of migration of
these wastes to ground water, surface
water, or air so as to protect human health
and the environment (6).
For incineration of dioxin bearing
wastes, in addition to the RCRA operation
and performance requirements (10), the in-
cinerator must achieve a destruction and
removal efficiency (ORE) of 99.9999% for
each principal organic hazardous constitu-
ent (POHC) designated in its permit. This
performance must be demonstrated on POHCs
that are more difficult to incinerate than
tetra, penta-, and hexachlorodibenzo-p~
dioxins and dibenzofurans (6). ORE is
determined from the following equation:
= (U-in - Wn.it.)
Win
x 100%
where:
W-jn = mass feed rate of one
POHC in the waste stream
feeding the incinerator
Wout = ma55 emission rate of the
same POHC present in ex-
haust emissions prior to
release to the atmosphere.
285
-------
Of course, the HSWA amendments contain
a strong presumption against land disposal.
Therefore, one would expect that the
Agency's implementation of the HSWA provi-
sions on land restrictions will place
primary emphasis on incineration for dis-
posal of dioxin bearing wastes. EPA has
also stated that operators of thermal
treatment devices must demonstrate that
they can meet the performance requirements
for incinerators in order to treat dioxin
bearing wastes.
ALTERNATIVE TREATMENT TECHNOLOGIES FOR
DIOXIN WASTES
As indicated in Table 1, there are
several thermal and non-thermal processes
that appear feasible for treating dioxin
wastes. While some of these processes have
already been tested using dioxin contami-
nated waste streams, others are either
under development, or have been tested with
other chlorinated waste streams. The
processes are presented in order of the
availability of data for treating actual
dioxin wastes. Those processes which have
actually treated dioxin wastes are pre-
sented first. Those that appear to have
potential but have not actually been used
to treat dioxin wastes are presented in
subsequent sections. Discussions of each
of the processes follow. The names of the
EPA contacts in Table 1 are individuals
who are familiar with the processes.
Reference or failure to reference a
particular process does not imply that the
process is recommended or not recommended
by the U.S. Environmental Protection Agency
for treating dioxin contaminated wastes.
The discussion that follows is a compila-
tion of processes that, from a technical
perspective, the authors judge to have
potential to treat such wastes. To the ex-
tent that test data are available they are
presented here. Each specific application
of each technology is subject to EPA and
state and local regulatory and permitting
requirements on a case-by-case basis.
PROCESSES WHICH HAVE BEEN USED TO TREAT
DIOXIN WASTES
EPA Mobile Incinerator
The EPA's Office of Research and De-
velopment has developed a mobile rotary
kiln incinerator capable of incinerating
hazardous waste at the point of generation.
The unit, called the Mobile Incineration
System, consists of three major subsystems:
(1) combustion and air pollution control
(APC) equipment mounted on three trailers;
(2) continuous flue gas monitoring equip-
ment located in a fourth trailer; and (3)
support equipment.
The incineration system consists of
a primary combustion chamber (rotary kiln)
where solid or liquid wastes are intro-
duced, volatilized, and partly oxidized and
a clean fuel oil-fired secondary combustion
chamber (SCC) for completing the combustion
process initiated in the kiln. The combus-
tion flue gases pass through a gas cleaning
unit designed to collect and remove parti-
culate matter and acid gases prior to
discharge through the stack. The initial
treatment of the combustion gases is accom-
plished by water quenching in the quench
elbow, which is then folowed by particulate
removal in a cleanable high efficiency air
filter (CHEAP). (The primary function of
the elbow is to lower the gas temperature
from 120QOC (220QOF) to adiabatic satura-
tion, at approximately 85°C (185°F), and
thus reduce the gas volume; some parti cu-
late and acid gas removal does occur in the
quench elbow.) The final step in the gas
treatment process is the removal of acid
gases in a cross-flow packed-bed mass-
transfer scrubber. The pH of each process
water stream is controlled by the addition
of sodium carbonate to enhance acid gas
removal and to minimize corrosion of the
gas cleaning equipment. The prime gas
mover for the incineration system is a
dieselpowered induced-draft fan that
maintains a negative pressure in the en-
tire system so as to prevent the leakage
of toxic vapors. The process gas is dis-
charged through a 12 m (40 ft) stack (12).
In 1983, trial burns were conducted
in Edison, New Jersey on RCRA-listed
surrogates, including dichlorobenzene,
trichlorobenzene, tetrachlorobenzene,
Aroclor 1260, and tetrachloromethane
(CC14). After a solids feed system was
installed and tested in December 1984, ad-
ditional laboratory tests were conducted.
Currently, the mobile incinerator is in-
stalled at the Denny Farm site near
McDowell, Missouri, where "cold" and "hot"
tests were conducted using clean soil and
soil contaminated with surrogates similar
to those employed in the earlier liquid
waste tests (CC14 and hexachloroethane).
Tests using dioxin contaminated liquid
286
-------
TABLE 1
TREATMENT PROCESSES FOR DIOXIN WASTES
Treatment Process
EPA Mobile Incineration System
Huber Advanced Electric Reactor
Shirco Infrared System
EPA Contact
Frank Freestone^
212-321-6632
Harry Freeman^
513-569-7529
Harry Freeman
Liquid Injection Incinerators
Photolysis
Rotary Kiln Incineration
GA Technologies Circulating Bed Combustor
Chemical Dechlorination
UV Photolysis/APEG Chemical Detoxification
Supercritical Fluids
Kite Rot Funps
Don Oberacker^
513-569-7431
Charles J. Roger s^
513-569-7757
Don Oberacker
Harry Freeman
Qarry
513-569-7756
Charles Rogers
Robert Olexsey2
513-569-7717
P. R. Sferra 2
513-569-7618
1. Address: Hazardous Waste Engineering Research Laboratory
Edison, New Jersey 08837
2. Address: Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
wastes and soil verified the ORE and the
effectiveness of the control devices.
Interim delisting guidelines were esta-
blished and analyses were conducted on
ash, treated soils, filter materials, and
process/quench water to ascertain if the
guidelines were attainable (13).
The dioxin trial burns were success-
ful, with DREs exceeding 99.9999% for the
POHCs burned. During these tests 3.84
pounds of 2,3,7,8-TCDD, contained in
1,750 gallons of liquids and over 40 tons
of soil, were destroyed. Particulate
emission permit limitations (<180
7% 02) were achieved in three of four
test runs. During the fourth run, parti-
culate emissions exceeded the prescribed
limit slightly, possibly due to the ac-
cumulation of submicron-sized particles
in the air pollution control system. The
observed CO emission values (1.3-7.7 ppm)
are equivalent to those from the best
available incineration technologies and are
indicative of very complete combustion
(C.E.s = 99.993-99.999% (13).
Finally, the treated soil (ash) and
process wastewater from the trial burn
were analyzed for a series of specific
constituents considered as likely con-
taminants. The results of these analyses
287
-------
were used to support an application to
"delist" residues from a planned larger
scale burn of similar wastes (14).
Advanced Electric Reactor
The Advanced Electric Reactor (AER) is
a process designed specifically for on-site
detoxification of soils. It is owned by
the J. M. Huber Company, Borger, Texas.
The reactor employs a new technology
to rapidly heat materials to temperatures
between 4000 - 5000°F using intense thermal
radiation in the near infrared region. The
reactants, which can be in gaseous, liquid,
or solid form, are isolated from the reac-
tor core walls by means of a gaseous blan-
ket formed by flowing nitrogen radially
inward through porous graphite core walls.
Carbon electrodes are heated and in turn
heat the reactor core to incandescence so
that heat transfer is accomplished by
thermal radiative coupling from the core to
the feed materials. Destruction is accom-
plished by pyrolysis rather than oxidation
(15).
For solid waste treatment, the solid
feed steam is introduced at the top of the
reactor by means of a metered screw feeder
connecting the air-tight feed hopper to the
reactor. Nitrogen is introduced primarily
at two points in the reactor annulus formed
by the external containing vessel and the
porous graphite core which creates the
fluid barriers. The solid feed passes
through the reactor where pyrolysis occurs
at approximately 4000°F. After leaving the
reactor, the product gas and waste solids
pass through two post-reactor treatment
zones (PRTZ). The first PRTZ is an insu-
lated vessel to provide additional high
tenperature (in excess of 1100°C, 2000°F)
and residence time (5 to 10 seconds) while
the second PRTZ is watercooled. It also
provides additional residence time (approx-
imately 10 seconds) but primarily cools the
gas to less than 100Q°F prior to downstream
particulate cleanup (16, 17).
Solids exiting the PRTZ are collected
in a solids bin which is sealed to the
atmosphere. Additional solids 1n the prod-
uct gas are removed by a cyclone and routed
back to the solids bin. The product gas
then enters a bag house for fine particle
removal followed by an aqueous caustic
scrubber for chlorine removal. Any resid-
ual organic and chlorine are removed by
passing the product gas through activated
carbon beds just upstream of the emission
stack. The organic, particulate, and
chlorine-free product gas conposed almost
entirely of nitrogen is then emitted to the
atmosphere through the process stack.
Among the potential advantages of this
process are transportability, extremely
high treatment efficiencies (because of the
high process temperatures and long resid-
ence times), intrinsic safty features (such
as the activated carbon beds, electrically
driven solids feeder, and large amount of
thermal inertia in the reactor), and the
ability to detoxify wastes in a pyrolytic
atmosphere, thereby minimizing particulate
emissions.
Huber has reported several relevant
tests that have been carried out over the
past two years. A test in the Spring of
1984 to determine the destruction effici-
ency and destruction and removal effici-
ency (ORE) for the process with carbon
tetrachloride over a wide range of operat-
ing conditions demonstrated greater than
six 9's (99.9999%) ORE. (15) A September
1983 test on PCB contaminated solids demon-
strated better than seven 9's ORE and re-
sulted in Huber's facility at Borger, Texas
being certified to destroy PCB contaminated
solids (17).
In November of 1984 a mobile version
of the Huber unit was used to treat soil
containing 2,3,7,8-TCDD in Times Beach Mis-
souri. Reports of this test state that no
2,3,7,8-TCDD was detected in the treatment
soil at detection limits of .11 ppb from
the 2,3,7,8-TCDD and that no 2,3,7,8-TCDD
was detected in the stack gas at detection
limits of .55 ng/M3 (18).
The process's most recent test was at
a U.S. Navy facility in Mississippi in
which some 1,000 pounds of soil contami-
nated with Agent Orange were pyrolyzed.
Results for these tests are not yet avail-
able.
Shirco Infrared System
The Shirco unit treats waste by meter-
ing the waste streams onto a woven metal
alloy conveyer belt and passing the waste
under infrared heating elements and then
burning combustible off gases in a second-
dary chamber. The unit currently being
tested by Shirco is a portable pilot-scale
288
-------
system that is contained in a 45 foot
trailer (19).
The primary chamber is rectangular
(2.5 ft. x 9 ft. x 7 ft.) and is lined with
multiple layers of ceramic fiber blanket
insulation. Infrared energy is provided
by heating elements equally spaced along
the length of the furnace. The elec-
trically powered heating element can pro-
vide capabilities of 50Q°-185Q°F of process
temperature with material residence times
variable between ten and one hundred and
eighty minutes. Oxidizing, reducing, or
neutral atmospheres can be provided (19).
A secondary chamber is equipped with a pro-
pane burner and is sized to provide gas
residence time of up to 2.2 seconds. The
process temperature capability is 2300°F.
Scrubbers are provided for off-gas cleanup
(19, 20).
Shirco has reported successful use of
the system to treat municipal and industri-
al sludges and a simulated creosote pit
waste. During the week of July 8, 1985 the
unit was used to treat dioxin contaminated
soil at Times Beach, Missouri. A summary.
of the soil decontamination test results is
shown in Tables 3 and 4 (19).
Liquid Injection Incinerators
A liquid injection system consists of
a refractory-lined combustion chamber and a
series of atomizing nozzles. It 1s capable
of burning virtually any combustible waste
which can be pumped. Wastes to be burned
are usually blended in mixing tanks prior
to atomization to improve either their
pumpability or combustibility, and then are
atomized and burned in suspension. The
capacity will vary depending upon the
energy value of the 0.5 to 2 seconds and
13000 to 3000QF respectively. In addition
to being the primary part of a waste com-
bustion system, a liquid injection inciner-
ator is often used as an afterburner to
complete the combustion of waste gases
following burning in other incinerators
such as rotary kilns.
The advantages of liquid injection
units include their capability to inciner-
ate a wide range of wastes and their rela-
tively low maintenance costs due to the few
moving parts in the system. The primary
disadvantage is that they can burn only
pumpable liquids, and are susceptible to
being shut down because of clogged nozzles.
Liquid injection systems are also usually
designed to burn specific waste streams and
consequently are not often used for multi-
purpose facilities.
These facilities have been used to
destroy a variety of wastes including
phenols, PCBs, still and reactor bottoms,
solvents, polymer wastes, herbicides,
and pesticides. They are not recommended
for burning heavy metals, high-moisture
content wastes, or materials with high,
inorganic content.
Liquid injection is a technology in
daily use throughout the country both at
industrial locations and at central treat-
ment facilities. It is the most commonly
used incinerator for hazardous waste de-
struction (21), comprising some 64% of the
market for incinerators (22).
The incinerators aboard the M/T Vul-
canus, an incinerator ship designed to burn
chemical wastes on the open ocean, are
liquid injection units. The units are
vertically mounted with a volume of 131
cubic meters. Each unit has three rotary
TABLE 3
OPERATING CONDITIONS FOR DECONTAMINATION OF DIOXIN CONTAMINATED SOIL
Shirco System
Test No.
Solid Phase
Residence Time
(minutes)
Solid Fuel -__-__ Temperature ------
Rate Zone A Zone B Secondary Chamber
(ib/hr) (OF) (OF) (OF)
30
15
47.68
48.12
1560
1490
1550
1490
2250
2235
289
-------
TABLE 4
DIOXIN CONTAMINATED SOIL DECONTAMINATION TEST RESULTS
Shirco System
Test Number 1 2
Emissions Sampling Duration (hr) 7 2.5
Particulate* at 7% 02 (gr/dscf) 0.0010 0.0002
Gas Phase ORE of 2,3,7,8 TCDD >99.999996 >99.999989
Detection Limit (picograms) 14 8.4
Ash Analysis for 2,3,7,8 TCDD ND ND
Detection Limit (ppb) 0.038 0.033
* Particulate Filter Only - Without Train Rinse
> Indicates ORE Calculations at Detection Limit
ND Indicates Non-Detectable
cup vortex burners firing into a cylin-
drical refractory lined combustion chamber.
Each incinerator is capable of burning
some 1650 gallons/hr of liquid wastes (23).
In July - September 1977 the liquid
injection incinerators on the Vulcanus
were used to destroy Herbicide Orange, a
herbicide that contains several chlorinated
substances including TCDD's. The reported
destruction efficiency (DE) for the TCDD
was greater than 99.93%, which was the
highest DE that could be measured given the
detection limits of the TCDD concentration
in the exhaust gas (23). More recent test
burns in similar liquid injection units on
the Vulcanus II, a sister ship, demon-
strated greater than 99.998 DE for the
organochlorine CCl/j, CHC13, and 1,1,2-TCE
(23). The EPA concluded for its evaluation
of the 1977 and earlier test burns that
oceanic incineration had the potential to
be an environmentally acceptable alterna-
tive to other means of disposal (23).
Although incinerator ships are cur-
rently available for use, they are neither
TSCA or RCRA permitted. However, specific
permits will be required by the EPA before
any operation in this country can occur.
Photolysis
Photolysis is a process that involves
energy. Photolysis studies have shown that
the breakdown of a chemical by light chlo-
rinated dioxins may be photolytically de-
graded in the environment by sunlight (2).
Polychlorinated dibenzo-p-dioxins
undergo photolysis rapidly in the presence
of hydrogen donors such as alcohols,
ethers, hydrocarbons, and natural oils and
waxes. The photolytic degradation appears
to occur in a stepwise loss of one chlorine
atom at a time. Sunlight is effective in
activating the reaction, and hence of all
the possible natural detoxification mecha-
nisms, photolysis appears to be the most
significant. The ultraviolet wavelengths
of light have been shown to be the most
effective in causing the degradation reac-
tion (22).
Under appropriate conditions, certain
wave lengths of light convert the more
toxic dioxins to less toxic forms by remov-
ing halogen substituents. Applications of
this principle will, therefore, be limited
to decontamination and partial degradation,
rather than to complete disposal.
290
-------
The effectiveness of dloxin photolysis
Is enhanced by solution of the dioxins in a
light-transmitting film, the presence of a
hydrogen donor in the solution, and the use
of ultraviolet light (25).
IT Enviroscience, Knoxville, TN, has
developed a solvent extraction photolysis
process that was applied to dioxin-laden
sludge in Verona, MO, in 1980. in this
process the dioxin was extracted from a
sludge using hexane as solvent. The re-
sulting extract was then passed under eight
10-kw lamps to effect the desired reduc-
tion.
A report of the work states that "160
gal batches of the waste were run through
the destruction process in five 2-week
operating periods. Six extractions of the
waste reduced the dioxin content from 34
ppm to 0.2 ppm (34 mg/1 to 0.2 mg/1).
Generally, photolysis reduced the dioxin
concentration to less than 100 PPB (0.1
mg/1) with optimum photolysis times of
about 20 hours. Air sampling at the
single process vent confirmed that no
dioxin (less than .01 ng/m^) was
emitted to the atmosphere during operation"
(26).
PROCESSES WHICH HAVE POTENTIAL TO TREAT
DIOXIN WASTES
Rotary Kiln Incinerators
The rotary kiln system is the most
nearly universal of waste disposal systems.
It can be used for a wide variety of solid
and sludge waste disposal, and for the in-
cineration of liquid and gaseous wastes
(27).
Rotary kiln incinerators are refrac-
tory-lined cylinders mounted with their
axes inclined at a slight angle from the
horizontal. This type of incinerator has
a length-to-diameter ratio between 2 and
10; a rotational speed in the range of 1
to 5 fpm (measured at kiln periphery); an
incline ratio between 1/16 and 1/4 in./ft;
an operating temperature upper limit of
30000F (1650°C), (although typical temper
atures are lower) and a residence time
that can vary from seconds to hours. It
can be "used to dispose of solids, sludges,
liquids, and gases. The speed of rotation
may be used to control the residence time
and mixing with combustor air.
The primary function of the kiln is to
convert, through partial burning and vola-
tilization, solid wastes to gases and
ash/residue. The ash is removed and, if
found to be free of unacceptable levels of
hazardous wastes, is put in a landfill. An
afterburner using gaseous or liquid fuels
to generate a hightemperature oxidizing
environment is almost always required to
complete the gas-phase combustion reac-
tions. The afterburner is connected direc-
tly to the discharge end of the kiln, where
the gases exiting the kiln turn from a
nearly horizontal flow path to a vertical
flow path upwards to the afterburner cham-
ber. The afterburner itself may be hori-
zontally or vertically aligned (28).
The system is designed so that a nega-
tive pressure can be maintained in the kiln
in order to minimize emissions at the end
seals of the rotating section.
Both the afterburner and kiln are usu-
ally equipped with an auxiliary fuel firing
system to bring the units up to the desired
operating temperatures. The auxiliary fuel
system may consist of separate burners for
auxiliary fuel, dual-liquid burners de-
signed for combined waste/fuel firing, or
single-liquid burners equipped with a
premix system. Fuel flow is gradually
turned down, and liquid waste flow is
initiated after the desired operating tem-
perature is attained (28).
In rotary kilns, liquid wastes may be
fired either at the feed or discharge end
of the unit; both cocurrent and countercur-
rent firing designs are used.
Rotary kilns have the advantages of
being able to incinerate a wide variety
of liquid and solid wastes, being able to
accept drum and bulk containers, and being
able to retain waste materials sufficiently
long to accomplish very high destruction
rates. Rotary kilns are capital intensive
and need significant maintenance to main-
tain seals and refractory.
There are no reported data for the
combustion of TCDD in stationary rotary
kilns.' However, there are several rotary
kilns that have been permitted to burn PCB
wastes. These kilns have at least the
potential for incinerating dioxin waste
streams. Three commercial rotary kilns
with PCB permits are:
291
-------
1. ENSCO, El Dorado, AR
2. Rollins, Deer Park, TX
3. SCA, Chicago, IL
Each of these facilities as part of
the trial burns associated with their
permit application reported DRE's for RGB's
in excess of 99.9999*.
Circulating Bed Combustion
Circulating bed combustion technology
is a modification of traditional fluidized
bed incineration developed by GA Technol-
ogies, San Diego, CA. Fluidized bed incin-
erators are vessels, usually some 20 to 25
feet In diameter and thirty feet high, that
contain a bed of inert granular material,
usually sand. The bed is kept at a temper-
ature of 840°F to 1560°F. Fluidizing air
1s passed through a distributing plate
below the bed to agitate and heat the
granular material. Waste materials and
auxiliary fuel are injected radially in
small amounts and mixed with the bed mate-
rial which transfer heat to the waste. The
waste in turn combusts and returns energy
to the bed.
Circulating bed technology extends
fluidized bed technology by increased
turbulence and throughput per unit area.
The incinerator uses high air velocity
and circulating solids to create a highly
turbulent combustion zone. Because of the
high air velocity, solids are entrained and
combustion takes place along the entire
height of the combustion section. Solids
are separated from off-gases by an integral
cyclone and returned to the combustion zone
through a nonmeehanical seal. Temperatures
are uniform with +_ 5Q°F variance throughout
this loop. This uniform temperature and
high solids turbulence avoids the ash
slagging encountered in other types of
incinerators. Turbulence is a prime factor
1n the zone. Nitrogen oxide (NOX) and
carbon monoxide (CO) emissions are reported
to be well controlled by the good mixing,
relatively low temperatues (1,450-1,500°F),
and staged combustion achieved by injecting
secondary air at higher locations in the
combustor (29).
The unit is designed to control pollu-
tion by the efficient combustion/ destruc-
tion of principal organic hazardous con-
stituents (POHCs); efficient retention of
halogens, phosphates, and sulfur; and col-
lection of particulate matter. Circulating
bed combustion burns solids, liquids, and
sludges in the presence of dry limestone to
control acid gases without wet scrubbers.
Since most particles circulate until
combustion is complete, ash volume is mini-
mized. Effluent from the circulating bed
incinerator consists of dry ash. Upset
response in the circulating bed is to slump
the bed, thus retaining pollutants in the
stagnant bed of limestone and ash. Experi-
ence in full-scale commercial plants shows
that the slunped bed is easily and quickly
brought back to turbulent circulation (29).
Because of the fines recycle system,
circulating bed incinerators use fewer
mechanical conponents in the system. The
developer claims that this results in a
more efficient, less costly unit than the
conventional fluidized bed combustor.
The high combustion efficiency and
good heat transfer in the circulating bed
incinerator allows heat to be recovered
from low-grade fuels and wastes. The
combustion chamber itself is of "water-
wall" construction, so that cooling tubes
need not be located in the direct path of
solids.
A trial burn of soils contaminated
with PCB was carried out in the GA Techno-
logies pilot scale 16" diameter Circulating
Bed Combustor on May 7, 1985. The goal of
the tests was to demonstrate that the unit
could decontaminate soil contaminated with
10,000 ppm PCB. The material used for the
tests was soil "spiked" with PCB.
Results for these tests are presented
in Table 5. Dioxins and furans are pos-
sible products of inconplete combustion.
In the tests parts per billion quantities
of chlorinated dioxins were found in the
feed soils only. No indication of the
compounds was found in any by-product of
the combustion process (30).
CHEMICAL DECHLORINATION
Dechlorination processes use chemical
reagents to remove chlorine from chlorin-
ated molecules, to break apart chlorinated
molecules, or to change the molecular
structure of the molecules. Metallic sodi-
um is typically the reagent used to strip
the chlorine away from constituents to form
sodium chloride. Other proprietary re-
agents are often used in conjunction with
metallic sodium to further treat the
292
-------
TABLE 5
PCS TRIAL BURN RESULTS
GA Technologies Circulating Combustor
Parameter Test 1 Test 2
411.5
1,646
12,000
99.95
1.18
1,805
Soil feed rate, Ib/hr
Total soil fed, Ib
PCB concentration in feed, ppm
Combustion efficiency, %
Residence time, sec
Destruction temperature, °F
ORE, *
Particulate concentration
Dry, gr/dscf
Wet, gr/acf
dechlorinated constituents. The majority
of the dechlorination research has been
aimed at the detoxification of polychlori-
nated biphenyls (PCBs). This research is
applicable to many chlorinated organic
molecules, including 2,3,7,8-tetrachlorodi-
benzo-p-dioxin.
Several conpanies and research in-
stitutions have dechlorination processes.
Three of the dechlorination technologies
include the Acurex process, PPM process,
and Sunohio PCBX process. These processes
have been found effective in reducing the
toxicity of PCBs and other chlorinated or-
ganics.
The Acurex process uses a sodium re-
agent with a proprietary constituent. The
treatment stream is filtered prior to fur-
ther treatment. After filtration, the
stream is transferred to a reaction tank
where it is mixed with the sodium/ proprie-
tary constituent mixture. The products of
this reaction are a sodium chloride stream
and another non-toxic effluent.
The Sunohio PCBX process, unlike the
Acurex process is a continuous treatment
operation. This process uses a proprietary
reagent to convert the PCB molecules to
metal chlorides and polyphenyl compounds.
>99.9999 >99.9999 >99.9999
N/R
N/R
0.0425
0.0227
0.0024
0.0013
The PPM process uses a proprietary so-
dium reagent to dechlorinate organic mole-
cules. This process is a batch process. A
solid polymer is generated from the reac-
tion and this substance, although regu-
lated, can be more easily disposed of than
the chlorinated organics.
The above described processes may be
applicable to many chlorinated organics,
such as PCBs, dioxins, solvents, and pesti-
cides (31).
The Acurex process has been tested for
removal of chlorinated waste from soils. A
commercial unit designed to treat soil is
1n the early stages of development. In
transformer oil, 2,3,7,8-TCDD has been re-
duced from 200-400 ppt to 40 +_20 ppt using
the Acurex process. This system cannot be
used on aqueous streams (31).
The Sunohio PCBX process has reduced
PCB contaminated transformer oil from 225
ppm to 1 ppm in one pass through the sys-
tem. It is believed that, by passing the
oil through the system three times, PCBs
can be reduced from 3000 ppm to less than 2
ppm. This process was used at a number of
sites, including Maxwell Laboratories in
San Diego, where 163,000 gallons of PCB
contaminated oil were processed, and
293
-------
Chevron in El Segundo, CA» where a total
of 5,250 gallons of oil initially contain-
ing 420 ppm to 1500 ppm PCBs were processed
(31, 32).
The PPM process has reduced the PCB
concentration in contaminated oil from 200
ppm to below detection levels. As with all
of the sodium processes, this process
cannot be used on aqueous wastes (31),
As part of its program to identify
and encourage new and effective processes
for treating dioxin wastes, the USEPA's
Hazardous Waste Engineering Research
Laboratory (HWERL) has supported research
with the use of alkaline polyethylene
glycols (APEGs) to destroy dioxins. The
ultimate objective of the APEG program is
to develop reliable and cost-effective
techniques for the destruction and/or
detoxification of toxic halogenated organ-
1cs, Including 2,3,7,8-TCDD. The specific
objective of current research is to deter-
mine the degree of effectiveness of APEG
reagents singly and in conjunction with UV
Irradiation (14).
Laboratory experiments using 1,2,3,4-
TCDD have demonstrated the ability of APEG
reagents to chemically reduce concentra-
tions of that isomer in wastes. While
laboratory studies have investigated in-
sltu applications of APEG to contaminated
soils, recent research has focused on slur-
ry application of the reagent to dioxin
contaminated wastes and soils. The slurry
approach incorporates reagent recovery and
recycle.
While the chemical dechlorination
processes are at least theoretically
applicable to dioxin waste streams, more
research and testing is required before
adequate technical comparisons can be
drawn between the capabilities of the
thermal destruction processes and the
chemical dechlorination processes. The
fact that the sodium based processes
cannot be applied to aqueous waste
streams will limit the applicability of
these systems.
Supercritical Fluids
Some supercritical fluids, i.e. liq-
uids or gases at or above their critical
temperatures and pressures, have been found
to be excellent solvents for organic
wastes.
In the supercritical water process de-
veloped by the MODAR Company, Natick, MA,
an aqueous waste stream is subjected to
temperatures and pressures above the criti-
cal point of water, i.e., that point at
which the densities of the liquid and vapor
phase are identical. (For water the criti-
cal point is 714°F and 219 atmospheres).
In this supercritical region water exhibits
unusual properties that enhance its cap-
ability as a waste destruction medium.
Because oxygen is completely miscible with
supercritical water, the oxidation rate for
organics is greatly enhanced. Also, inor-
ganics are practically insoluble in super-
critical water. This factor should allow
the inorganics to be removed from the waste
streams. The ideal result is that organics
are oxidized rapidly and the resultant
stream is virtually free of inorganics
(14, 15).
A brief summary of this concept:
a. Waste is slurried with make-up
water to provide a mixture of 5
percent organics. The mixture is
heated using previously processed
supercritical water and then
pressurized.
b. Air or oxygen is pressurized and
mixed with the feed. Organics are
oxidized in a rapid reaction.
(Reaction time is less than 1
minute.) For a feed rate of 5
weight percent organics, the heat
of combustion is sufficient to
raise the oxidizer effluent to
93QOF.
c. The effluent from the oxidizer is
fed to a salt separator where in-
organics are precipitated out.
d. Waste heat from the process can be
reclaimed to provide sufficient
energy for power generation and
high pressure steam.
Bench scale experiments with the
supercritical water process have yield-
ed organic carbon destruction efficiencies
of from 99.970 to 99.996 percent for a
range of chlorinated organic compounds.
For PCBs in transformer oil and a methyl
ethyl ketone (MEK) matrix, average organic
carbon destruction efficiency was 99.991
percent (33).
294
-------
Another process, using supercritical
has been proposed to extract PCBs from
soil particles (34), In this process,
which is still in laboratory development,
over 90% of PCB contamination in soil was
extracted in under one minute. Work is
continuing on this medium of treatment.
Given the low levels of TCOD contamin-
ation normally found in dioxin contaminated
wastes, it is unlikely that supercritical
C02 will prove to be viable for TCDD waste
treatment. If supercritical water can
demonstrate specific TCDD compound destruc-
tion efficiencies in the range of those for
organic carbon in the previous laboratory
work, the supercritical water process may
have potential, particularly for highly
aqueous dioxin contaminated wastes or
leachates that present a problem for the
sodium based dechlorination processes.
White Rot Fungus
Through its biodegradation research
program, the USEPA is supporting work to
identify, develop, and test microorganisms
capable of degrading highly toxic and
refractory organohalide pollutants, includ-
ing 2,3,7,8-TCDD.
Although research has not yet identi-
fied organisms considered capable of treat-
ing 2,3,7,8-TCDD, it has identified an or-
ganism, the white rot fungus, Phanerochaete
chrysosporium, that appears very promising.
The fungus secretes a unique hydrogen
peroxide-dependent oxidant capable of de-
grading lignin. Recent reports on this
research indicate that the substrate is not
only effective in degrading lignin, but
also is effective in degrading organo-
halides such as lindane, DDT, 4,5,6-tri-
chlorophenol and 2,4,6 trichlorophenol
(35).
It is proposed that the fungus may
prove to be suited for use in biotreatment
processes for the degradation of refractory
organic pollutants. To date work with this
primary alternative has been confined to
the laboratory. The EPA's HWERL is plan-
ning to test the enzyme system at several
contaminated sites in the future (14).
SUMMARY
This article has identified and dis-
cussed many alternative technology options
for treating dioxin wastes. As the country
continues to move away from land disposal
towards more acceptable techniques for man-
aging hazardous wastes, processes such as
the ones discussed will become more common
as waste management options. While many of
the processes described here may now appear
to be expensive or exotic, as the EPA moves
to implement the landfill restrictions
imposed by Congress, these processes may
assume more practicality. Most certainly,
the adoption of advanced technologies such
as those described here will measurably
improve the quality of the environment.
REFERENCES
1. The Hazardous and Solid Waste Amend
ments of 1984. Approved by Congress,
October 1984.
2. Esposito, M. P., T. 0. Tiernan, and
Forrest Dryden. Dioxins. EPA Publica-
tion « EPA-600/2-80-197, November
1980.
3. Rappe, Christopher. "Analysis of
Polychlorinated Dioxins and Furans,"
Environmental Science and Technology,
Vol. 8, No, 3, 1984, pp. 78A-90A.
4. Milnes, M. H. "Formation of 2,3,7,8-
Tetrachlorodibenzodioxin by Thermal
Decomposition of Sodium 2,4,5-Trichlo-
rophenate." Nature (condon), August
6, 1971.
5. Langer, H. G., T. P. Brady, and P. R.
Briggs. Formation of Dibenzo-Dioxins
and Other Condensation Products for
Chlorinated Phenols, and Derviation
Environmental Health Perspectives,
5:3-9, September 1973.
6. U.S. Environmental Protection Agency,
40 CFR, Parts 261, 264, 265, 270 and
775. Hazardous Waste Management Sys-
tem, Dioxin Containing Wastes, Rule.
January 14, 1985.
7. McGaughey, J. F., et al. Final Report
in EPA Contract i 68-02-3148, Work
Assignment 10. USEPA, Office of Solid
Waste, September 1984.
8. Lee, Anthony. Analysis of Technical
Information to Support RCRA Rules for
Dioxin Containing Waste Streams. Sub-
mitted to P.E. desRosiers, USEPA,
Contract # SW-6242, NASX, July 1985.
295
-------
9. U.S. Environmental Protection Aency,
40 CFR Part 264, Subparts J, K, L, M
and N.
10. U.S. Environmental Protection Agency,
40 CFR Part 264, Subpart 0.
11. Yezzi, J. J., J. E. Brugger, I.
Wilder, F. Freestone, R. A. Miller, C.
Pfrommer and P. Lovell. "Results of
the Initial Trial Burn of the EPAORD
Mobile Incineration System." Proceed-
ings of the 1982 National Waste Proc-
essing Conference Engineering: the
Solution, ASME, New York.
12. Brugger, J. E., et al. "The EPA-ORD
Mobile Incineration Systems Present
Status," 1982 Hazardous Material
Spills Conference.
13. "Dioxin Trial Burn Data Package: EPA
Mobile Incineration System at the
James Denney Farm Site, McDowell, Mis-
souri." Prepared by IT Corporation
for EPA HWERL-Cincinnati, EPA Contract
# 68-03-3069, June 21, 1985.
14. Status of Dioxin Research in the U.S.
Environmental Protection Agency. A
report prepared by the Office of Re-
search and Development for the Science
Advisory Board. Prepared by the USEPA
"Dioxin Team," August 1985.
15. Freeman, H. M. "Innovative Thermal
Processes for the Destruction of Haz-
ardous Waste." Separation of Heavy
Metals and Other Trace Contaminants.
AIChE Symposium Series, Number 243,
Volume 81, 1985.
16. Headley, Kenneth N., et al. Para-
metric Study for the Destruction of
Carbon Tetrachloride in the Huber Ad-
vanced Electric Reactor. Report for
the Texas Department of Water Re-
sources, June 1984.
17. Lee, K. W. and D. S. Lewis. J. M.
Huber PCS Destruction Process Trial
Burn Report. Report by Radian Corpor-
ation, October 1983.
18. Field Demonstration of the Destruction
of Dioxin in Contaminated Soil Using
the J. H. Huber Corporation Advanced
Electric Reactor. Roy F. Weston.
Inc., February 11, 1985.
19. Daily, P. L. Performance Assessment
of Portable Infrared Incinerator"
HMCRI Conference, Washington,. 1985.
20. Charlesworth, David, et al. "Electri-
cally Fired Incineration of Combusti-
ble Radioactive Waste." Paper pre
sented at the 1985 National Conference
on Environmental Engineering, Boston,
July 1985.
21. Stoddard, S. K., et al. Alternatives
to the Land Disposal of Hazardous
Hastes; An Assessment for California.
Governor's Office of Appropriate
Technology, 1981.
22. Mitre Hazardous Waste Control Technol-
ogy Data Base - A Profile of Incinera-
tion Facilities. USEPA Contract
# 68-03-3021, November 1982.
23. Assessment of Incineration as a Treat-
ment Method for Liquid Organic Hazard-
ous Waste. Background Report 1: De-
scription of Incineration Technology,
USEPA, Office of Planning and Evalua-
tion, March 1985.
24. Ackerman, D. 6., et al. Project Sum-
mary. At-Sea Incineration of PCB
Containing Wastes on Board the M/T
Vulcanus, EPA 600/57-83-024, June
1983.
25. Wilkinson, R. R., 6. L. Kelson and F.
C. Hopkins. State of the Art Report -
Pesticide Disposal Research, USEPA,
OR&D, Washington, EPA 600/2-08-183,
September 1978.
26. Exner, J. H., et al. "Process for De-
stroying Tetrachlorodibenzo-p-dioxin
in a Hazardous Waste." Detoxification
of Hazardous Waste, Ann Arbor Science,
T98T:
27. Brunner, C. E. Incineration System
Selection and Design, Van Nostrand
Reinhold Company, New York, 1984.
28. Engineering Handbook for Hazardous
Waste Incineration. EPA Report # SW-
889, September 1981.
296
-------
29. Richman, W. S., et al. "Circulating
Bed Incineration of Hazardous Wastes."
Paper presented at AIChE 1984 Summer
National Meeting, Philadelphia, PA,
August 1984
30. Process Demonstration Test Report for
Trial Burn of PCB Contaminated Soils.
SA Technologies, Submitted to USEPA
Office of Toxic Substances, August
1985.
31. Briefing Technologies Applicable to
Hazardous Waste. Metcalf and Eddy,
Inc., Prepared for USEPA, HWERL, May
1985.
32. Radimsky, Oan and Arvind Shah. Evalu-
ation of Emerging Technologies for the
Destruction of Hazardous Maste. EPA
Cooperative Agreement R-808908. Haz-
ardous Waste Research Laboratory,
January 1985.
33. Olexsey, R. A. "Supercritical Fluids
Processing of Industrial Wastes,"
Presented at the 54th Annual Confer-
ence, Water Pollution Control Federa-
tion, Detroit, Michigan, 1981.
34. Brady, B. 0., et al. "Supercritical
Extraction of PCB Contaminated Soils."
Paper presented at the International
Conference on New Frontiers for Haz-
ardous Waste Management.
35. Bumpus, J. A., et al. "Biodegradation
of Environmental Pollutants by the
White Rot Fungus Phanerochaete Chry-
posporium." Paper presented at the
EPA HWERL llth Annual Research Sympo-
sium, Cincinnati, Ohio, 1985.
297
-------
EVALUATION OF ON-SITE INCINERATION
FOR CLEANUP OF DIOXIN-CONTAMINATED MATERIALS
F. Freestone, R. Hazel, I. Wilder, J. Brugger, and J. Yezzi
US Environmental Protection Agency
and
R. Miller, C. Pfrommer, R. Helsel, and E. Alperin
IT Corporation
ABSTRACT
The EPA Mobile Incineration System (consisting of a kiln, secondary combustion
chamber, and air pollution control unit, each individually trailer-mounted, and a
separate trailer fitted with continuous stack gas analysis capabilities) was
rigorously tested in 1982-1983 using PCB-contaminated liquids and other chlorinated
organic fluids. Destruction and removal efficiencies of at least 99.99991 were
consistently attained at a heat release of 10 million Btu/hr. Based upon these
performance data, a project was initiated to evaluate the technical, economic, and
administrative viability of on-site incineration of dioxin-contaminated materials
in southwest Missouri. During 1984, the system was extensively modified for field
use. Solids shredding, conveying, and weighing units were performance-tested in
Edison, NJ, with a wide variety of uncontaminated soils and other solid wastes.
Additionally, permit documentation, operating and safety protocols, site
modification planning documentation, and a risk assessment for the planned trial
burn in Missouri were prepared. An aggressive public information activity was
conducted by EPA Region VII to disseminate planning and permitting information to
Missouri residents.
Simultaneously, laboratory and pilot plant studies were carried out to
establish optimum kiln conditions for decontamination of soils by fully
volatilizing the organic contaminant, i.e., dioxin. These studies indicated that
the conditions necessary to decontaminate soils thermally could be achieved in the
kiln. Previous laboratory efforts by others established that dioxins could be
destroyed using time, excess air, and temperature conditions achievable in the
secondary combustion chamber (SCC) of the EPA system. The EPA system was judged to
be more than adequate for detoxifying dioxin-contaminated solids and liquids, and
could be reasonably expected to achieve a successful dioxin trial burn.
Accordingly, the system was transported to the Denney Farm site in McDowell,
Missouri.
From January to March 1985, extensive field shakedown activities were
conducted, followed by a trial burn on dioxin-contaminated liquids and solids in
April. Results indicated that destruction and removal efficiencies exceeding
99.99992 were achieved for 2,3,7,8-TCDD; no products of incomplete combustion were
detected. Additionally, the kiln ash and process wastewater byproducts were shown
to be "dioxin free" and in accordance with guidelines proposed by EPA's Office of
Solid Waste.
A field demonstration is currently underway and a further trial burn on solid
RCRA- and TSCA-designated materials will be conducted during the Spring of 1986.
This paper will present a detailed discussion of the field application of the EPA
Mobile Incineration System for destruction of dioxin-contaminated materials at the
Denney Farm site.
298
-------
INTRODUCTION
The continued discovery of abandoned
hazardous waste sites as a result of
Superfund investigations has placed
increasing pressure on the US Environmental
Protection Agency (EPA) to find alternate
solutions for treating and disposing of
toxic and hazardous wastes. The decreasing
availability of landfill sites and the
increasing public opposition to toxic and
hazardous waste transport have added to the
pressure. The treatment and disposal
problem is particularly acute in the case
of the highly toxic dioxin isomer
2,3,7,8-TCDD. In recognition of these
difficulties, EPA's "Dioxin Strategy"
(November 28, 1983) indicated that
high-temperature incineration is a
potential dioxin treatment method
warranting further evaluation by the Office
of Research and Development (ORD).
A promising, publicly acceptable
approach to waste site problems is the use
of mobile or transportable cleanup systems
that can be brought to a waste site, used
to treat or destroy the hazardous
materials, and then removed from the site.
Public objections are reduced because there
is neither the creation of a permanent
waste disposal site to which other wastes
could be brought nor the continuing
transport of wastes through a community.
Mobile incineration is a currently
available, on-site cleanup technology that
is quite promising. High-temperature
incineration is an acceptable method for
the destruction of hazardous substances
cited under the Resource Conservation and
Recovery Act (RCRA) regulations and of PCBs
identified by the Toxic Substances Control
Act (TSCA) regulations.
Accordingly, the ORD Releases Control
Branch (RCB), a component of the Hazardous
Waste Engineering Research Laboratory
(HWERL), at the request of EPA Region VII,
embarked on a field validation project to
evaluate the existing EPA-ORD Mobile
Incineration System (MIS) for on-site
treatment and disposal of toxic and
hazardous wastes, particularly soils
contaminated with 2,3,7,8-TCDD. This
project is part of an ongoing
Superfund-sponsored program to evaluate and
promote the commercialization of processes,
methods, and prototype devices for on-site
cleanup of Superfund sites. Specifically,
the goals of this project are to evaluate
the technical and economic feasibility of
the Mobile Incineration System, establish
procedures for obtaining Federal, State,
and local permits, and gauge the reactions
of the public to the use of this system.
This paper is an interim report on the
evaluation of incineration for on-site
treatment of hazardous substances and
dioxin-contaminated materials.
Laboratory and pilot-scale work to
determine the treatability of
dioxin-contaminated soils in the MIS proved
very promising. As a result, the MIS was
prepared for field operations to conduct a
very carefully controlled trial burn on
dioxin-contaminated liquid and solid
materials at the Denney Farm site in
southwestern Missouri. Subsequently, field
validation tests on a variety of
dioxin-contaminated feeds began in July,
1985. The field work should be completed
during the winter of 1986 and a detailed
Final Report will be available thereafter.
All documentation will be publicly
available.
The final report will include detailed
descriptions of the activities of this
project and detailed laboratory study
information on the MIS, such as the minimum
times and temperatures for thermally
removing dioxins from soils. Detailed
analyses of lessons learned from the field
operations with the MIS will be provided.
The report will include evaluations of the
technical, economic, and institutional
viability of on-site incineration,
including detailed total and unit cost data
and the identification of permitting
problems. The report will also contain an
annotated index to the additional
documentation arising from or related to
this project, including (1) the detailed
permit application materials for RCRA,
TSCA, and NEPA, (2) the delisting petition,
(3) risk assessment documentation, (4)
299
-------
detailed operation and maintenance manuals
for the MIS, (5) safety manuals for field
operations, (6) detailed plans and
specifications for constructing the MIS
(including detailed cost estimates), (7)
detailed plans and specifications for the
field installation, (8) detailed trial burn
reports arising from the liquids and solids
trial burns, and (9) a feasibility study on
a modular, transportable incineration
system. All documentation will be publicly
available through the National Technical
Information Service.
BACKGROUND
The Mobile Incineration System consists
of a refractory-lined rotary kiln, a.
secondary combustion chamber (SCC), and air
pollution control equipment mounted on
three heavy-duty semi-trailers. Monitoring
equipment is carried by a fourth, smaller
trailer. Other ancillary equipment is
assembled at the site, as needed. A
detailed description of the Mobile
Incineration System has been previously
published (1-4).
LIQUID TRIAL BURN
The ability of the MIS to destroy toxic
and hazardous liquid organic wastes while
complying with applicable Federal and State
regulations was demonstrated by carrying
out. a Liquid Trial Burn of five tests
conducted in three phases from September,
1982 through January, 1983 at the EPA
facility in Edison, NJ. The tests
evaluated the ability of the MIS to destroy
tetrachloromethane (carbon tetrachloride),
dichlorobenzene, trichlorobenzenes,
tetrachlorobenzenes, and PCBs while
controlling the emissions of HC1 and
particulate matter. A total of 25 test
runs was conducted during which the
incinerator's operating conditions were
monitored and an extensive sampling and
analytical program was conducted. Federal
and/or State observers were on site
throughout the entire trial burn to ensure
that the incineration system was operated
safely and in accordance with the trial
hum permits.
The high combustion and destruction
efficiencies measured during the Liquid
Trial Burn clearly demonstrated that the
EPA Mobile Incineration System is an
effective device for the destruction of
hazardous organic material. In fact, the
level of combustion and destruction
reported was essentially based on
analytical limitations of measurement
rather than on the actual finding of
hazardous components in the stack
emissions. The results of the trial burn
indicate that the system met or exceeded
all applicable Federal requirements for
incineration systems (Table 1). The system
was then deemed to be ready for field
operations on liquids.
MISSOURI PROJECT DESIGN
The first step toward field operations
was selection of a site for the
demonstration. Agreements were reached in
April, 1984 to operate the Mobile
Incineration System on the Denney Farm near
McDowell, MO, where over 90 drums of buried
dioxin-contaminated wastes had been
excavated and stored in a diked shelter. A
second covered concrete basin on the site
contained over 240 cubic yards of soil that
had become contaminated when the buried
drums leaked. The site was considered
desirable by Syntex (current owner of the
wastes), Region VII, the Missouri
Department of Natural Resources (MDNR), and
HWERL because of its remoteness —
approximately one mile from a public road,
and in a rural area.
A wide variety of dioxin-contaminated
wastes, such as soil, liquids, drums,
trash, and chemical solids, were on the
site and could be used to demonstrate the
versatility of the Mobile Incineration
System. A variety of nearby soil types
could be used to demonstrate that
incineration could decontaminate any
dioxin-containing soil found anywhere in
Missouri.
The overall project design for
evaluation of the MIS in Missouri included
the following elements, each of which is
discussed in greater detail below.
300
-------
TABLE 1. RESULTS OF LIQUID TRIAL BURN
Parameter
Reg.
Limit
2b
4d
Kiln
Temp., oc 900
890
1180 1180
7.5 8.8
1.96 2.10
99.999 99.999
sec
Temp., °C 1200+/-100f
02, % - 3lf
Retention Time, sec 2^
Combustion Efficiency, % 99.9f
HC1 Removal Efficiency, % 999
ORE
CClJ 99.999
DiClB 99.99
TriClB 99.99
TetradB 99.99
PCB h
Participate mg/Nm3 @ 7% 02 1809 6.9
42.7
900
1190
8.0
2.06
99.999
99.95
99,99996
99.99998
22.1
980
1220
7.2
1.97
99.999
99.98
99.9998
99.9994
99.9998
64.9
960
1250
6.9
2.03
99.999
99.99
99.99993
99.9998
99.99991
35.2
a Test 1
b Test 2
c Test 3
d Test 4
e Test 5
f TSCA
9 RCRA
h No ORE limit for liquid PCB
fuel oil only - baseline - all results average of 3 test runs
1.2% Iron oxide in fuel oil for particulate test
21% CC14, 29« DiClB in fuel oil
11% Askarel in fuel oil
39« Askarel in fuel oil
* CC14 = tetrachloromethane; DiClB = o-dichlorobenzene; TriClB = trichlorobenzene;
TetraClB = tetrachlorobenzene; PCB = polychlorinated biphenyls.
1. Planning and Permits
2. Legal and Public Relations Activities
3. Modifications to the MIS
4. Laboratory and Pilot Studies
5. Operating Procedures and Operator
Training
6. Site Preparation and Logistics
7. Field Shakedown and Preliminary Testing
8. RCRA/TSCA Trial Burn
9. Field Demonstration
10. Final Reports
11. Future Use
PLANNING AND PERMITS
This activity included the preparation
of: detailed contractors work plans;
applications for Federal permits (Clean Air
Act, Toxic Substances Control Act, National
Environmental Policy Act, Resource
Conservation and Recovery Act) and State
permits (Missouri Department of Natural
Resources, (MDNR)); "delisting petitions"
for assuring the cleanliness of MIS solid
and liquid byproducts for both the trial
burn (Table 2) and the field demonstration;
and a risk assessment of MIS operations on
dioxins and a "health survey" required by
301
-------
TABLE 2. MDNR AND US EPA DELISTING PARAMETERS
(FOR TRIAL BURN ONLY)
A. Per 40 CFR 261 Subpart C
Ash Solids
Ignitability
Corrosiyity
Reactivity
EP Toxicity
pH = 2.0-12.5
Not reactive with water
As per 40 CFR, 261,24,
Table 1, and App. II,
except mercury0
Scrubber Waste Liquids
pH = 2,0-12,5
Not reactive with water
ICP scan-heavy metals,
except mercury as per
Table 1=
B. Specific Substances
Toxic Constituent
Dioxins/Dibenzofurans"
2,3,4-Trichlorophenol
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
2,5-Dichlorophenol
3,4-Dichlorophenol
2,3,4,5-Tetrachlorophenol
2,3,4,6-Tetrachlorophenol
1,2,4,5-Tetrachlorobenzene
1,2,3,5-Tetrachlorobenzene
Hexachlorophene
Poly chlorinated Biphenyls
Benz(a)pyrene
Benz(a)anthracene
Chrysene
Dibenzo(a,h)anthracene
Indeno(1,2,3-c,d Jpyrene
Benz(b)fluoranthene
Solids
1 ppb
100 ppm
100 ppm
1 ppm
350 ppb
100 ppm
1 ppm
1 ppm
100 ppm
100 ppm
200 ppm
2 ppm
5 ppm
5 ppm
50 ppm
5 ppm
5 ppm
5 ppm
Concentration
Scrubber Water
10 ppt
10 ppm
10 ppm
50 ppb
15 ppb
10 ppm
50 ppb
50 ppb
10 ppm
10 ppm
5 ppm
1 ppm
10 ppb
10 ppb
1 ppm
10 ppb
10 ppb
10 ppb
a The ash would not be ignitable after having passed through a kiln
and having reached approximately 750°C at the time of discharge.
b The scrubber wastewater, is not considered ignitable.
c Analysis will be for the following metals: arsenic, barium, cadmium,
chromium, lead, nickel, selenium, silver.
<* Weighted average of tetra-, penta-, and hexa-isomers using weighting
factors related to toxicity.
MDNR (the risk assessment and health survey
documents were prepared by ORD's Exposure
Assessment Group).
LEGAL AND PUBLIC RELATIONS ACTIVITIES
Legal Arrangements
The activities conducted at the Denney
Farm site have required contractual
agreements between the two principal
parties, namely, EPA and Syntex
302
-------
Agribusiness, Inc. (Syntex) and between the
EPA's operating contractor, IT Corporation,
and both of the principal parties. Three
agreements were prepared, each of which was
interdependent.
Additionally, an access agreement with
the landowner, the James Denney family, was
modified to enable operations to take place
at the farm. These complex arrangements
among Syntex, IT, and the Denneys were
negotiated by Region VII's legal counsel
with technical inputs from EPA's HWERL
Releases Control Branch and IT Corporation.
Public Relations Activities
As a result of an agressive public
information effort conducted by Region VII,
the Mobile Incineration System was welcomed
in Missouri and was operated in a favorable
atmosphere of increasing public acceptance
of incineration. The activities included
press conferences, presentations to civic
organizations, an open house and tour of
the site installation prior to operations,
the required RCRA/MDNR permit public
hearing, and a "technical day" during which
preliminary trial burn data were publicly
released and a site tour conducted for
interested parties. These activities were
critically important to the overall success
of the project, and received the personal
attention of the Regional Administrator --
a key element in the success of the project.
MODIFICATIONS TO THE MIS
Several changes were made to the
original MIS design, including general
modifications affecting the refractory; the
burner controls; the stack gas monitoring
system; the electrical system; and the
design, specification, procurement,
installation, and shakedown of a solids
feed system. Both hot and cold tests were
performed.
The main purpose of the cold test was
to evaluate the factors that affect the
retention time of solids in the kiln,
including rotational speed and inclination
angle of the kiln, type and rate of feed of
material to the kiln, and the use of
flights at the inlet and heat-economizer
chains at the outlet of the kiln. Based on
the observations made during the cold test,
design and operational changes were made to
improve the flow of solids through the kiln.
As a result of the hot tests, it was
concluded at the time that the kiln
capacity was limited by the solids feed
system and that particulate emissions would
not be a problem. The study suggested
changes in operating procedures, such as
operation of the kiln in a sloped position,
decreasing the rotational speed, and
operating with less excess air.
Upon completion of the hot tests, the
kiln and SCC were opened and inspected.
The examination revealed that a large
amount (up to 9 in) of solids had been
carried over into and deposited at the
front end of the SCC. Although some
carryover had been expected, the quantity
and location of the deposits were not. It
appeared that the "spin vane," a device
designed to create swirling turbulence in
the gas stream and thus enhance combustion
efficiency, was a major factor in the
deposition of particulates at the front, end
of the SCC. At that time, it was decided
to make no changes in the spin vane, but to
carefully monitor the operations and
frequently check for the presence of solids
buildup in the SCC. A number of
modifications were made to correct several
minor problems, and a second hot test was
run to ensure that these problems had been
corrected. Following the hot test, the
system was dismantled and prepared for
shipment to Missouri.
LABORATORY AND PILOT STUDIES
The goals of these studies were to
determine whether the objective of
decontaminating the soil to less than 1 ppb
dioxin was feasible, given the operating
limits of the MIS, and to develop
recommended operating conditions for the
demonstration run.
Laboratory Studies
In treating contaminated soil with an
incinerator or other thermal treatment
device, two processes can achieve
decontamination: volatilization with
resulting vapor separation followed by gas
phase combustion, and thermal decomposition
within the solid soil matrix. The
mechanisms that determine the relative
importance of these "treatment" processes
are volatilization, diffusion, and
thermochemical reaction, each of which
occurs at different temperature-dependent
rates.
303
-------
Theoretical and empirical studies of
volatilization and diffusion of chemicals
within a static soil at normal
environmental conditions are not directly
translatable to treatment under dynamic,
high-temperature conditions, such as exist
in a rotary kiln. Thermochemical behavior
of specific chemicals has been studied to a
limited extent; estimates of thermal
stability and decomposition kinetics are
important in establishing guidelines for
Incineration performance requirements.
However, these studies have dealt with gas
phase reactions and pure (single component)
systems. Also, alterations to the chemical
and physical characteristics of the primary
natural organic components and of certain
inorganic constituents of soil occur at
temperatures far lower than typical
incineration conditions. The complexity
and variability of a soil and its
interaction with specific chemicals
frustrates any straightforward analysis of
the effect of typical incineration
conditions on a particular contaminated
soil.
Prior to the initiation of laboratory
treatabllity tests, information on the
soils at the various confirmed
dioxin-contaminated sites in Missouri was
collected. From this list, four sites were
selected as candidates for treatment. Two
of these soils, Piazza Road and Denney
Farm, were selected for laboratory
treatability testing since they covered a
wide range in pH, conductivity, organic
matter, and particle size distribution.
Samples of these two soils were
prepared for laboratory treatability
testing by air drying, screening through a
2-mm (10-mesh) sieve, and blending
thoroughly. Analysis for 2,3,7,8-TCDD in
triplicate aliquots of each prepared soil
showed relative standard deviations of less
than n. The initial average 2,3,7,8-TCDD
concentration was 563 ppb for Denney Farm
and 338 ppb for Piazza Road soils. These
relatively high levels were advantageous in
enabling the maximum range of treatability
to be investigated.
The laboratory experimental program was
divided into three series of separate
treatability tests with a total of 31
tests. The first series, using principally
Denney Farm soil, explored broad ranges of
residence time and temperature to define
the appropriate combination of conditions
that would produce desired treatment
efficiencies, as measured by a final
2,3,7,8-TCDD concentration of 1 ppb or
less. The second series used a fixed
time-temperature condition and evaluated
the effect of soil type (Piazza Road vs
Denney Farm), initial soil moisture
content, and gas phase composition on
treatability. The final series included
selected additional treatment conditions to
fill in data gaps and also several special
tests in which 5 cm "cubes" of Piazza Road
soil were prepared and subjected to various
test conditions.
In the second series of tests, there
was no significant correlation between
treatability and either moisture or
atmosphere. Soil type had an influence,
but this was determined to be due primarily
to the temperature increase that occured
with Piazza Road soil, presumably as a
result of exothermic reactions of the
organic matter that was present at much
higher levels than in the Denney Farm
soil. Once the effect of temperature was
factored into the statistical evaluation,
the true effect of soil type was relatively
minor.
In the third series of tests, using 5
cm cubes, the substantial lag in achieving
the target test temperature within the core
was due largely to the drying process. The
evaporation rate of the initial 20%
moisture content from the cube is dependent
on the heat and mass transfer
characteristics of the cube and the
external gas temperature, which in the MIS
kiln would be higher than 500°C. Another
test performed with the furnace temperature
at 800°C resulted in a significant
reduction of heat-up time of the cube; the
core reached 400°C in 19 minutes vs. 36
minutes in the 500°C furnace. Separate
analyses for 2,3,7,8-TCDD of the core and
exterior sections of the treated cubes
demonstrated that 1 ppb could be achieved
throughout and illustrated the effect of
the transient temperature condition. At an
800°C furnace temperature, a total
residence time of 33 minutes resulted in
non-detectable (less than 0.1 ppb)
2,3,7,8-TCDD concentration, whereas a
shorter residence time of about 20 minutes
showed more than 7 ppb 2,3,7,8-TCDD
remaining in the core section. The test at
500°C furnace temperature provided for 40
minutes for treatment and resulted in
approximately 1 ppb throughout the cube.
304
-------
A linear regression analysis of the
treatability data for Denney Farm soil
produced two different mathematical
relationships to predict the final
2,3,7,8-TCDD concentration at different
time-temperature conditions. The simplest
of these expressions shows a logarithmic
dependence of 2,3,7,8-TCDD concentration on
the "time integral of vapor pressure,"
which is defined by integrating the
calculated vapor pressure of 2,3,7,8-TCDD
using the specific temperature-time
profile. The vapor pressure can be
estimated using Antoine constants developed
by Schroy et al. experimentally. Figure 1
and Table 3 based on this expression depict
the time-temperature requirements to
achieve different decontamination of the
Denney Farm soil initially containing 563
ppb 2,3,7,8-TCDD.
Other observations made during the
laboratory program offered insight into the
thermochemical phenomena that occur in soil
at incineration temperatures.
Agglomeration, slagging, or fusion of soil
particles was not observed for either soil.
The cubes developed cracks and became
friable particularly within the core.
Weight-loss measurements and visual
inspection indicated that decomposition of
soil organic matter could lead to highly
variable results, depending on the exposure
TABLE 3. SUMMARY OF TIME-TEMPERATURE EFFECT
ON REMOVAL OF 2,3,7,8-TCDD
Nominal
Test
Temperature
(°C)
429
430
429
428
429
475
478
477
479
550
550
554
616
616
616
803
808
803
Time at
Test
Temperature3
(min)
0 -
15
30
90
90
0
15
30
30
0
0
15
0
15
30
30
30
90
Soil
Typeb
A
A
A
A
B
A
A
A
A
A
A
A
A
A
A
A
B
A
Residual
2,3,7,8-TCDD
Concentration
(ppb)
377
60
30.8
10.2
2.86
67
8.4
3.7
3.37/3.3QC
24
27.5
0.16
0.2
NO (0.08)
ND (0.06)
ND (0.02)
ND (0.04)
ND (0.08)
a This time begins when the target test temperature is reached; therefore,
zero time is actually six to nine minutes after start of heat-up.
b A: Denney Farm Soil; B: Reference Soil.
c Analytical duplicate; separate aliquots of treated soil were analyzed.
305
-------
900-
800-
700-
-------
or accessibility of the soil particles. In
general, weight loss in excess of that
attributable to free moisture and organic
matter was 1-9%. Loss of bound moisture
from the clay mineral content of these
soils is suspected to be the primary source
of this excess loss. Higher temperatures
resulted in greater weight loss, and Piazza
Road soil tended to have higher losses.
However, the cube experiments using Piazza
Road soil showed lower non-moisture-related
weight loss than the experiments using a
loose, thin layer of the same soil, and the
interior of the cubes turned black.
Evidently, the significant concentration of
organic matter in the Piazza Road soil was
partially retained as non-volatile products
(char) of thermal decomposition.
The results from the laboratory testing
demonstrated that the clean-up criterion of
1 ppb could be achieved at reasonable kiln
operating conditions and provided part of
the information needed to project the
specific treatment regimen (kiln residence
time and temperature) for various feed
rates and feed conditions for the MIS.
Pilot Testing and Computer Modeling
Pi lot-scale experimentation was
conducted on uncontaminated Missouri soil
from the Denney Farm site. Tests included
firing in a pilot-scale rotary kiln;.bulk
density and screen analysis; dynamic angle
of repose, thermogravimetric and
differential scanning calorimeter analysis;
and soil moisture analysis. These tests
were conducted at the Allis-Chalmers
research and test center in Oak Creek,
Wisconsin on August 28-30, 1984.
The first pilot-scale run was performed
at a constant gas temperature of 982°C
(1800°F). The gas temperature was
maintained throughout the course of the
60-minute run by varying the fuel rate at a
constant air flow. The initial moisture
content of the soil was 19i. Subsequent
runs were conducted by maintaining set
natural gas and air flow rates throughout
the run. The set flow rates were
determined from temperature profiles of
previous runs and were designed to give the
desired steady-state temperatures and
excess oxygen levels.
A significant result of the runs was
the apparent correlation of soil moisture
content to the observed fractionation of
soil chunks. In all trials with 19%
moisture, complete fractionation was
observed, yielding soil particles that were
smaller than 2 cm on any dimension. In the
single run conducted at 12.3% moisture,
slightly less fractionation occurred,
although it was similar to the 19% moisture
runs. It is apparent that the
fractionation mechanism is at least
partially dependent on the forces exerted
by steam escaping from the soil during
heating.
Trials were performed at bed
temperatures of up to 1Q36°C (18970F),
with no slagging or agglomeration
observed. This demonstrates that slagging
and agglomeration should not occur when the
MIS is used to treat Missouri soils.
The nine runs of the pilot test
provided data to evaluate the heat transfer
computer program, which contains standard
constants for the various heat transfer
modes in a rotary kiln. By operating the
kiln at various steady-state conditions, a
data base on the time-temperature profile
of Missouri soil in the batch kiln was
obtained. The heat transfer program could
be modified to predict this profile and the
difference between the observed and
predicted results used to more closely
model actual conditions.
The computer program calculates kiln
solids capacity and retention time if all
other variables are specified. For this
study, the kiln capacity was set by the
height of the corbel at the feed end of the
kiln. More correctly, the height of the
material at the kiln entrance cannot exceed
the height of the corbel. The same height
at the feed end of the kiln can result in a
variety of mass-flow kiln capacities,
depending on the kiln slope, rpm, etc.
According to the program, this variation
was not great; the calculated kiln
capacities varied from 6-81 of the kiln
volume. This agreed with earlier tests
that indicated a maximum kiln capacity of
7% of the kiln."volume.
The mass-flow computer program was run
for a matrix of kiln slopes, rotational
speeds, and feed rates. As the kiln slope
increases, its capacity also increases as
the result of a more consistent solids
depth along the length of the kiln. For a
kiln with a low slope, the slope of the
solids provides the driving force for
307
-------
solids flow. In a (din with a steeper
slope, the kiln slope provides the driving
force, allowing for more consistent solids
depth. It was therefore recommended that
the kiln be operated at its maximum slope.
For the MIS on a level pad, this slope is
approximately 2-3%.
Flights along the kiln length increase
solids throughput at a given slope and
rpm. This limits the minimum feed rate and
reduces the effective kiln length. For
these reasons, flights have been removed
from the length of the kiln. However, the
eight flights at the feed end of the kiln
prevent spillage over the corbel remain.
OPERATING PROCEDURES AND OPERATOR TRAINING
Operating and safety procedures for the
Mobile Incineration System were updated
based on the new solids feed capability,
the experience gained during previous
operations, the special considerations
required for handling dioxins, and the
site-specific requirements for Missouri.
The operating and safety manuals were
revised to incorporate the new procedures.
All operators and other key personnel went
through an extensive training course that
included system operation, operating
safety, and other site safety procedures.
Further, an industrial hygienist
remained onsite during the Dioxin Trial
Burn. He observed and corrected operating
and sampling practices, conducted personnel
monitoring, and made daily site safety
inspections. All site subcontractors,
government observers, and visitors to the
site for any length of time were given
site-specific safety presentations.
SITE PREPARATION AND LOGISTICS
Site Planning and Preparations
Upon selection of the specific site for
the Mobile Incineration System
demonstration, detailed engineering and
design was commenced to satisfy
geographical, operating, and permitting
requirements.
To establish geographical requirements,
a site survey was conducted that considered
grades, elevations, property boundary
lines, access routes, and existing road
conditions. Maximum consideration was
given to leaving "untouched," to the
greatest extent possible, the natural
contours, vegetation, and woodlands. The
final site layout is shown in Figure 2.
The physical dimensions of the solids
feed handling system and the location of
the contaminated materials in the Drum
Storage Building (Figure 2) determined the
relative location of the incinerator. The
concept was to maintain the incinerator in
a "clean area" while violating the
integrity of the secure, "contaminated"
boundary (fenced area around existing
buildings) as little as possible. This
would contain the contaminants within the
incinerator system and prevent any increase
in contaminated area. All the equipment
was placed on a 4-inch-thick, poured
concrete pad surrounded by a 6-inch-high
concrete dike to facilitate containment and
clean-up in the event of a spill or leakage
of contaminated material in the incinerator
area.
The remote, undeveloped nature of the
site necessitated installation of a deep
well and the water supply system required
for operations. Fuel oil, propane, water,
and wastewater tanks were installed.
Portable office and storage trailers were
used extensively to minimize the
construction of fixed structures. The MIS
was enclosed In a 40 X 225 ft prefabricated
shed building for personnel and equipment
weather protection. A guard service
company was contracted to provide site
security and to control access of personnel
during operations. A log was kept for all
persons as they entered and left the site.
Six underground telephone lines were
installed; the telephones were located
throughout the site. For backup emergency
communications, two FM radio base stations
were installed that operated on the Monett,
MO, Police Department frequency.
Transportationand Setup of Incinerator
The MIS was transported to and set up
on the Denney Farm site in mid-December,
1984. In addition to the four main
trailers of the incineration system, five
other trailers were required, both to
complement the operation (e.g., a personnel
decontamination trailer) and to transport
the auxiliary equipment and spare parts.
Additional support equipment was either
provided by Syntex, purchased, or leased to
complete the incineration system setup.
308
-------
•\
\
\
1
t
1
1
1
1
1
,'s
/
/
/
/
f i
H
p
LJ
?
3
4
imt
&
f
1
1
j
0[
a
D
Q
0
in
1
1
*
•t
t
j
i
/">
-£>
O
m.
1
1
i
i
i
i
1
I
I
/t
IB.
en °
16
ts
/*
n
a
f
S9 . V
1
1
|
I
1
'I
i
1 " 1
D ">
•nnu<«r/t4«xW* I
^T' I
O,.
V
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
?6
LEGEND
DESCRIPTION
Office trailer
Operations trailer
Sampling and Analytical trailers
;Utility equipment and.storage area
Spare parts and storage trailer .
.Mobile incinerator building
Stack monitoring trailer
? Clarifier tanks with
filtration skids
2 15,000-Gal wastewater
storage tanks :
7,500-Gal process water
storage tank
7,500-Gal alkaline storage.tank
11,000-Gal fuel oil storage tank
HEPA filter
Hydraulic skid for shredder
1,000-Gal waste oil feed tank
. Drum storage building (Syntex)
Personnel decon. station
Operating and Sampling crew
change trailer
Microbiological degradation
building No, 2 (Syntex)
Microbiological degradation
building No. 1 (Syntex}
Water heater .shed (Syntex)
Equipment decon, building (Syntex)
Personnel decon, shed (Syntex)
Guard shed (Syntex)
EPA conference trailer
Mobile Incineration System
Figure 2... Si,te layout for MIS at Denney Farm.
Syntex provided four storage tanks ranging
from 10,000-15,000 gal capacity, to be used
for makeup water, alkaline solution, and
wastewater storage. Seven trailers were
leased to provide lunch and rest space for
the operators, office and work space, and
shelter for the site guard service. Other
leased auxiliary equipment included such
items as forklifts, a backhoe, power
generators, air compressors, storage vans,
and space heaters. Fuel oil and propane
storage tanks were furnished by local
suppliers.
FIELD SHAKEDOWN AND PRELIMINARY TESTING
Final preparations,, component checks,
and on-site personnel safety training were
completed by early January and the
incineration system was then started up
with fuel oil to check its performance
after transport from New Jersey. The
startup proceeded relatively smoothly and
the system was brought to operating
temperature within two days; however, the
solids feed shredder jammed during
below-freezing temperatures shortly after
being started with clean Missouri soil.
The failure was caused by a broken spline
drive resulting from the high viscosity of
the hydraulic drive fluid and differential
contraction of internal motor components,
both resulting from the subfreezing
temperatures.
Extremely cold weather (7:15°F,
309
-------
-50°F wind chill) that lasted for several
days created a number of problems with
freezing water lines and gelled fuel oil;
the operating crew spent much time
struggling to keep the system operating.
The incineration system was restarted early
in February, but a number of different
problems hampered initiation of the trial
burns. Most of these problems involved
mechanical difficulties, or were
weather-related. The mechanical problems
were ultimately solved, the ambient
temperature rose, and the shredder has
performed well ever since.
As part of the startup plan, a blend of
15* by volume sodium sulfate and clean
local soil was fed to the incinerator to
evaluate whether that specific
concentration of salt would cause slagging
and ash removal problems in the rotary
kiln. The test was conducted because of
the potential for incineration of 23 cubic
yards of dioxin-contaminated sodium sulfate
at the Syntex Verona, MO, plant. The salt
built up as a heavy slag in the kiln;
therefore it must be fed at much lower
concentrations, if at all, in order to
minimize slag build-up in the kiln. (The
quantity of soil to be fed during the field
demonstration is insufficient to handle all
the sodium sulfate at the acceptable low
concentration.)
In preparation for Test 1 of the trial
burn (see Trial Burn Plan, below), the
tetrachloromethane/methanol liquid blend
and PCB/hexachloroethane/ montmorillonite
solids blend were fed to the incinerator.
However, before the system could be brought
to equilibrium to start sampling, a number
of mechanical problems developed, primarily
in the quench system of the air pollution
control section.
After repairing the quench pump and
installing a hydrocyclone in the quench
sump (to remove accumulations of fine
particulates being carried into the quench
system from the SCO, high rates of
tetrachloromethane/methanol were
successfully fed to the incinerator to
verify that the air pollution control
system was functioning properly.
- Additionally, preliminary stack
particulate testing with a high feed rate
Of chlorine (equal to that scheduled for
Test 1) revealed corrosion of the
carbon-steel sound attenuator stack section
by hydrogen chloride. The high corrosion
rate caused fine particles to be emitted
from the stack, thus causing the system to
exceed RCRA particulate standards. The
attenuator was subsequently replaced by a
stainless steel stack section so that the
system would comply with RCRA when high
chlorine feeds were to be used.
At this time, after consultation with
Region VII management and incinerator
operating personnel, HWERL decided to defer
the high-chlorine Test 1 and conduct the
Dioxin Trial Burn (Test 2) first.
Accordingly, dioxin-contaminated liquids
and solids were fed to the incinerator for
the first time at the end of February.
Additional minor problems were encountered,
and the system was shut down for
correction. The restart run was conducted
with minimal problems and became the first
completed dioxin trial burn run.
The February operation was difficult
for the operating crewi as well as for the
EPA and observers, primarily because of all
the minor problems that delayed the
testing. Entrained solids carryover from
the kiln became a problem because of their
accumulation in the secondary combustion
chamber in front of the lower of the two
fuel oil burners. The mass of solids
became a glass-like slag that interfered
with the burner operation. The burner was
removed each night and the solids that
could be reached were removed manually by
chipping with a sledgehammer and a steel
pipe. When the burner was removed after
the completed test run, a large hole was
found in the internal spin vane that
induces swirling turbulence to the gases
that enter the SCC. The Inconel (a
high-temperature nickel alloy) vane
evidently melted because of flame
impingement that resulted when the slag
buildup deflected the burner flame onto the
vane. Because the vane was regarded as a
critical component, the system was again
shut down.
After consultation with incineration
experts, a plan was developed in March to
rebuild the spin vane, make minor operating
adjustments, and continue with the Dioxin
Trial Burn while carefully observing the
solids buildup in the SCC. The plan
further called for more extensive
modifications during the shutdown period
after the trial burn and before the field
demonstration.
310
-------
The SCC was totally cleaned out, the
new spin vane was fabricated and installed,
and routine maintenance was performed
during March. The incinerator was
restarted April 1, and three more dioxin
trial burn test runs were executed in a
period of 4 days. There continued to be
some minor problems with the solids feed
and with solids carrover, but the runs
generally went smoothly. After completing
the dioxin test, the Syntex lagoon sludge
evaluation, Test 3, was conducted. The
sludge was the first combustible solid of
any quantity to be fed to the Mobile
Incineration System. The test went well
but, when the SCC was checked after the
test, a large amount of solids was again
observed near the spin vane.
RCRA/TSCA/DIOXINS TRIAL BURN
Trial Burn Plan
The purpose of the trial burn is to
obtain data, which, when combined with data
from the Liquid Trial Burn in Edison, NJ,
would verify that 0) dioxins and other
hazardous organic liquid and solid
materials are destroyed by incineration in
the EPA Mobile Incineration System to a
residual ash concentration of less than 1
ppb, and (2) the resulting stack emissions
do not pose an unacceptable health or
safety risk to the surrounding
communities. The trial burn was designed
to provide data to support the issuance of
the State and Federal permits required for
use of the incinerator at the Denney Farm
site in southwestern Missouri and to
provide sufficient data to enable future
use of the EPA Mobile Incineration System
on practically any hazardous organic
material at Superfund and other hazardous
waste sites.
At the Denney Farm site and at other
locations in Missouri, dioxins, PCBs, and
RCRA-regulated substances are stored or
otherwise exist as liquids and solids.
Therefore, trial burn data on these
materials in both liquid and solid form
were necessary for evaluating the
performance of the MIS. EPA and State
regulatory offices currently require trial
burn data on both liquids and solids
because of the processing differences
between liquids and solids in the kiln of a
kiln-plus-secondary-combustion-chamber-type
incinerator. In part, this arises because
organic liquids can be passed through a
burner and directly destroyed whereas
solids must first be heated by thermal
radiation and by contact with refractory in
a kiln to release organic contaminants that
are subsequently destroyed. Further,
because of the public sensitivity of
dioxins, a special dioxin trial burn — as
applied to liquids and solids — was
requested by the EPA Dioxin Disposal
Advisory Group and the Office of Solid
Waste in support of RCRA regulations
relating to the destruction of dioxins.
The trial burn program was planned to
consist of three tests. The tests each
consisted of three replicate sampling runs,
i.e., a total of nine sampling periods.
However, due to operational problems that
arose during the trial burn, Test 1 was not
executed. Test 1, a burn of PCBs and
chlorinated hydrocarbons on montmori11onite
soil, will be conducted at the end of the
field demonstration. The Trial Burn Plan
was extensively reviewed by scientific and
regulatory reviewers and formed a major
portion of the RCRA/MDNR permit application.
1. Test 1
A mixture of solid hexachloroethane and
montmori11onite (a major source of
bentonite) with adsorbed PCBs, will be
fed to the rotary kiln. In addition,
two liquid streams will be fed to the
rotary kiln, one an organic liquid
consisting of tetrachloromethane and
methane! having a low heat value, and
the other an aqueous stream containing
a portion of the organic liquid. The
objective of this test will be to
obtain for the MIS (1) a TSCA
non-liquid PCB permit, (2) a RCRA
non-liquid permit, and (3) a RCRA
liquids permit to allow the
incineration of low-heat-value organic
liquids and contaminated aqueous
liquids. This test will consist of
three runs of 8-12 hours each to
collect the required 4 m^ of stack
gas sample.
2. Test 2
During this test, 2,3,7,8-TCDD
(dioxin)-contaminated soil and dioxin-
containing waste liquids were fed to
the rotary kiln to demonstrate the
Mobile Incineration System's capability
of destroying dioxin at the 99.9999%
ORE level.. This test consisted of four
successful runs of 8-12 hours each to
collect the required 4 m^ of stack
gas sample.
311
-------
3. Test 3
Bromine-contaminated sludge was fed to
the rotary kiln to demonstrate the
Mobile Incineration System's
capabilities of controlling
bromine/hydrogen bromide emissions
while incinerating bromine-contaminated
wastes. The test consisted of three
runs of 0.5'hour each.
Quality Assurance Project Plan
Preparation activities for the trial
burn included the preparation of a Quality
Assurance Project Plan (QAPP) as required
of all EPA-sponsored research programs.
The QAPP focuses on the details of
collecting, handling, and reporting the
trial burn data and results. The QAPP also
establishes quality criteria to ensure that
the trial burn data and results are
technically sound and acceptable to
regulatory offices. This plan underwent
extensive review by technical experts, both
inside and outside of the Agency, and by
regulatory officials. The QAPP encompasses
two separate aspects of the trial burn:
the operation of incineration equipment and
the collection of routine operating data.
The plan also describes in great detail the
collection and analysis of special samples
associated with the trial burn.
The sampling techniques range from the
very simple, e.g., collecting waste water
samples for standard analyses, to the very
complex, e.g., collecting stack gas samples
for measurement of nanogram quantities of
2,3,7,8-TCDD. The Sampling and Analysis
plan also provides all the specific
procedures to be used for sample analyses.
Standard analytical protocols are used
whenever possible, but the determination of
2,3,7,8-TCDD in incinerator effluents
requires state-of-the-art analyses to
demonstrate the required destruction and
removal efficiencies. The samples were
analyzed in two laboratories to provide
independent verification of test results.
Detailed QA/QC data were sent to the EPA
Environmental Monitoring and Support
Laboratory in Las Vegas and to Region VII
for review.
Dioxin Trial Burn Results
The feed materials for the trial burn
consisted of dioxin-contaminated liquid
still bottoms and soil for Test 2, and
dioxin-contaminated lagoon sediment for
Test 3. The still bottoms, originally from
a trichlorophenol purification still, were
blended with solvents to control the
concentation of 2,3,7,8-TCDD within permit
limitations. These solutions were analyzed
prior to the trial burn to verify the
2,3,7,8-TCDD concentration and to select an
appropriate liquid waste feed rate. The
dioxin-contaminated soil came from the
original Denney Farm trench and had been
kept in a covered concrete basin as loose
soil. The soil was placed in 55-gallon
drums that were subsequently dumped into
the solids feed system. The lagoon"
sediment came from an industrial waste
storage lagoon in Springfield, MO, and
contained low concentrations of brominated
naphthalenes and 1-20 ppb 2,3,7,8-TCDD.
This sediment was stabilized with calcium
oxide and transferred to the Denney Farm
site in 55-gallon drums for processing.
The incinerator operating conditions
during the Dioxin Trial Burn were
essentially the same as those during the
previous Liquid Trial Burn successfully
conducted in New Jersey (Table 4). Waste
liquids and solids were fed to the rotary
kiln. The solids were retained in the
rotary kiln, operated at gas exit
temperature of 845° to 955°C, for
approximately 30 minutes before being
discharged into drums. The solids achieved
approximately 750°C upon discharge. The
gases from the combustion of wastes flowed
into the secondary combustion chamber where
they were heated to 1150 to 1230°C. In
the secondary chamber, the combustion gases
were mixed with excess oxygen (air) (to a
control level of 4-7% 02) and were
retained for 2,4 to 3 seconds.* The
combustion gases then passed through three
stages of air pollution control equipment
used to cool, filter, and remove acid gases
(by-products from the combustion of wastes)
and particulate matter (primarily from the
solid wastes processed). Other process
effluent streams (kiln ash, CHEAP filter
mat, and purge water) were collected and
analyzed in accordance with delisting
guidelines and the Trial Burn Plan to
determine whether any hazardous materials
were discharged from the incineration
system.
The performance of the incinerator
during the trial burn was accurately
determined by monitoring all feed and
eff-luent streams. The performance criteria
for Test 2 were the destruction and removal
312
-------
TABLE 4. INCINERATION SYSTEM OPERATING CONDITIONS SUMMARY
Parameter
Rotary Kiln
outlet gas temperature, °F
pressure, in. H£0 vacuum
combustion air flow, scfm*
atomization air flow, scfm
diesel fuel flow, Ib/hr
liquid waste flow, Ib/hr
solid waste flow, Ib/hr
ash discharge rate, Ib/hr
gas residence time, sec
excess air level, %
Secondary Combustion Chamber
outlet gas temperature, °F
pressure, in. H20 vacuum
combustion air flow, scfm*
diesel fuel flow, Ib/hr
gas residence time, sec
oxygen, vol. %
carbon dioxide, vol. %
carbon monoxide, vol. ppm
Quench System
process liquor pH
quench recycle flow, gal /mi n
makeup water flow, gal /mi n
CHEAP
process liquor pH
CHEAP recycle flow, gal /mi n
pressure differential, in. HgO
MX Scrubber
process liquor pH
MX recycle flow, gal /mi n
Stack
temperature, °F
pressure, in. h"20
gas flow rate, dscfm
Run K
1,650
0.7
1,029
79
104
233
1,158
933
2.30
57
2,178
0.8
857
211
2.56
7.70
10.9
4.1
8.7
45
12
8.6
17
19
8.7
127
164
+0.62
3,649
Test 2
Run #3
1,600
0.3
923
89
45
236
1,363
989
2.99
48
2,190
0.4
717
178
3.18
6.97
11.3
2.0
8.9
46
7
9
9
13
8.5
171
168
+0.59
2,876
Average
Run 14
1,678
0.2
1,177
90
106
234
2,068
1,531
2.05
43
2,187
0.7
758
202
2.49
6.44
11.4
2.3
8.6
41
9
9.8
13
15
8.6
163
179
+0.59
3,167
Run #5
1,729
0.3
1,056
89
45
247
1,322
1,105
2.35
49
2,196
0.8
932
238
2.55
6.42
11.1
2.5
8.3
50
9
9.1
13
16
8.7
161
171
+0.59
3,299
Test 3 Average
Runs 1, 2, & 3
1,587
0.1
1,098
90
60
0
973
463
2.33
52
2,172
0.7
1,041
251
2.35
7.55
11.0
3.5
8
54
12
8.8
15
16
8.6
145
173
+0.54
3,361
*Combustion gas retention time and excess air levels calculated by material balance
efficiency (ORE) for 2,3,7,8-TCDD and the
particulate emission rate (Table 5).
Bromine emission control was monitored
during Test 3, and the incinerator
demonstrated exceptional performance in
destroying organic waste materials. This
ability has now been demonstrated on both
solid and liquid waste materials. In fact,
the incinerator's performance is actually
better than that reported since the actual
emissions were lower than what is
measurable by current sampling and
analytical technology. No 2,3,7,8-TCDD was
detected in the stack, using
state-of-the-art high resolution mass
spectrometry. The lowest ORE for
2,3,7,8-TCDD was 99.999973%; the best DRE
measured during the trial burn occurred in
Test 2, Run 4, which had the greatest
analytical sensitivity (Table 6). The DRE
for this run was 99.9999951, which is
fifteen times better than State and Federal
requirements. Although the spin vane
probably failed prior to Test 2, Run 2, the
results meet all permit requirements.
Some problems were encountered with the
measured levels of particulates in the
313
-------
TABLE 5. TCDD AND PARTICIPATE MATTER EMISSION SUMMARY
Parameter
Permit Test 2 Test 3
Limit Run #2 Run #3 Run #4 Run #5 Combined
Waste Feed Characteristics
liquid waste feed
total flow, Ib/hr
2,3,7,8-TCDD cone., ppm
2,3,7,8-TCDD feed rate, g/hr
solid waste feed
total flow, Ib/hr
2,3,7,8-TCDD cone., ppb
2,3,7,8-TCDD feed rate, g/hr
Performance Standards - Stack
ORE for 2,3,7,8-TCDD
emission rate, rag/day
ORE, %
Bromine emissions
emission rate (kg/hr)
removal efficiency (%)
Participate matter
emission rate (mg/Nm3 7% Og)
400
27.2
2,000
233
249
26.3
1,158
101
0.05
236
357
38.3
1,363
382
0.24
234
264
28.0
2,068
1010
0.95
247
225
25.2
1,322
770
0.46
0.169 0.127 0.031 0.064
99.9999 99.99997 99.99998 99.99999 99.99998
180
134.3
147.3
145.6
201.5
973
0.12
99.0
stack, with one run of four being somewhat
in excess of the RCRA standard of 180
mg/m3 (@ 7% 02). This was subsequently
attributed to a buildup of submicron-sized
particles in the mass transfer scrubber,
the last element in the air pollution
control train. (A 50 u cartridge filter
was installed on the water recycle system
for the mass transfer scrubber, and
subsequent particulate testing was
conducted in the stack during the field
demonstration with results well below the
RCRA particulate limits.) The CO emission
values are equivalent to those from the
best available incineration technologies
and are indicative of very complete
combustion.
The results of Test 3 were also
satisfactory in that no bromine or chlorine
was detected in the stack gas.
FIELD DEMONSTRATION
A number of activities remain to be
completed as of this writing (1-86),
including the field demonstration and Test
1 of the trial burn.
The objective of the field
demonstration is to determine the rates at
which various types of dioxin-contaminated
liquids and solids can be fed into the
system and decontaminated. In addition to
evaluating the capability of the Mobile
Incineration System, the demonstration will
result in the cleanup of the majority of
dioxin-contaminated material in
southwestern Missouri. The specific
materials to be incinerated are shown in
Table 7, which includes the quantity and
estimated dioxin concentration for each
material. As of January 10, 1986, more
than 1,700,000 pounds of solids and 140,000
pounds of liquids have been incinerated
FINAL REPORTS
The final report will evaluate the
economic, technical, and institutional
viability of mobile incineration, and will
include a feasibility analysis of a larger,
"transportable" incinerator. Preliminary
analyses are presented here based on
results to date.
Economic Factors
The economics of on-site cleanup are
complex and do not lend themselves to
simple cost estimating approaches. The
concept of a mobile incineration system
comes with inherent constraints that must
be observed during the design, engineering,
314
-------
TABLE 6. RESULTS OF MISSOURI TRIAL BURN
Parameter
Permit Test Number
Limit 2-2a 2-3 2-4 2-5
3-1,2,
Kiln
temp., °C/op
SCC
temp., °C/OF
02, vol. *
CO, vol. ppm
retention time, sec
1400/1900 900
2050/2400
4
100
HC1 removal efficiency, % 99
ORE, %
2,3,7,8-TCDDC 99.9999
Participate Emission Rate 180
mg/Nm3, 1% Og
1190
7.7
4.1
2.5
NA
870
1200
7.0
2.0
3.2
NA
910
1200
6.4
2.3
2.5
NA
940
1200
6.4
2.5
2.6
NA
860
1190
7.6
3.5
NC
99.0
99.99997 99.99998 99.99999 99.99998
134 147 146 202
a Test 2 liquid feed - TCDD-contaminated trichlorophenol still bottoms
solid feed - TCDD/trichlorophenol-contaminated soil
b Test 3 liquid feed - fuel oil
solid feed - TCDD and brominated naphthalene-contaminated lagoon sludge
combined results for three 0.5 hour runs
c 2,3,7,8-TCDD = 2,3,7,8-tetrachlorodibenzo-p-dioxin
NC - not calculated
NA - not analyzed
and operation of the system. In order to
be mobile, maximum size and weight
limitations are imposed on the equipment
design. Thus, the MIS is denied the
"economies of scale," in that an equivalent
crew size could be used to operate a
stationary (or transportable) system that
has a much larger throughput. On the other
hand, one crew could potentially operate
two or more mobile units on the same site.
Since labor is a substantial portion of the
operating cost, increasing the capacity (by
increasing either the size or the number of
systems, or by modifications to the design)
significantly reduces the unit cost of
waste material destroyed.
Other factors that affect the economics
of the system are (1) heat and moisture
content of the material being processed,
(2) the operating factor, or percentage of
time that the system is able to process
material once it has been set up, checked
out, and put on stream, and (3) setup costs
and the duration of a given operation.
A detailed procedure for estimating the
unit costs of a mobile incinerator or any
other complex on-site treatment system will
be presented as part of the final report.
This procedure can then be used to compare
candidate technologies for Superfund use,
thus assuring that all key elements are
included in the comparative estimates.
Technical Factors
Design and operating decisions can
315
-------
TABLE 7. MATERIAL TO BE INCINERATED DURING FIELD DEMONSTRATION
Material
Penney Farm
MDB Soil
mixed solvents and water
chemical solids and soils
drum remnants and trash
Verona
hexane/i sopropanol
methanol
extracted still bottoms
activated carbon
decontamination solvents
sodium sulfate salt cake**
miscellaneous trash
Neosho
spill area soil
bunker soil/residue
tank asphaltic material
Erwin Farm
contaminated soil
loose soil
contaminated liquid
contaminated trash
Rusha Farm
spill area soil
contaminated liquid
contaminated trash
Talley Farm
spill area soil
contaminated liquid
contaminated trash
Eastern Missouri
Times Beach soil sample
Piazza Road soil sample
Estimated Quantity
210 cu yd*
2590 gal
31,150 Ib
84 85-gal overpack drums
10,000 gal
5,000 gal
5,000 gal
5,000 Ib
1,000 gal
23 cu yd
84 55-gal drums
282 cu yd
15 drums
75 gal
1264 drums
160,000 Ib
18 drums
16 drums
31.5 cu yd
1 drum
2 drums
176.5 cu yd
6 drums
16 drums
3 cu yd
3 cu yd
2,3.7,8-TCDO Cone.
500 ppb
Low
1 ppb - 2 ppm
Unknown
0.2 ppm
ppt
0.2 ppm
Unknown
Unknown
1 ppb
Unknown
60 ppb
2 ppm
2 ppm
8 ppb
8 ppb
Unknown
Unknown
Unknown
Unknown
Unknown
6 ppb
Unknown
Unknown
500 ppb
1600 ppb
*1 cu yd - 1 ton
**Will not be incinerated.
316
-------
significantly affect reliability, operating
factors and, therefore, costs. The
shakedown for the trial burn in Missouri,
as noted above, was conducted during
January through March, 1985. These
experiences indicate that careful planning,
conservative designs, a full complement of
spare parts, and adequate redundancy are
necessary for successful field operations.
As a temporary facility, the MIS
installation did not incorporate many
features found in a permanent facility,
such as a high-quality access road or
"line" power. Lack of such features
retarded progress during adverse weather
conditions. Each site will have its own
unique conditions that must be accommodated
to reduce system downtime. The cost for
specific site and equipment improvements
must be weighed against system downtime
costs. This trade-off also applies to the
number of spare parts and the quantity of
backup or standby equipment that is
available.
In general, however, the technology
elements for on-site cleanup by
incineration are currently available and in
common industrial use. There are no exotic
or high technology elements in the MIS.
The results of the field demonstration will
include suggestions or design improvements
that could be made by future, private
sector incinerator owners.
Institutional Factors
In order to use mobile incineration
systems, a number of permitting issues must
be addressed. Many of these issues
surfaced during the initial permitting of
the unit in New Jersey, although fewer
problems arose during the permitting for
the Missouri operation. Some of the
problems encountered were:
(1) Changes in personnel involved in
the permit application review at both
the State and Federal levels.
(2) Changes in the regulations between
the time the permit applications were
initially submitted and the time the
permits were issued.
(3) The interdependence of compliance
requirements, e.g., in order to comply
with the National Environmental Policy
Act (NEPA), the applicant must first
comply with the National Historic
Preservation Act, the Endangered
Species Act, and the Fish and Wildlife
Coordination Act.
These problems required considerable
documentation and correspondence with
various government organizations, and
resulted in long delays in granting permits.
There appears to be room for
improvement in the permitting process.
Toward that end, EPA is moving aggressively
to facilitate and simplify the permitting
process for mobile treatment systems.
Comparison with the Transportable
Incinerator
Although the EPA Mobile Incineration
System has clearly demonstrated the" ability
to destroy liquid and solid hazardous and
toxic organic wastes, its relatively low
capacity severely limits its use for the
massive task of cleaning up the myriad
abandoned hazardous waste sites already
discovered. Since the capacity of a single
mobile incineration system cannot be
significantly increased without the loss of
mobility, development of a modular,
transportable incineration system appears
to be a logical alternative for increasing
capacity. Such a system would be
transported to a site, assembled and
operated on-site, and then dismantled and
transported to another site. Accordingly,
a project was undertaken in 1984 to examine
the technical, administrative, and economic
feasibility of the use of modular,
transportable incineration systems for the
destruction of toxic organic wastes at
Superfund sites in the US. The operating
premise of such a system is that it will:
o have a capacity of 5-10 times that
of the EPA Mobile Incineration System
o be operated on a very large
Superfund site for 1-3 years, and
o be constructed from currently
available system components.
A rotary kiln with a secondary
combustion chamber appears to be the most
suitable type of incineration equipment for
handling the broad spectrum of hazardous
materials at Superfund sites. A 75 MM
Btu/hr system is presently estimated to be
the largest incineration system that could
317
-------
be transported without extensive field
fabrication.
Two key variables that affect the
system economics are primary and secondary
combustion chamber temperatures and the
types of wastes. A higher temperature
operation requires more auxiliary fuel than
a lower temperature, especially when little
fuel value waste is available as would be
the case at most Superfund sites.
Therefore, if PCBs or dioxins are not
present or are handled separately,
significant operating economies could ensue
by operating at lower temperatures.
The estimated cost for incinerating
waste materials with a modular
transportable system ranges from $52/ton
for contaminated soil at the lower set of
temperature conditions up to $561/ton for
industrial solid wastes at the set of high
temperature conditions. These estimates
Include capital and operating costs. Waste
excavation, transportation, and other
non-direct costs are not included.
FUTURE USE OF THE MIS
Further use of the EPA Mobile
Incineration System, after the field
demonstration at Denney Farm, will be at
the direction of the EPA Office of Solid
Waste and Emergency Response.
The intention of future operations with
the MIS is more to encourage
commercialization of on-site cleanup
technologies than it is to use the system
consistently for cleanup activities. As a
result of EPA experiences and operating
information, the private sector will likely
build improved, more reliable, larger
capacity, lower cost systems of at least
equivalent performance for use in routine
cleanup operations.
REFERENCES
Brugger, J. E., Yezzi, J. J., Jr.,
Wilder, I., et al. The EPA-ORD Mobile
Incineration System: Present Status.
In: Proceedings of the 1982 Hazardous
Materials Spills Conference, Milwaukee,
WI, April 19-22, 1982, pp. 116-126.
Yezzi, J. J., Jr., Brugger, J. E.,
Wilder, I., et al. The EPA-ORD Mobile
Incineration System. In: Proceedings
of the 1982 National Waste Processing
Conference, ASME, New York, NY,
May 2-5, 1982, pp. 199-212.
3. Yezzi, J. J., Jr., Brugger, J. E.,
Wilder, I., et al. The EPA-ORD Mobile
Incineration Trial Burn. In:
Proceedings of the 1984 Hazardous
Materials Spills Conference, Nashville,
TN, April 9-12, 1984.
4. Yezzi, J. J., Jr., Brugger, J. E.,
Wilder, I., et al. Results of the
Initial Trial Burn of the EPA-ORD
Mobile Incineration System. In:
Proceedings of the 1984 National Waste
Processing Conference, ASME, Orlando,
FL, June 4-7, 1984.
318
-------
REVIEW OF ALTERNATIVE TREATMENT PROCESSES
FOR METAL-BEARING HAZARDOUS WASTE STREAMS
Douglas W. Grosse
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Environmental Protection Agency (EPA) has initiated measures to incrementally
ban the land disposal of hazardous wastes generated in the United States. Wastes con-
taining toxic metals and cynaide complexes have been selected as a group to be restrict-
ed. Many of these wastes have been identified as significant sources contributing to
groundwater contamination when landfilled or land treated. EPA has been directed to
identify and quantify applicable alternative treatment technologies which have the poten-
tial to treat, reduce or render nonhazardous these wastes. A review of existing alter-
native treatment processes will be presented.
INTRODUCTION
An estimated 7.9 billion gallons of
metal bearing wastes, 4.7 billion gallons
of cyanide wastes and 0.8 billion gallons
of reactive wastes containing cyanides
were generated in the United States in
1983 (1). Currently 85% of all metal and
cyanide bearing wastes that are subjected
to land disposal are aqueous (_< 1% total
solids). Due to the high generation rate
associated with this category, a large
capacity of waste treatment processing
will be required. Technologies to be
considered as alternatives to land dis-
posal practices must demonstrate an
acceptable level of treatment before
these processes will be permitted for
operation. Treatment standards for many
waste codes are being established through
case studies, demonstration projects and
pilot studies.
Discussion of how various treatment
processes can achieve acceptable treatment
objectives can be generalized by the ap-
proach presented in Figure 1. Although
this simplification of waste treatment
options does not account for all metal
bearing wastes, a majority of these wastes
wi 11 be treated by some aspect of
the prescribed waste treatment strategy.
Waste Description
Hazardous wastes which can be cate-
gorized as metal and cyanide bearing wastes
are included in each of the five source des-
ignations for hazardous wastes as designat-
ed by EPA (2). The source categories
are:
D - wastes which are hazardous because
they exhibit a particular hazard-
ous characteristic, such as
toxicity;
F - wastes from non-specific sources;
K - wastes from specific sources;
P - acutely hazardous constituents and
U - toxic constituents
The majority of metal/cyanide wastes
generated in this country are characteris-
tic toxic metal wastes (D) and wastes that
result from electroplating and metal treat-
ing processes (F). Table 1 illustrates
the generated waste distribution for both
metal and cyanide wastes.
Other hazardous wastes containing met-
als and/or cyanide constituents are gener-
ated from specific industrial sources
(RCRA Code K); such as inorganic chemicals
and pigments (KQQ2 thru 8), petroleum
319
-------
AQUEOUS
METAL
WASTE
\>
DESTRUCTION
TREATMENT
I
CYANIDE
DESTRUCTION
i.e. ALKALINE
CHLORINATION
POLISHING
ADSORPTION
ION-EXCHANGE,
ACTIVATED
ALUMINA
CARBON, R/0
PRECIPITATION
LIME, SULFIDE,
SULFATE, CHROME
REDUCTION, ETC.
JZ
FILTRATION
MIXED-MEDIA
CARBON, SAND,
ULTRAFILTRATION
SLUDGE
DEWATERING
STABILIZATION
SOLIDIFICATION
RECOVERY
SPENT
RINSE
BATH
ELECTROLYTIC
ION-EXCHANGE
REVERSE
OSMOSIS
r
DISCHARGE
Figure 1. A General Treatment Approach for Aqueous Metal/
Cyanide Bearing Waste Streams
TABLE 1. GENERATED WASTE DISTRIBUTION
Metals
Hazardous
Waste
Group
D Wastes
F Wastes
K Wastes
P Wastes
U Wastes
Total
Annual
Volume
(million
gals.)
3,685
3,920
219
28.3
5.2
7,858
% of
Total
Metals
46.9
49.9
2.8
.36
.07
No. of
Generators
3,860
2,091
402
405
365
6,586
Annual
Vol ume
(mil lion
gals.)
Est. 750
3,920
572
226
3.92
4,723
Cyanides
% of
all
CN
,_
83
12.1
4.8
.08
•
No. of
Generators
Not available
2,091
10
sno
306
2,781
Source: Reference 1
320
-------
refining (K052), iron and steel (K061,
KQ62), secondary lead (K069, K100), and ink
formulations (K086). Salts of the various
metal ions and cyanide complexes are listed
in Part 261.33 (e) and (f) as discrete
chemical constituents of discarded com-
mercial chemical products, off-specifica-
tion species, container residues and
spill residues of acutely hazardous (P)
and toxic (U) wastes, respectively. Metal
ions are prevalent in many forms based upon
the type of industrial processes and uses
employed (2).
A1ternative Treatment Processes
Due to the wealth of information
found in the literature describing pretreat-
ment practices associated with specific
waste discharges, the following discussion
will be limited to two basic waste manage-
ment approaches:
0 waste treatment processes
0 waste reduction and recovery
Processes will be listed and synopsized for
each main category including a description
of their capabilities, advantages and limi-
tations encountered and significant results
lending relevance to standards and regula-
tions. The processes listed in this review
are to have potential for use to treat the
listed wastes. Regulatory and permitting
requirements for installation and operation
of the technologies presented may vary
among local, state and federal authorities.
Waste Treatment Processes
Traditional approaches used to treat
aqueous metal bearing wastes other than
land disposal include: precipitation,
reduction, stabilization/solidification,
cyanide destruction and biological treat-
ment. Some of these approaches will be
discussed here.
Precipitation
Precipitation is the most preferred
treatment process employed to remove toxic
heavy metals from electroplating waste-
waters. This method is used by approxi-
mately 75% of the electroplating facilities
treating aqueous metal bearing wastes. The
bulk of electroplating wastewater treatment
sludges are -usually in the form of metal
hydroxides (4). Precipitation techniques
currently being practiced on spent process
waters include hydroxide, lime and/or
sulfide treatment. Table 2 lists
generically and specifically the
types of wastes treated by sulfide
and lime precipitation, respectively.
TABLE 2. LISTING OF SPECIFIC
HAZARDOUS WASTES
AMENABLE TO PRECIPITATION
Waste Hazardous Wastes
Category Amenable to Precipitation
1.
Source
EP toxic:
(DOXX)
Sulfide
0002
0006
0008
0009
0011
Lime
0002
0005
D006
D007
2. Non-specific F006
industrial: F007
(FOXX) F008
F006
3.
4.
Specific
industrial :
(KXXX)
Toxic (UXXX)
and Hazardous
K021
K044
K046
K059
K071
K100
K106
(PXXX) Constituents:
K008
K061
K062
Salts of:
Mercury
Lead
Thallium
Arsenic
Cadmium
Silver
Antimony
Vanadium
Source: Reference 1
(1) Sulfide Precipitation - Sulfide
precipitation is a process that produces
the precipitation of a metal ion as a metal
sulfide (MS) through the reaction of the
metal ion (M+2) and a sulfide ion (S~2):
M+2 + s -2 __>
MS
(1)
Most heavy metals associated with electro-
plating wastes will form relatively stable
321
-------
metal Ions. Exceptions to this rule include
the trlvalent chromic and ferric ions. The
two processes currently employed to precip-
itate metals as sulfides are soluble sulfide
precipitation (SSP) and insoluble sulfide
precipitation (ISP).
SSP processes use water-soluble sulfide
compounds, such as, NaHS which yield rela-
tively higher concentrations of dissolved
sulfide than the ISP process. As a result,
rapid precipitation of dissolved metals
occurs. A disadvantage of the process is
that under certain pH conditions (pH <9),
HgS can evolve from dissolved sulfide ions.
The ISP process mixes the wastewater
with a slurry of slightly soluble FeS which
dissociates to satisfy its solubility
product. This has the effect of control-
ling the level of dissolved sulfide at
concentrations low enough to eliminate any
detectable emission of HgS. One advantage
of the ISP process is the ability of com-
bined sulfide and ferrous ions to reduce
hexavalent chromium to its trivalent state.
This eliminates the need to segregate and
treat chromium wastes separately. A gener-
alized treatment scheme for the ISP process
is illustrated by Figure 2 (5).
filtration. Effluents and residual waste-
waters containing excess hydroxide may
require neutralization before discharge.
Sludges precipitated from treatment may
require fixation or encapsulation to ensure
post disposal non-1eachability. Because
the optimum pH for hydroxide precipitation
is different for each metal ion, treatment
of mixed-metal aqueous wastes may require
some adjustment,
Disadvantages encountered when employ-
ing lime precipitation are: limitation of
metals removal due to solubility constraints
of the metal hydroxides for variable waste-
water matrices; interferences associated
with complexing agents, e.g., cyanide,
ethylene diamine tetraacetic acid (EDTA)
etc., and difficulty encountered when
stabilizing metal hydroxide sludges.
{3} Magnesium Oxide Precipitation -
One precipitation agent which has shown
promise for removal of heavy metal ions is
magnesium oxide. To improve technology for
treating process waters Schiller, et a!., of
the U.S. Bureau of Mines have shown in a
research study that magnesium oxide (MgO)
has many advantages over lime or caustic
soda for precipitating metal hydroxides. (6)
Neuiratization
Clarification
Figure 2. Insoluble Sulfide Process
(2) Lime Precipitation - Lime precipita-
tion is a well-established process, perhaps
the most commonly used process for treat-
ment of aqueous metal bearing wastes. Lime
(in slurry or solution) or caustic soda is
added to the waste as a source of hydroxide
1on raising the pH to a suitable level for
optimum precipitation of the metal hydrox-
ides. (3) The metal hydroxide precipitate
forms a floe which is afterwards removed
from the water by gravity settling or
Due to the decreased solubility of the
magnesium complex, lower sludge volumes are
produced as the hydroxide sludge becomes
more compacted. The MgO-metal hydroxide
sludge also cements together upon standing,
thus hindering resuspension of the metal
ions. One factor contributing to this
phenomenon is the positive surface charge
on MgO complementing the negative surface
charge on most heavy metal hydroxides,
Hence, water is expelled from the spaces
322
-------
between .the particles yielding a more
condensed solid. Also, MgO-hydroxide sludge
is more easily dewatered.
Reduction
Chemical reduction has gained accept-
ance in industry for reducing complexed
metals, such as nickel, copper, hexavalent
chromium waste, soluble lead, silver,
metal-containing cyanide and mercury. Chem-
ical reduction is most effective when the
aqueous metal wastes are relatively devoid
of organic compounds. Chromium is typically
reduced from the hexavalent to trivalent
form by addition of sodium sulfite salts,
sulfur dioxide and sulfuric acid. Iron,
aluminum and zinc have shown potential
for reducing chromium wastes. Sodium
borohydride (NaBH/j) has recently shown
promise for reducing and removing soluble
lead, including organo-lead salts, from
tetra-alkyl lead manufacturing wastes (7).
Typical hazardous wastes treated by
chromate reduction are listed in Table 3.
TABLE 3. HAZARDOUS WASTES TREATED BY
CHROMATE REDUCTION
Waste Code Description
Soluble U032 Calcium chromate
D007 Chromium (hexa-
valent with
chromic acid)
F019 Sludges from
chemical conversion
coatings
Insoluble KQ02 Production sludges
from: chrome yellow
K003 molybdate orange
K004 zinc yellow
K005 chrome green oxide
K006 chrome green oxide
K008 oven residues from chrome
green oxide
K086 pigments & inks
F006 insoluble chromates
Several aqueous based processes which are
commercially practiced are listed below: (1)
° Hydrosulfide reduction - a sulfide ion
is added to the chromate waste, with pH
adjustment to 7. After allowing, a few
hours for reaction, the pH is adjusted
to the alkaline side where chromic
hydroxide is precipitated and filtered.
° Sulfur Dioxide Reduction - Gaseous sul-
fur dioxide is added under pressure to
the chromate containing solution, where-
by, the pH is lowered to an acidic
level. Chromic hydroxide is precip-
itated after alkali addition and filter-
ed.
° Ferrous Ion Precipitation - Waste pic-
kle liquor is frequently used to ini-
tiate a rapid reaction with ferrous
ions under acidic conditions. Chromic
hydroxide is formed with iron hydroxides
and is precipitated under alkaline
conditions. The resultant precipitate
is filtered for removal.
Adsorption
Adsorption is a separation process
which concentrates molecules at a contact-
ing surface without undergoing a chemical
reaction. In general, adsorption systems
have seen limited application in treating
higher strength metal bearing hazardous
wastes. Although adsorption systems have
traditionally been used to polish effluents
from aqueous waste streams containing organ-
ic compounds, recent studies have shown
potential for use in treating and polishing
aqueous metal bearing wastes.
Laboratory work has been conducted to
demonstrate the feasibility of using acti-
vated carbon to treat cadmium (?.) - contain-
ing wastes (D006) (8). Two schemes were pro-
posed for the treatment of this waste: (1)
completely mixed flow reactor (CMFR) and
(2) column reactor (CR). In the CMFR,
powdered activated carbon (PAC) was added
to effect treatment and in the CR activated
carbon beads containing 70% PAC to 30%
polyvinyl alcohol and glutaraldehyde were
utilized. It was demonstrated that activat-
ed carbon appeared to be economically
comparble to ion-exchange,and alkaline
neutralization/precipitation methods. One
disadvantage encountered was that removal
of cadmium is inhibited by the presence of
cyanide which can form relatively non-
adsorbable cadmium (2) - cyanide complexes.
The process is generally pH sensitive.
Another study by Rosenblum and Clifford
(9) has demonstrated the effective removal
of Arsenic 5 (0004) by adsorption onto an
323
-------
activated alumina bed. Activated alumina
has in the past proven to be successful in
removing fluoride from groundwater. Equili-
brium isotherms were obtained for both
batch and mini column studies. It was shown
that the optimum pH for adsorption of
Arsenic 5 onto activated alumina was at pH
6. One disadvantage encountered was that
Arsenic adsorption was significantly re-
duced in the presence of competing fluoride
and sulfide anions.
Mayenkar and Lagvankar, in another
study, (10) presented findings on the
removal of chelated nickel and copper ions
from boiler tube chemical cleaning wastes
onto iron filings. Conventional physical/
chemical processes were not able to remove
nickel to a desired level (£1.0 mg/L), due
in part to the presence of oxidizing and
chelating agents (EDTA). Investigative
studies lead to the development of a pro-
cedure using iron oxides as an adsorbent.
Nickel (>60%), copper (100%) and iron
(85%) removals were observed. Iron fil-
ings adsorption of aqueous wastes contain-
ing complexed metals appears to be a very
promising alternative to conventional treat-
ment when considering the cost and restric-
tions of sludge disposal.
Cyanide Destruction
One of the major sources of generated
cyanide waste is the electroplating indus-
try. Cyanide baths are used to hold metal-
lic ions such as zinc and cadmium in solu-
tion. When the "drag-out" of plating
solution containing cyanide ions contami-
nates rinsing baths, a wastewater is gener-
ated which must be treated.
Cyanide appears in waste residuals,
in two forms as an "ionized cyanide"
in the form of sodium cyanide (NaCN) and
hydrocyanic acid (HCN) or as a complex with
heavy metals, e.g. ferro-cyanide, nickel
cyanide or copper cyanide. Other cyanide
species come in the form of inorganic com-
pounds such as thiocyanates and cyanates
which are easily removed from waste streams
by using an assortment of treatment process-
es. Various methods for treating cyanide
wastes have been employed, with cyanide
destruction via alkaline chlorination being
used most frequently. Unit processes report-
ed to successfully treat cyanide and cyanide
complexes include incineration, alkaline
hydrolysis of cyanohalides, pyrolysis of
copper cyanide, peroxide oxidation
(KastoneW) and iron cyanide pyrolysis (1).
Although natural degradation including oxi-
dation, biodegradation and photo-decomposi-
tion will reduce cyanide concentrations in
aqueous wastes given enough time, a recent
study has reported successes with ozonation,
ion-exchange, activated carbon, reverse
osmosis, electrodialysis, electrolytic hy-
drolysis and titanium oxide/UV.H Two of
the most preferred treatment options are
listed below.
(1) Alkaline Chlorination - As the most
prominent, widely used treatment option
employed to date, alkaline chlorination may
be accomplished by direct addition of
sodium hypochlorite or by addition of
chlorine gas with sodium hydroxide. Hypo-
chlorite, perhaps the most widely used
oxidizing agent, reacts with cyanide in
the following manner:
(1) ' CN" H20 + OCL-- —> C1-CN+20H" (2)
conversion of CN" to dissolved
cyanogen chloride
(2) CL - CN + 20H- —> GOT (3)
+ CL" + H20 hydrolysis of
cyanogen chloride
(3) 20CN- + 30CL" --> 2C02 + N2 + (4)
3CL" + OH conversion to C0? and
nitrogen
If thiocyanates are present, the demand
for chlorine is preferential, therefore,
requiring more oxidant. This process
cannot effectively treat simple or com-
plex copper cyanides, zinc cyanide and iron
cyanides due to their low solubilities.
The following cyanide bearing wastes
are amenable to alkaline chlorination:
RCRA
Waste
Code
P106
P098
P021
P013
P083
P074
Waste
Sodium cyanide
Potassium cyanide
Calcium cyanide
Barium" cyanide (the treated waste
requires subsequent treatment by
the sulfate precipitation process)
Hydrogen cyanide (in aqueous solu-
tion)
Nickel cyanide
324
-------
P099 Potassium silver cyanide (sludges
generated may be reprocessed to
recover silver)
P104 Silver cyanide (sludges generated
are reprocessed for silver
recovery)
K060 Lime sludge from coking operations
F006 - Cyanide bearing plating solutions
F012 and sludges F012 (some treated
sludges may be reprocessed to
recovery metals; sludges of no
value may require subsequent
treatment by the encapsulation
process
P030 Miscellaneous cyanides
(2) Incineration - Incineration has been
used to destroy gaseous hydrogen cyanide,
gaseous cyanogen, organonitriles and organic
isocyanates at elevated temperatures.
Several types of incinerators such as
liquid injection, rotary kiln and fluidized
bed are capable of treating these wastes.
Isocyanates are converted to C02, N2, and
HoO leaving little if any toxic residuals.
Trie following cyanide wastes may be treated
by incineration, although provision must be
made to control stack gas emissions.
RCRA
Uaste
Code Haste
P083 Hydrogen cyanide (as gas)
U003 Acetonitrile
U152 Methyl acrylonitrile
U149 Propane nitrile
U009 Acrylonitrile
P101 Propionitrile
P064 Methyl isocyanate
P069 2-methylacetonitrile
P027 3-Chloropropanenitrile
P069 2-hydroxy-2-methylpropanenitrile
P031 Cyanogen
K011 - 14 Uaste streams for acrylonitrile
and acetonitrile production
K087 Tar sludges from coking operations
U223 Toluene diisocyanate
UASTE REDUCTION AND RECOVERY
Interest in establishing centralized
treatment and recovery systems for treating
heavy metal wastes in large metropolitan
areas has spurred many studies to demons-
trate waste reduction and recovery technol-
ogies. With the high cost of raw materials
and pending land disposal restrictions
reclamation of rau materials fponr
waste streams can be the treatment option
of choice. EPA studies indicate that a
large part of the operational costs would
be subsidized by the sale of the recovered
metals (12).
Economic feasibility of metals recovery
operations are primarily dependent upon
three factors:
1. Metal type and concentration;
?,. Availability of significant quan-
tities of material within close
proximity and
3. Economical (profitable) recovery
process.
Justification for recovery processes is not
exclusive to raw material savings. Reduced
operating costs for "end-of-pipe" pollution
control can be a significant factor, par-
ticularly when considering that these costs
may rise in response to the HSUA. Four
reduction and recovery processes will be
discussed here: evaporation processes,
ion-exchange systems, membrane technology
and electrolytic deposition.
Evaporation
Evaporators have traditionally been
used to concentrate rinse water and recycle
"drag-out" back to the plating bath. This
in-plant separation process represents a
sound waste minimization practice. Figure 3
illustrates two types of systems which have
been employed: 1) a closed-loop system
which totally eliminates a water discharge
from the plating process, and requires sub-
stantial make-up water to satisfy rinsing
requirements and 2) an open-loop system with
lower operating costs as a result of less
make-up water. However, with the open-loop
system a wastewater discharge is generated.
Ion-Exchange
Ion-exchange (I/E) and resin adsorp-
tion systems offer a versatile separation
process with wide application to the metal
finishing industry. I/E has all the advan-
tages of evaporative processes for recovery.
In addition, it can serve as a polishing
step after conventional wastewater treatment
and purification of liquids (acids, bases,
water, etc.). By definition, I/E is a
separation technology which removes various
ionic species from solution via interchang-
ing reversible ions between the solution
(aqueous waste) and the exchanger (e.g.
325
-------
PRODUCT FLOW
AND DRAG-OUT
STEAM OR HOT WATER
COOLING WATER
cr
TV---/-
U
PLATING
BATH
P *-
TV / n'
U|
RINSE //]
.-—.» EVAPOR
1
TO WASTE TREATMENT
CONCtHTRATID PLATING
CXUICAIS
DISTILLATE
I
CONCENTRATE
2—Open-Loop Evaporative Recovery of Chromium Plating Drag-Out
1—Ctoietf-Loop Evaporative Recovery System.
Figure 3. Evaporative Recovery Systems
resin). The I/E process can be manifested
as a resin system, membrane selection
process (Donnan Dialysis) or a liquid-
liquid extraction procedure. Exchangers
may be comprised of synthetic resins, in-
soluble salts, molecular sieves or even
liquid membranes. Liquid ion exchange,
where "organic polar molecules" are added
to complexed metal ions, appears to have
some promise (13).
Several studies have been conducted
to demonstrate specific application of
various resins for treating synthetic and
Industrial aqueous metal wastewaters.
Tare, et. al . (14) described the kinetics
for metal ion removal by chelating I/E
resins. The presence of strongly complex-
Ing organics (such as citrate, EDTA, and
tartrate) and inorganic ligands (CL, F,
P04, etc.) in dilute metals waste make
chemical separation processes (i.e. precip-
itation) not feasible. Chelating I/E
resins offer an alternative to treating
these complexed wastes.
In another study, Maeda (15) investi-
gated the influence of porosity for several
macroreticular chelating resins (RST) con-
taining triethylene-tetramine with differ-
ent porosities. Results indicated that
large pore radius RST's proved to be more
practical for removal and recovery of heavy
metal ions from an industrial metal bearing
waste.
Liquid ion-exchange is a relatively
new process which combines the concepts of
I/E and liquid-liquid extraction for the
recovery of metal ions from wastes.1° This
application has shown promise for treating
the more dilute metal wastewaters. Like
resin systems, the actual exchange takes
place when the solution of metal-ions, upon
contact with the extractant, produces an
unstable liquid-liquid dispersion.
Membrane Processes
Membrane processes typically used to
treat aqueous metal wastes are reverse
osmosis (RO) and electrodialysis. RO is a
pressure driven membrane process where
water molecules are forced through the
"microscopic pores" of a semi-permeable
membrane. The production rate of an RO
membrane is a function of dissolved solids
concentration, temperature, pressure, pH
and chemistry of the solution to be treated.
No commercially available membrane polymer
has demonstrated tolerance to all extreme
chemical factors, such as, pH, strong
oxidizing agents and aromatic hydrocarbons.
A study by Crampton (12) successfully
demonstrated the use of a cellulose acetate
membrane to concentrate the "drag-out" from
plating solutions. Both the concentrate
and purified water are recycled. However,
this type of membrane is pH sensitive and
is primarily used to treat Watt's nickel
326
-------
plating baths. Other drag-out recovery
applications successfully treated copper,
zinc and chromic acid baths.
In "An Update on Reverse Osmosis for
Metal Finishing" Cartwright defines various
membrane types and their applications (17).
Some applications of RO membranes to various
waste streams are listed below:
Nickel
Copper sulfate
Zinc sulfate
Brass cyanide
Copper cyanide
Chromium (VI)
Other
cellulose acetate
hollow-film polyamide,
cellulose triacetate and
sprial-wound-thin film
composite
spiral-wound-thin film
composite
polyamide cellulose,
triacetate hollow fiber
polyamide hollow fiber
spiral-wound-thin film
composite
thin film composite (TFC)
Table 4 presents RO test data for six
common plating bath rinses (17).
TABLE 4. RO TEST DATA FOR SIX COMMON
PLATING BATH RINSES
Plating
Bath
Watts Ni
Copper
cyanide
Zinz
cyan ice
Brass
cyanide
Bath
Temp.
°C/°F
60/140
60/140
27/80
27/80
Toxic
Contam-
inant
Ni2+
Cu+
ZN2+
CN-
ZN2+
Cu+
Percent
Rejec-
tion
99+
85
46
96
80
98 .
97
CN-
Decorative Cr
Hard Cr
43/110
55/130
90
92
Potential problems with RO systems
stem from membrane fouling due to suspended
precipitated solids coating the membrane
surface, precipitation of salts in the
concentrate, membrane deterioration due to
chemical attack and for higher strength
wastes, high levels of dissolved solids in
the permeate.
In electrodialysis systems an electri-
cal potential is applied across the mem-
branes to provide the driving force for ion
passage through the membranes. Membranes
used in this process are "thin sheets" of
the same polymeric network used to make I/E
resins. Unlike RO systems these membranes
are tolerant of most chemical environments.
Another variance exhibited by electrodialy-
sis over RO systems is the propensity to
separate water from the salt by selective
removal of the salt, rather than concentra-
ting the salt from solution.
Electrolytic Recovery
Electrolytic recovery is one of the
many technologies utilized to remove and
concentrate metals from process waste
streams. This process uses electricity to
pass a current through an aqueous metal
bearing solution between a cathode plate
and an insoluble anode. Positively charged
metallic ions cling to the negatively
charged cathodes leaving behind a metal
deposit which is strippable and recoverable.
Initially used to recover gold and silver,
its use in recovering less precious metals
has been limited. However, new interest
has been given to this promising recovery
technology.
Walters and Vilagliana of the Univer-
sity of Maryland have demonstrated the
electrolytic recovery of zinc from metal
finishing rinse waters (18). The deposition
of copper ions after contacting diluted
aqueous metals solution onto a packed
graphite bed was studied. Results of the
study indicated that zinc recovery from
plating bath rinse waters, which consisted
of a dilute zinc cyanide solution, was
achievable. A batch electrochemical reactor
with stainless steel electrodes was em-
ployed. The controlling factor in achiev-
ing high rates of zinc deposition appeared
to be agitation. The study cited that
mechanical mixing and nitrogen gas aeratiot.
were both effective. Nad was added to
maintain a minimal conductivity level.
Higher current densities elicited higher
deposition rates as opposed to lower current
densities. One noticeable disadvantage was
that corrosion could play a significant
limiting factor. Electrodes would have to
be replaced frequently.
327
-------
REFERENCES
1. Preliminary Draft: EPA Contract No.
68-03-3149, Versar, Inc. (1984).
2. 40 CFR Protection of Environment Part
261.0, 1984.
3. Preliminary Draft Report: EPA Contract
No. 68-03-3166 (1985).
4. EPA Draft Report: Environ - Character-
ization of Waste Streams Listed in 40
CFR Section 261 Waste Profiles, Vol. I,
(1985).
5. Summary Report: Control and Treatment
Technology for the Metal Finishing
Industry: Sulfide Precipitation,
EPA 625/8-80-003, (1980).
6. Schiller, J.E. and Khalafalla, S.E.,
"Magnesium Oxide for Improved Heavy
Metals Removal," Mining Engineering,
173, (1984).
7. Briefing: Technologies Applicable to
Hazardous Waste, Metcalf & Eddy,
Inc., Prepared for the U.S. EPA,
Section 2.18 (1985).
8. EPA-600/S2-83-061 Project Summary;
Activated Carbon Process for the
Treatment of Cadmium (11) - Containing
Wastewaters," Huang, C.P.
9. EPA-600/S2-83-107, Project Summary:
"The Equilibrium Arsenic Capacity of
Activated Alumina," Rosenblum &
Clifford.
10. Mayenkar, K. U. and Lagvankar, A.C.,
"Removal of Chelated Nickel from
Wastewaters," Proc. 39th Ind. Waste
Conf., Purdue University, 457 (1984).
11. Draft Report: Destruction of Cyanide
in Wastewaters: Review and Evaluation,
Weathington, C., Contract No. 68-03-
3037 and 68-03-3197, U.S. EPA (1985).
12. Crampton, P., "The Application of
Separation Processes in the Metal
Finishing Industry," EPA-600/2-81-028.
13. Eckert, et. al., "Research Planning
Task Group Study: Separation Tech-
nology," U.S. EPA, CR806819, University
of Illinois, Urbana-Champaign, (1982).
14. Tare, et. al., "Kinetics of Metal Re-
moval by Chelating Resin from a
Complex Synthetic Wastewater," Water,
Air and Soil Pollution, 22, 1984,
p. 429-439.
15. Maeda, H., "Studies of Selective Ad-
sorption Resins, XVIII. The Influence
of Porosity of Macroreticular Chelating
Resin on Adosrption of Heavy Metal
Ions in an Aqueous Solution," Journal
of Applied Polymer Science, Vol. 29,
(1984), pp 2281-2287.
16. Gallacher, L.V., "Liquid Ion Exchange
in Metal Recovery and Recycling,"
EPA-666/2-81-028.
17. Cartwright, P.S., "An Update on Re-
verse Osmosis for Metal Finishing,"
Plating and Surface Finishing, April
1984, p. 62-66.
18. Walters, R. and Vitagliano, D.,
"Electrolytic Recovery of Zinc from
Metal Finishing Rinse Waters." Proc.
of 16th Mid-Atlantic Ind. Waste Conf.,
1984.
328
-------
FIELD EVALUATION OF TREATMENT PROCESSES FOR
CORROSIVES AND METAL-BEARING WASTE STREAMS
Neville K. Chung, Mike A. Crawford, A William E. Shively
Metcalf X Eddy, Inc., Wakefield, MA 01880
ABSTRACT
The 1984 amendments to the Resource Conservation and Recovery Act (RCRA) require
that EPA ban land disposal of certain hazardous wastes, and establish levels or methods
of treatment. To implement the provisions of this legislation, EPA must determine
whether adequate treatment technologies exist, what wastes can be treated and how
effectively, what residuals and environmental discharges are produced, and what the
associated costs are. As part of the program to develop this information, field
evaluations of alternative technologies for treating or destroying wastes that have
been listed for priority action were conducted. This paper reports on field
evaluations of full scale processes as they are being applied at commercial hazardous
waste treatment facilities for treating corrosive and metal-bearing wastes.
INTRODUCTION
The RCRA Hazardous and Solid Waste
Amendments of 1984 mandate that a ban on
land disposal of hazardous wastes be
considered and that treatment standards
for banned wastes be established. As
set by the Amendments, the first waste
categories to be affected by the land
disposal ban are:
Liquid hazardous wastes,
including free liquids
associated with any solid or
sludge, containing cyanides;
Liquid hazardous wastes,
including free liquids
associated with any solid or
sludge, containing specified
concentrations of arsenic,
cadmium, hexavalent chromium,
lead, mercury, nickel, selenium
or thallium;
Hazardous wastes containing
halogenated organic compounds;
Dioxin-containing hazardous
wastes;
Solvent wastes.
In support of the efforts of the
Office of Solid Waste (OSW) to implement
the provisions of the RCRA land disposal
restrictions, the Hazardous Waste
Engineering Research Laboratory (HWERL)
has initiated several programs to
evaluate alternatives to land disposal
of hazardous wastes. A major component
of these programs has been the field
evaluation of existing, full-scale
technologies, with initial focus on the
waste categories designated for priority
action.
PURPOSE
The purpose of the field evaluation
program is to develop information
concerning existing processes being
applied to the treatment of hazardous
wastes, the effectiveness of those
processes, and the environmental impacts
in terms of residues and emissions.
Although many waste types, including
most of the priority wastes, were
included in the field evaluations, this
paper discusses primarily corrosive and
heavy metal wastes. The paper presents
general observations from 14 site visits
conducted at commercial treatment,
storage and/or disposal facilities
(TSDFs) and focuses in particular on
available results from testing at two
facilities.
329
-------
APPROACH
Several approaches were considered
appropriate for obtaining the desired
data on waste characteristics, treatment
process performance and treatment
residuals, to supplement existing
data. They include onsite pilot testing
of selected industrial process waste
streams using EPA-owned or other pilot
equipment; field evaluation of onsite
(I.e. generator owned) treatment
processes; and field evaluation of
offsite (i.e. commercial) hazardous
waste treatment facilities. The last
approach was selected for initial
testing activities because it was
considered to offer the best opportunity
to quickly identify, select and test
processes that are currently being
applied to treat wastes that are shipped
off-site for disposal, potentially to
land disposal.
The approach for.selecting and
testing treatment facilities consisted
of:
1. Screening commercial hazardous waste
treatment facilities through the use
of waste directories and telephone
contact to determine the types of
waste handled, management practices
employed, and willingness to
participate in the program.
?. Site visits to selected facilities
that are treating the priority
wastes of interest, and appear to be
wel1 operated.
3. Evaluation of the site visit data to
select sites for testing.
4. Performance of field sampling and
process monitoring, following
concurrence of EPA with the site
selection, site specific Test Plan,
and Quality Assurance Project Plan.
For each selected test site, sample
locations were selected to characterize
the waste streams and assess the
performance of unit operations. Most
operations were batch processes. For
this reason, grab samples were collected
at the beginning and completion of each
batch unit operation, and composited at
the end of each 24-hour period according
to the number of batches treated.
Effluent from semi-continuous processes
were composited to account for possible
variations in effluent quality over time
as a function of the treatment cycle.
Samples from tank trucks, drums, storage
tanks and reactor vessels were collected
by means of composite liquid waste
samplers (COUWASAS) In order to ensure
collection of representative samples.
The parameters selected for
analysis were based on information
gathered from the manifest accompanying
the waste, or the plant effluent permit
requirements. Volatile organic and
extractable organic fractions from the
contract laboratory programs hazardous
substance list were also analyzed.
RESULTS
fieneral Observations
The following observations from
visits to 14 commercial treatment,
storage and disposal facilities (TSDFs)
are characteristic of practices that
were encountered in the commercial
treatment industry, but in many cases
are applicable in general to the
handling and treatment of corrosive and
metal containing wastes.
1. Except for three specialized solvent
reclaimers, all of the facilities
visited accepted corrosive and
metal-bearing wastes for treatment
and disposal. Almost all facilities
place some restrictions on waste
characteristics that are accepted,
based on their ability to
effectively handle the waste with
the equipment and processes
employed. For example, some
facilities do not accept sludges
because they do not have sludge
dewatering/disposal capability; and
they themselves generate and ship
sludges offsite to other TSDFs for
ultimate disposal. Some facilities
have restrictions on mixtures of
acids, and on the maximum heavy
metal content of wastes. The
presence of chelating agents in
metal-containing wastes was a major
concern at each of the facilities
visited. As may be expected, most
will not accept dioxin, PCBs, and
wastes that are explosive,
radioactive or infectious.
330
-------
2. Incoming wastes are analyzed to
verify the composition and the
treatability of the wastes. A
number of facilities either refused
to accept, or minimized the quantity
of nitric acid accepted, due to its
aggressiveness. One facility based
acceptance of corrosives on the
total acidity or alkalinity content
as measured by a neutralization
index. A number of facilities spot
test for the presence of dictating
agents by performing settleability
tests on the incoming wastes.
3. Generally, the commercial TSDFs
allow waste to accumulate onsite
until sufficient volume accrues for
cost-effective treatment. For
facilities that accept a wide
variety of wastes, extensive tank
and/or drum storage capacity is
necessary to accommodate waste
segregation.
4. A large percentage of the pumps and
tankage at the commercial TSOFs
visited is used equipment which has
been retrofitted to fulfill process
needs. Operating and maintaining
this equipment often increases the
labor requirements in what is
already a predominantly manual,
materials handling industry.
5. In view of the relatively small
quantities and highly variable
nature of the wastes, batch
treatment is usually most
appropriate. The required treatment
must often be determined for each
batch by jar testing.
6. To the extent possible, strong waste
acids and caustics are used for
gross pH adjustment, e.g. to lower
pH for chrome reduction or cracking
of emulsions, or to elevate pH for
alkaline chlon*nation of cyanide
wastes. However, since these waste
acids and caustics are often
contaminated with heavy metals,
other inorganics and sometimes
organics, they are not suitable for
final treatment steps. Excess acids
and caustics that cannot be used in
the treatment process are used to
neutralize each other, with any
deficiency made up by virgin
reagents.
7. Concentrated wastes must usually be
diluted, - slowly in the case of
strong acids to avoid excessive
temperature rise due to heat of
dilution. High temperatures are
detrimental to equipment,
(especially fiberglass reinforced
plastic tanks that are commonly used
for corrosion resistance), and can
cause undesirable emissions.
8. Three facilities visited neutralized
waste acids and caustics in clay-
lined lagoons. This allowed more
rapid neutralization of corrosives
since the heat of reaction was
easily dissipated.
9. Lime precipitation followed by
either sedimentation or filtration
in a filter press was the most
common aqueous metals treatment
process observed. Sulfide
precipitation was used at three
facilities.
10. Two facilities incorporated
evaporation to treat chelated
wastes. Another facility treated
these wastes by lowering the pH to
weaken the metal-complex bond, and
adding ferrous sulfate or ferric
chloride to substitute iron for the
complexed metal ion.
11. The majority of TSOFs that treat
aqueous wastes have an effluent
holding tank at the end of the
treatment process. This affords the
plant the opportunity to confirm
that the effluent quality is within
the discharge limits before
releasing .it to the sewer.
Table 1 summarizes general
observations and unit processes
incorporated at the 11 facilities
visited that treat corrosive and metal-
containing wastes.
Plant A
Faci1ity Descriptlon. The unit
processes employed at Plant A are shown
in Figure 1. Cyanide-containing wastes,
331
-------
TABLE 1. PROCESS SUMNR OF TSBF* VISITS
GJ
CO
ro
Pl80t
I
2
3
4
5
Rest? I efforts On
lncc-lr>q Wflst«s!
Do not accept corrosive or
Mfal-bear Ing sludges.
Acids with pH less than 2 and
certain mtals In excess of
Col 1 fornlo waste I In Its ore not
accepted.
Avoid cheating agents and high
arananla concentrations.
Limit acceptance of nitric acid
and other mixed acids based on a
neutralization index; limit
concentrations.
Acceptance based on traatabl 1 Ity
analysis.
Corrosive Voter U$ag*
car TrestBtAt
Gross $« adjostiNKttj neutralization
of fields sod caustics In
polyethylene tanks,
NeutraMza waste acids and caustics
In a ct ay™ 11 nod lagoon.
Neutralize waste acids and
caustics In concrete equalization
tank. Some waste acid used In
emulsified oil treatment*
Gross pH adjustment; neutralization
of acids and caustics In continuous
flotf reactor.
Gross pH adjustment; neutralization
of acids and caustics In closed top
Precipitation
Process
Lie* precipitation
NA
line, sulflde, and carbonate
precipitation
L!«e and/or sulflde
precipitation
Lima and/or sulfldo
precipitation
Liquid-Solids
Separation Process
Gravity sedI*estBt|«i end
eefltrJfugatlen
Evaporation In surface Ippoundiwtt
Gravity sedlnentatlonf sand
filtration, vacuira filtration
Cavity sedlBentatlon,
continuous belt filter press
Gravity sedimentation, plate and
frame filter press, centrlfugatton
Ultlsate
0!«h«rge
Treated liquid to POTV; sludge
to approved landfill or ratals
reclaimer
Zero liquid discharge;
stabilized sludge to
landfill.
Treated liquid to POTVf; ! (de-
stabilized sludge to landfill.
Treated liquid to PQI>fj
stabilized sludge to landfill.
Treated, liquid to POTW; filter
cake to landfill.
Acceptance based on treatabtllty
analysis.
No restrictions on corrosive
or Betal-baarlng wastes.
Do not accept chelated wastes.
process tank.
Gross pH adjustment
Line precipitation
Gross pH adjuslmantj neutralization LIsne precipitation
of acids and caustics In a clay-lined
lagoon.
Gross pH adjustment! neutralization
of acids and caustics In reactor
vessels.
Line precipitation
Plate and frame filter press,
multi-media filtration, and
carbon adsorption
Gravity separation and evaporation
In surface Impoundment
Gravity separation, filter presses
Treated liquid to POTH; sludge
to approved landfill.
Zero liquid discharge;
stabilized sludge land
disposed.
Treated 11quId to
recalvlnq stream; sludge
to approved landf II I.
Acceptance based on treatablIIty
analysis* Do not accept
Gross pH adjustment; neutralization
of acids and caustics In steel
Lime and limited sulflde
precipitation
Gravity separation (plate and
frame press under construction)
Treated liquid to POTW;
thickened sludge to solar
to
11
corrosive or metal-bearing
sludges.
No restrictions on corrosive
or metal-bearing wastes.
or metal-bearing wastes.
tanks.
Gross fH adjustment; neutralization Lime, alum or waste
of acids and caustics In a alkaline precipitation
c E ay-i 1 ned 1 agoon.
Recycle waste sulfurlc acid If quality Lima precipitation
permits; neutralization of acids and
reactor vessels.
disposal.
In a surface Impoundment receiving stream;
stabilized sludge land
disposed.
All neutralized wastes are Zero liquid discharge;
stabilized for land disposal. stabilized sludgs to
landfill.
1 - Restrictions based on corroslvity and natal-bearing characteristics only.
NA - Not applicable
-------
Caustic •
Hypochlorite
Cyanide Bearing
Aqueous Waste
Dewatered Sludge
To Landfill
Aqueous Waste pH Reduction
Receiving Tanks Tanks
Non-Cyanide Bearing
Aqueous Waste
Treated
Effluent to
Sewer System
Lime Flocculation
Tank
Filter Press
Surge Tank
Plate and Frame
Filter Press
Virgin HCI
Effluent
Surge
Tank
Carbon
Adsorption
Tanks
pH Adjustment
Tank
(g) Sample Locations
FIGURE 1. WASTE TREATMENT FACILITY-PLANT A
which also contain heavy metals, are
treated by alkaline chlorination prior
to being combined with non-cyanide
metal-bearing wastes in the influent
aqueous waste receiving tanks. Cyanide
wastes containing greater than one
percent cyanide are diluted to avoid
excessive heat of reaction that would
damage fiberglass reinforced plastic
(FRP) tanks. The process is operated as
a single stage alkaline chlorination
process, converting cyanide to
cyanate. Cyanate is hydro!yzed to
carbon dioxide and nitrogen upon
addition of acid in the pH reduction
step of the aqueous waste treatment
system.
Unit processes for treatment of the
oxidized cyanide waste and noncyanide
aqueous wastes include storage,
blending, pH adjustment, chemical
reduction, lime precipitation of heavy
metals, filtration, neutralization and
carbon adsorption. Blended wastewater
from the influent equalization tanks is
pumped to one of two pH reduction
tanks. The pH of the waste is adjusted
to two with dilute (approximately 20
percent) mixed waste acid (sulfuric,
nitric, hydrochloric, or chromic acid)
and/or virgin hydrochloric acid. This
operation serves to: introduce the acid
wastes into the aqueous waste treatment
system for dilution and subsequent
neutralization and metal removal; lower
the pH for chrome reduction; and
reportedly to also break certain metal-
organic chelate complexes, facilitating
metals precipitation in the next unit
operation. If chromic acid is added to
the waste, or if chrome is already
present in the waste, then bisulfite is
added to reduce hexavalent chromium to
trivalent chromium. When sulfuric acid
is added to the waste, calcium sulfate
is formed, which enhances the downstream
filter press operation but also
increases the quantity of sludge
generated. After pH adjustment and
chrome reduction, which requires two to
three hours, the wastewater is pumped to
a steel reactor vessel to which lime has
been added. This raises the pH to
approximately 11, resulting in the
formation of insoluble metal hydroxides.
Following lime addition, the
wastewater (now a slurry of metal
precipitates in water) is transferred
from the precipitation tank to one of
three aqueous waste slurry tanks. The
333
-------
slurry is agitated in the surge tanks
and is pumped to a plate and frame
filter press for dewatering. This press
produces a sludge filter cake with a
solids content of about 45 percent. The
press uses polypropylene fabric filter
material and no chemical conditioning.
The sludge is trucked offsite for land
disposal. The filtrate from the press
is neutralized with reagent grade
hydrochloric acid in a pH adjustment
tank and then passed through two multi-
media filters in series, followed by two
beds of granular activated carbon (GAC)
1n series before discharge to the local
sewer system.
TMLE 2. CTWIOE TREATMENT 8* ALKALINE CHLORINAT10N (PLANT A)
Period One
YOllM
Totll
TOC
70X
Untts
fta*
Waste
* Gallons 5000
Cyinldc ng/1
og/1
% by
might
Samplinq
5800
20,000
0.30
Results.
Treated
Waste
tooo
<5.0
5700
0.68
Analytical
Perl od
Raw
Waste
3500
11,000
20,000
0.19
Two
Treated
Haste
6300
<5.n
6900
0.60
results from alkaline chlorination
process samples collected at PI ant'A are
summarized in Table 2. Two samples of
untreated cyanide wastes contained
11,000 rag/1 and 5,800 mg/1 of cyanide,
and TOC of 2 percent each. After the
waste was diluted for treatment (about a
1:1 hypochlorite solution to waste
ratio) and cyanide oxidation was
complete, the cyanide concentration in
both effluent samples was less than 5
mg/1. The effluent TOC concentration
averaged 6,000 mg/1. This decrease from
the influent TOC concentration can only
partly be accounted for by dilution with
water and treatment chemicals. The
remaining reduction in organic carbon
content may be due to oxidation by
chlorine, volatilization of organics
and/or adsorption of organics on metal
hydroxide precipitates formed in the
treatment tank.
Total organic halide (TOX) content
of the untreated and treated cyanide
waste was measured to determine whether
chlorinated organic compounds were
formed in the alkaline chlon"nation
process. The data indicates that after
alkaline chlorination the total
chlorinated organic content of the waste
increased from 0.3 percent to 0.68
percent in Sample 1 and from 0.19
percent to 0.60 percent in Sample 2.
Taking dilution from chemical addition
into account, this indicates that the
mass of TOX more than quadrupled in both
cases.
Five samples were collected daily
from the aqueous metal-bearing waste
treatment process during the three day
sampling campaign. Analytical results
are presented in Table 3. Generally,
the aqueous influent samples had High
concentrations of suspended solids,
varying amounts of heavy metals, and
fairly high concentrations of
organics. The most concentrated of the
three samples contained 180 mg/1
cadmium, 1500 mg/1 copper, 500 mg/1
nickel, 1700 mg/1 zinc and 19,000 mg/1
TOC, After precipitation and
filtration, the filtrate contained 50
mg/1 cadmium, 380 mg/1 copper, 22 mg/1
nickel, 46 mg/1 zinc and 5000 mg/1
TOC. Following GAC treatment, the
effluent concentrations were 1.4 mg/1
cadmium, 0.56 mg/1 copper, 1.8 mg/1
nickel, 0.7 mg/1 zinc and 1,400 mg/1
TOC.
TABLE 3. H£»Vl METALS TRESTMEMT - PLANT »
Period 1
Cadoilum
Hex. chromium
Total chromium
Copper
Lead
Nickel
Zlno
TOC
Period 2
Cadmium
Hex. ehroalum
Total ehrooilusj
Copper
Lead
Nickel
Zinc
TOC
Period 3
Cadmium
Hex. chromium
Total chroralura
Copper
Lead
Nickel
Zinc
TOC
Mixed
Aqueous
Influent
(mg/1)
5.7
<20 ,
12
200
1.5
7.3
HOC
5,600
3.9
<5
12
. 150
1.1
«.3
60
12,000
180
<5
18
1,500
3.8
500
1,700
19,000
Filter
Press
Influent
(mg/1)
NS
US
US
NS
MS
US
NS
US
NS
MS
NS -
NS
NS
NS
7,100
NS
NS
KS
»S
NS
NS
NS
10,000
Filter
Press
Sludge
(mg/kg)
12,000
NS
10,000
50,000
1,300
9,300
35,000
5,200
6,700
NS
23,000
13,000
790
6,<)00
11,000
NS
• 2,900
NS
2,200
25,000
830
1,1400
11,000
NS
Filtrate
(mg/1)
1.2
<10 .
2.1
160
1.4
'2.2'
21
NS
7.9 "
NS
1.9
170
0.82
2,3
16
2,'IOfl
50
NS
0,07
380
0.67
22
«
5,000
CAC
Effluent
(mg/1)
0.15
<5
0.09
0.32
0.56
2.2
1.1
1,200
. 0.51
<5
0.12
0.07
0.71
2.0
1.6
1,500
1.1
<5
0.09
0.56
0.63
1.8
0.70
1,400
US - Net sampled
334
-------
TABLE 4.' TCLP AND EP TOXICITY ANALYTICAL RESULTS FOR A METAL HYDROXIDE SLUDGE
Sludge
Compositional
Analysis
Pollutant (mg/kg)
1,1.1 -Trichl oroethane
Methylene chloride
Tri chl oroethyl ene
Tetrachl oroethyl ene
To! uene
230
50
95
200
690
TCLP
Results
(«g/D
3.1
2.8
1-.7
0.66
7.2
EP Toxicity
Results
(mg/1)
NA
NA
NA
NA
NA
Proposal Proposed
Health Based Liner Protection
Threshold Threshold
(rag/1 ) (mg/1 )
1,300
1.2
0.1
0.015
22.0
2.0
2.0
Maximum EP
Toxicity
Threshold
(mg/D
NA
NA
NA
NA
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selen i urn
Silver
NS
NS
7,200
11,200
970
NS
• NS
•NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
0.13
0.26
0.38
<0.10
NS
NS
NS
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
NS - Not sampled
NA - Not applicable
Samples were collected of two waste
acids used in the pH adjustment
process. Waste acid Sample 1 was
identified as a strong hydrochloric
acid. Heavy metals present 'in the acid,
and therefore introduced to the aqueous
waste, included 15,000 mg/1 zinc, 5800
mg/1 copper, and lesser concentrations
of cadmium, chromium, lead, and
nickel. Waste acid Sample ? was
identified as a waste sulfuric acid.
Chromium was present at a concentration
of 24,000 mg/1.
The clay-like filter press sludge
samples were very dry. High
concentrations of metals were reported
in each sample. In addition, fairly
high amounts of organics were found to
have concentrated in the hydroxide
sludge. This was consistent for each of
the samples and duplicate samples
collected. Toluene and 1,1,1-
trichloroethane were the organic
compounds present in the sludge samples
in the highest concentrations for each
sampling period. The sludge sample was
also subject to the extraction procedure
toxicity test (EP Toxicity) and the
proposed toxicity characteristic
leaching procedure (TCLP) test. Table 4
summarizes the analytical results and
lists the maximum EP toxicity threshold
concentrations and the proposed TCLP
health based and liner protection
threshold concentrations for
comparison. It is interesting to note
that although this sludge meets the
maximum EP toxicity threshold criteria
for the metals tested, it is not in
compliance with the proposed health
based threshold limits for methylene
chloride, trichloroethylene, and
tetrachloroethyl ene or within the
proposed liner protection threshold for
1,1,1-trichloroethane and toluene.
The GAC effluent, the finished
product of the aqueous waste treatment
system, was a clear, neutral. pH
effluent. With the exception"of acetone
in each of the .samples, and methylene
chloride and 1,1, 1-trichloroethane in
the Period Three sample, volatile and
extractable priority pollutant organics
were not detected. Di-n-butylphthalate
was detected in very low concen-
trations. Heavy metal concentrations
were low for each of the samples
collected. Lead, copper, and chromium
were consistently reported in
concentrations of less than 1 mg/1,
while cadmium, nickel and zinc were
consistently less than 2 mg/1.
Plant B
Facility Description. A schematic
of the treatment process employed at
Plant B for corrosives and metal-
containing wastes is shown in Figure
2. Typically the corrosive wastes are
diluted with well water and reagents
until a neutral pH is established.
335
-------
Process
Water
Lime
Slurry
Centrifuge
Incoming Aqueous
Waste
Batch Reactor
Tank
Centrate to
Sewer System
Dewatered Sludge to
Metals Reclaimer
or Landfill
Sample Locations
FIGURE 2. WASTE TREATMENT FACILITY - PLANT B
If the waste is extremely
corrosive, then process water and lime
slurry are charged to the reaction tank
first and the waste is slowly bled into
the stirred tank in order to prevent
excess heat generation. The majority of
acids treated by Plant B contain
significant amounts of metals. For these
wastes, precipitation of metal
hydroxides is accompli shed"by means of
Hme addition and mixing in polyethylene
tanks. Depending on the waste, sodium
silicate is added to enhance
flocculatlon. The waste is settled in
the same polyethylene tanks under
quiescent conditions. The decant is
pumped through a basket centrifuge to
the local municipal sewer system and the
gravity thickened sludge is pumped to
the same basket centrifuge for
dewatering. The centrate passes to the
sewer system and the dewatered sludge,
typically about 25% solids, is
transported to an approved landfill or
to a metals reclaimer.
Sampling Results. The waste
treated during the sampling period was a
nitric acid solution from printed
circuit board manufacturing which
contained about 7 percent copper. The
waste also contained significant amounts
of nickel, tin, lead and chromium, plus
chlorinated organic solvents.
Approximately 35 gallons of this nitric
acid copper contaminated waste was
treated over a two day period. The acid
waste was slowly added to about 400
gallons of lime slurry (150 pounds of
high calcium lime) and dilution water.
The tank was agitated with a paddle
mixer, and waste addition was
discontinued when the pH dropped to
8.0. Sodium silicate was added to
enhance flocculation and the pH was
adjusted to 7.5 with waste sulfuric
acid. Gravity separation was allowed to
proceed over a 16-hour period, and the
decant was pumped to the sewer system.
The precipitated hydroxide sludge was
pumped to a perforated basket centrifuge
that was lined with filter paper.
This is an example of how an
aggressive acid waste, i.e. nitric
acid, containing a very high
.concentration of metals, was accepted
and treated by applying known technology
and existing equipment. With no prior
treatability experience, jar tests were
followed by full scale batch treatment
with visual observation and simple
process monitoring (i.e. pH) being used
to determine the rate and amount of
treatment chemicals and the completion
of each treatment step. A primary
objective was to produce a copper sludge
that would be acceptable to a metal
reclaimer. This objective was
achieved. The treatment efficiency
however, was very low, due to poor
performance of the centrifuge.
Sampling results are summarized in
Table 5. Only about 50 percent removal
of heavy metals was achieved due to the
336
-------
TABLE 5. HEAVY METALS TREATMENT - PLANT B
Raw
Waste
(mg/1)
Raw Waste
After
Dilution*
(mg/1 )
Treated Effluent
(Cent rate)
Unfiltered
(mg/1)
Soluble
(mg/1)
Percent
Removal
(unflltered)
Hetals
Copper
Nickel
Tin
Lead
Chromium'
Zinc
VOCs
Methylene Chloride
1 ,l,l'-Trichloroethane
Tetrachloroethane
70,000
3,700
9,100
1,900
250
23
140
90
12
4900
'259
658
133
17.5
0.16
9.8
6.3
0.84
2500
30
320
92
8.8
0.34
0.12
0.013
<0.01
28
16
<1
2.0
<0.3
<0.2
NS
NS
NS
49
50
51
31
50
98.8
99.8
>98.8
poor capture of solids by the
centrifuge. Though the concentration of
metals was reduced by over 90 percent,
this was due primarily to dilution.
Dissolved metal residuals in the
cent rate indicate that total heavy metal
mass removals could he increased from 50
percent to 98 percent with improved or
additional solids capture.
Treatment was labor intensive,
requiring two days to treat a single
batch of 35 gallons of waste. The rate
of acid addition was limited by heat
generation. Approximately 17.5 gallons
of dewatered sludge and 500 gallons of
treated effluent by volume resulted from
the treatment of the 35 gallons of
nitric acid copper etchant. Solids
separation by the centrifugation process
was incomplete.
CONCLUSIONS
Corrosive and metal-containing
wastes are routinely treated by most
commercial hazardous waste treatment
facilities. Established technologies
are generally applied; however, the
application of these technologies to the
highly concentrated, mixed wastes that
are accepted for treatment by the
commercial TSDFs often requires special
process design and operating
procedures. Jar testing and "custom
designed" batch treatment procedures are
the norm.
Effluent quality acceptable for
discharge to municipal sewer systems was
achieved through treatment of the
corrosive and metal-containing wastes by
neutralization and precipitation
methods, with effluent polishing if
necessary by filtration and granular
activated carbon adsorption. Analysis
of a dewatered metal hydroxide sludge
from 1ime treatment of a concentrated
waste containing metals and organics
showed a high concentration of metals
and organics in the sludge. Extraction
tests showed that this sludge would not
meet proposed TCLP criteria for land
disposal based on the concentration of
several organic compounds in the extract
solution. In contrast, the metals
analyzed as part of the EP toxicity test
did not exceed the respective maximum
contaminant concentrations.
337
-------
RECOVERY OF METAL VALUES FROM METAL-FINISHING HYDROXIDE SLUDGES
BY PHOSPHATE PRECIPITATION
L. G. Twidwell, D. R. Dahnke, 8. W. Arthur, S. M. Nordwick
Department of Metallurgy and Mineral Processing Engineering
Montana College of Mineral Science and Technology
Butte, Montana 59701
ABSTRACT
A process for selectively recovering trivalent cations from electroplating hydroxide
sludge materials has been developed. The process is based on phosphate precipitation.
The advantages of the process are:
'simple selective precipitation
"easy solid/liquid separation
"effective conversion of the precipitated phosphates to other products with the
regeneration of phosphate reagent.
Application of the process to mixed metal hydroxide sludges will be discussed.
Experimental results illustrating selective separation, filterabiHty rates, and
conversion possibilities will also be presented.
INTRODUCTION
Metal bearing hydroxide sludge
material 1s generated by the metal
finishing industry. Environmental impact
and resource conservation considerations
have in recent years promoted development
of alternatives to disposing of these
hydroxide materials into hazardous waste
containment sites. Separation and
production of pure metal salts from mixed
metal sludges is one such alternative. In
many cases synthesis of these metal salts
is very attractive economically, with
purified forms demanding high market value.
In other instances, metal compounds can be
produced for reuse in the metal finishing
circuit.
Preliminary work on developing a
series of selective precipitations for the
separation and recovery of soluble metal
components common to electroplating sludges
has been completed. As a part of this work,
flowsheets were developed for separation of
iron, copper, zinc, chromium and nickel, and
the various unit operations of the flow-
sheets were tested at a flow rate of 200
liters per day. Although this treatment and
recovery system performed satisfactorily the
chromium extraction phase involved the very
costly step of chromium oxidation. Use of a
phosphate precipitation system, however, may
offer the advantages of allowing selective
precipitation of metal phosphates, producing
precipitates with better filterability char-
acteristics than hydroxide or sulfate so-
lids, and eliminating the chromium oxidation
step. The precipitations may be accomplish-
ed by exploiting the selective solubility as
a function of pH and solution temperature of
various metal phosphates. Typically, triva-
lent metal phosphates are very much less
soluble than divalent metal phosphates.
Metals such as iron or chromium can be ra-
pidly and cleanly separated from solutions
containing many grams per liter (g/liter) of
nickel, zinc or cadmium. These iron and
chromium phosphates form as dense, spherical
particles and yield excellent filterability
and washing characteristics.
A project is currently in progress at
Montana College of Mineral Science and
Technology to pursue the final development
of this process in terms of long-term
operation and integration with other, more
traditional hydrometallurgical unit
338
-------
operations. An extensive final report will
be issued upon completion of the project.
TREATMENT TECHNIQUES
Leaching
Sulfuric acid 1s a very effective
leaching agent for metal-bearing hydroxide
materials; one gram of HoSO, is required
for each gram of solids present in the
sludge. The leaching operation is
performed at ambient pressure and without
external heating. The heat of reaction is
sufficient to raise the temperature of the
leach slurry to about 50°C, which
undoubtedly aids the dissolution process.
Retention time is 30 minutest followed by
solid-liquid separation. The filtrate
typically contains better than 95% of the
metal values available from the sludge. A
small fraction of unleachable materiali
usually consisting of sand from filters,
wood chips, plastic, etc., comprises the
filter cake. The leach solution is usually
diluted to yield a total dissolved solids
level of 40 g/liter. Some water may be
added during the leach, with the remainder
supplied after solid-liquid separation.
Filter cake wash water is used in this
second dilution to enhance overall metal
recovery. The leaching operation produces
a feed stream with a pH value of
approximately 1.5.
Calcium Removal
Some sludge materials contain a
significant fraction of Ca, present either
from water deionizing system backwashing
operations, or from CaO used in waste water
treatment. When sulfuric acid is employed
to leach a high calcium bearing material,
gypsum (CaSO.'EHpO) will selectively
precipitate. The solubility limit for
calcium in the leach liquor is
approximately 0.6 g/liter. If a minimum
volume of dilution water is added before
solid-liquid separation, the total mass of
calcium rejected is increased and a final
feed solution bearing, 0,2 to 0.4 g/liter
calcium can be produced. The gypsum
precipitate enhances filterability of the
leach residue. The gypsum bearing filter
cake can either be disposed, or treated to
recover pure gypsum which could then be
calcined to produce plaster of Paris.
A high calcium bearing electroplating
sludge material was treated with sulfuric
acid to leach the metal values. A heavy
white precipitate, gypsum, formed during
the leach operation. This process rejected
approximately 88% of the total calcium from
the leach solution. Data in Table 1 repre-
sents the analysis of the sludge material
while data in Table 2 gives the analysis of
the leach solution. Note that if the total
calcium available had leached, a calcium
concentration of 6.0 g/liter would have
resulted. The gypsum produced in this test
was 93% pure CaSQ4"2H20 with iron being the
major contaminant.
Copper Recovery
High levels of copper present in the
leach solution,preclude removal of Fe b«
ferric phospate precipitation because Cu
will coprecipitate with the iron. For
those sludge materials which contain signif-
icant fractions of copper, the removal of
copper from the leach solution must precede
ferric phosphate precipitation. Hydrometal
lurgical treatment of copper is a well
developed commercial operation and these
technologies are readily applicable to
copper bearing hydroxide sludge leach solu-
tions. By far the simplest method for
copper recovery from acidic, aqueous media
is by cementation with iron. The cementa-
tion process is described by the reaction:
Fe° -1- Cu^
-> Cuc
(1)
In small scale operations the cemen-
tation process is carried out by suspend-
ing a permeable basket containing iron
filings in the Teach solution. Containment
of the iron filings In this fashion facili-
tates rapid removal of the cement copper
product from the process solution after a 15
to 30 minute retention time. Copper
cementation is not without its
drawbacks:
1. an impure copper is produced which
must further be refined by
smelting,
2. iron in solution is reduced to the
ferrous (Fe ) state, and
therefore must be oxidized before
precipitation as ferric (Fe )
phosphate.
Cementation of copper with iron powder
was tested in the laboratory to remove copper
from a mixed-metal electroplating sludge
leach solution(2). A large excess of iron
powder (Mr- /M~ =10) was slurried into the
mixed-metaf solution for one-half hour with a
339
-------
solution temperature of 57°C. Starting and
ending solution concentrations of metals
are reported In Table 3. Iron powder Is
obviously very effective In rapidly remov-
ing copper from solution. The composition
of the cemented copper 1s reported In Table
4. The large fraction of Iron contained In
this material 1s Indicative of having used
a large excess of Iron for this particular
test. Copper can successfully be cemented
from solution using much less Iron which
would result 1n a product with higher
copper values.
Another commercially practiced method
of recovering copper from aqueous solution
1s by solvent extraction. This technique
involves the selective transfer of cupric
(Cu ) ions from the aqueous leach solution
Into an immiscible organic phase. This
organic phase contains a reagent which is
capable of forming stable organometallic
complexes with copper. After the organic
1s loaded with copper by contact with the
slightly acidic feed solution, it is
contacted with a highly acidic aqueous
solution where the copper is transferred
from the organic into the acidic strip
phase. Copper is recovered from the strip
solution either by copper sulfate
crystallization or electrowinning. This
method produces a readily salable copper
product, but 1s Inherently more complex in
operation than cementation.
Extensive data has also been collected
on removing copper from mixed-metal
electroplating hydroxide sludge leach
solutions by solvent extraction.(3) A Bell
Engineering laboratory size continuous sol-
vent extraction system was set up with
three extraction and two strip cells. A
solution with 15 volume percent LIX-622
dissolved in Kermac 470-B kerosene was used
as the organic extraction agent in this
test work. Strip liquor was a solution of
200 g/Hter sulfuric acid and 30 to 40
g/Hter copper. This system was operated
through eleven, 40 liter batches of mixed-
metal leach solutions. Results of extrac-
tion efficiencies from this series of tests
are presented in Table 5 and indicate very
good extraction of copper.
Iron Precipitation
An extremely Important aspect of this
treatment scheme as a whole is the ease of
Iron removal via precipitation with
phosphate. Iron is a common constituent of
many hydrometallurgical process streams.
and 1s normally considered to be a
deleterious impurity because of its low
value. Two precipitation techniques are
currently employed industrially to remove
iron from process solutions. One technique
is the high temperature synthesis of ferric
hydroxide which yields a very high surface
area solid that is difficult to filter and
adsorbs many ionic species from solution
resulting in heavy contamination of the
iron bearing solid. The second method
entails the high temperature formation of a
basic hydrous alkali sulfate of iron, a
mineral named jarosite. This process re-
quires near boiling solution temperatures
for several hours to promote acceptable
reaction kinetics, and will reduce the Iron
concentration in aqueous solution to only
0.2 to 0.4 g/liter. Also, if chromic
(Cr ) cations are present, they will co-
precipitate with the iron and contaminate
the jarosite with a valuable metal. In
spite of these drawbacks, both of these
iron removal processes currently are
practiced industrially.
A significant, recent development in
iron removal technology 1s the precipitation
of ferric phosphate. Ferric phosphate pre-
cipitation prevails over the inadequacies of
the presently employed industrial tech-
niques. Rapid formation of ferric phosphate
occurs in acidic aqueous media at room tem-
perature, requiring an hour or less to re-
duce several grams per liter of iron to less
than 0.1 g/liter. The precipitated solids
are spherical, which is the geometric shape
with the lowest specific surface area, and,
therefore, the tendency to adsorb contamin-
ating ions from solution is much reduced.
This particle morphology also promotes
excellent filterability and easy washing of
the solids to remove entrapped solution.
Ferric phosphate of remarkable purity, i.e.,
better than 991 FePQ4'2H2Q can be precipi-
tated from a solution containing several
grams per liter each of a yariet^of other
ionic species inc3uding,Ni , Zn , Cr ,
Gd , S04 , N03~', Cl ' CrO ~z, etc. After
the iron phosphate has been filtered from
solution and washed, conversion to ferric
hydroxide is readily accomplished by slur-
rying the ferric phosphate into a high pH,
aqueous solution that is slightly elevated
in temperature. Conversion is complete in
one to one-and-one-half hours. This process
recovers the valuable phosphate reagent in a
form usable for pH control and phosphate
addition in the leach liquor treatment cir-
cuit. Ferric hydroxide produced in this
manner filters reasonably well, much better
340
-------
than does ferric hydroxide which is
precipitated directly from solution.
Extensive data have been collected on
the precipitation of ferric phosphate from
a wide variety of process solutions and
under a wide range of conditions. Details
of these experiments are presented in other
publications.(2,4) A summary of the solu-
bility of ferric phosphate as a function of
pH is presented in Figure 1 (the solid lines
represent the solubility of solid ferric
phosphate and ferric hydroxide). It is
imperative to note that this solubility
function for ferric phosphate is but little
affected by a wide variation in solution
compositions or temperatures. Ferric phos-
phate precipitates equally well from sulfate,
chloride, or nitrate systemsf and the solu-
bility is not significantly affected by
solution temperature.(5)
Work thus far performed on converting
ferric phosphate to ferric hydroxide sug-
gests that high pH and moderate solution
temperatures are required to achieve com-
plete conversion. A summary of preliminary
data is presented in Figure 2.(6)
Chromium Treatment
Trivalent chromium is removed from
solution by phosphate precipitation in a
manner very similar to iron removal. Un-
like the solubility of ferric phosphate,
however, the solubility of chromium phos-
phate exhibits a strong dependence on solu-
tion temperature, with increasing tempera-
tures promoting reduced solubilities. This
property is^explcited,to achieve a separa-
tion of Fe from Cr . At ambient temper-
atures, i.e., 20°C, ferric phosphate will
precipitate cleanly in the presence of
chromic cations through the pH range 1.5 to
2.1. After ferric phosphate is filtered
from solution, chromic phosphate can be
made to precipitate through.approximately •
the same pH range by heating the solution
to 50 to 60°C. The individual chromium
phosphate particles are spherical in shape
which results in excellent f1,lterability
and minimum surface adsorption of other
ionic species. . .
A market for chromic phosphate exists
in the pigment industry. This; market may
very well be exploited to consume the chro-
mic phosphate produced from hydroxide
sludge materials. An attractive alterna-
tive to the sale of chromic phosphate is,
the conversion to chromic hydroxide with
the resulting recovery of phosphate in the
very alkaline solution. The chromic hy-
droxide can be calcined to produce high
value chromic oxide ($5.50/lb.),(7)
Some hydroxide sludge materials
contain chromium in the hexavalent state;
sludge produced by electrochemical machin-
ing is one such example. In this case, it
is economically advantageous to recover
chromium in its oxidized chromate state.
This is easily accomplished by precipita-
tion of lead chromate from mildly acidic
solutions, i.e., pH > 4.0. Lead chromate
filters well and can be leached with sul-
furic acid to produce lead sulfate and
chromic acid. The lead sulfate so produced
can be returned to the leach liquor treat-
ment circuit for recovery of additional
chromate. Chromic acid produced by this
methodology can be returned to a plating
bath.
Sufficient data has thus far been
generated to characterize chromium phosphate
precipitation.(4) Solubility data for
chromium phosphate precipitation is sumrnai—
ized in the stability diagram presented in
Figure 3. Solubility data as a function of .
solution temperature is presented in Figure
4 and clearly shows the inverse relation
between solubility and temperature. It is
evident that several grams per liter chrom-
ium can be maintained in solution to pH
levels greater than four, with low solution
temperature. Precipitation of lead chro-
mate has been tested for removing chromate
ions from solution. Because of the great .
solubility of chromate as a function of pH,
this is usually the last ion to be removed
in the leach solution treatment scheme. A
leach solution produced from electrochemi-
cal machining sludge was treated for getal
value recovery, with 1.59 g/liter Cr
remaining in solution after ferric
phosphate and nickel hydroxide precipita-
tion. Chromate concentration was reduced
to 0.3 parts per million by adding ,a slight
stoichiometric excess of lead nitrate and
precipitating at a solution pH of 6.1.
Aluminum Precipitation
The solubility of trivalent aluminum
in aqueous solution is essentially
identical to that of chromium. Separation
of iron from an aluminum-chromium bearing
solution can be accomplished by precipita-
tion of ferric phosphate at low tempera-
ture. Heating the solution, after solid-
liquid separation to remove ferric phosphate
341
-------
solids, results in copreeipitation of alumi-
num and chromic phosphates. These two
metals can be separated by leaching the
mixed-metal phosphate, oxidization of chro-
mic cations to chromate anionst and repre-
cipltation of aluminum phosphate; this time
without contamination by chromium. The
chromate bearing filtrate which results
from filtration to remove aluminum phos-
phate can be treated by lead chromate pre-
cipitation followed by chromic acid and
lead sulfate production.
Separations by Valence
The preceeding discussion has facused,
on separating trivalent cations (Fe , Cr
and Al ) from each other and from various
divalent cations (Zn , M"**). Certain
instances arise where the isolation of each
Individual trivalent cation is not desir-
able. Instead, only removal of contaminat-
ing trivalent cations from a predominately
divalent specie bearing solution is desir-
able. In this case, iron, chromium, and
aluminum can be made to copreeipitate as
phosphates by elevating the solution tem-
perature to 50 to 60°C. This accomplishes
purification of the aqueous solution in a
single precipitation and solid-liquid sep-
aration operation. If present in only
small quantities, the trivalent, mixed-
metal phosphate material can be converted
to hydroxide to recover the phosphate with
subsequent disposal of the hydroxides. The
purified solution can be returned to the
plating circuit. An example of this appli-
cation is the recovery of pure zinc phos-
phate from the iron-zinc phosphate sludge
that typically forms in phosphatizing
tanks.
A mixed-metal leach solution was pro-
duced from electroplating siudge,and ,
treated to psecipitate Fe , Cr , and AT"
from the Zn rich solution. Data for this
test is presented in Table 6 and illus-
trates excellent separation of the
trivalent cations from the divalent zinc.
SUMMARY
A general overview for each of the
unit operations employed to recover metal
values from electroplating hydroxide sludge
materials has been given. These operations
are either applications of well developed
commercial technology or are simple
selective precipitations. Overall the
technology utilized, except maybe for
solvent extraction, does not extend beyond
the normal realm of waste water treatment
methods. All precipitations used in the
foregoing metal recovery scheme require pro-
per reagent addition, pH control, and control
of solution temperature. Indeed, these are
criteria very common in waste water
treatment.
The precipitation techniques described
herein are not equipment intensive. A series
of stirred vessels with appropriate heating
capabilities, pH sensors and reagent addition
equipment, coupled with a filtration system,
is essentially all that is required to imple-
ment this technology on an industrial scale.
These types of equipment are prevalent in the
water treatment industry. It is certainly
feasible that some sludge generators can
successfully implement at least some of these
precipitation techniques with very little
operational or equipment change and minimal
equipment acquisition. The return on in-
vestment can begin immediately, because the
volume of sludge requiring hazardous waste
handling and disposal can be very greatly
reduced, reagents can be regenerated for use
in the process line resulting in reduced
expenditure for raw materials, or marketable
metal salts can be produced. This, of
course, is all in addition to the
conservation of nonrenewable resources.
An EPA sponosored research program to
test these precipitation schemes on a pilot-
plant scale is currently underway at the
Montana College of Mineral Science and
Technology. A wide variety of sludge ma-
terials have been collected for treatment
in the test system. After analyses of
chemical composition and moisture content,
each sludge is processed through the system
in a series of repetitive tests. Data on
elemental distributions are collected and
careful mass balances are maintained.
Process variables are optimized to yield a
very low concentration of a particular metal
in solution and still maintain high purity of
the recovered metal product. Successful
completion of this test project should result
in an industry-ready method for recovering
metal values from metal bearing hydroxide
sludge materials.
1, Biswas, A. K., Davenport, W. 6.
Extractive Metallurgy of Copper,
Pergamon Press, Oxford, 1976, p. 272.
2. Dahnke, D. R., Removal of Iron from
Acidic Aqueous Solutions. M. S.
Thesis, Montana College of Mineral
342
-------
Science and Technology, May 1985, p.
137.
3. Twidwell, L. 6. Metal Value Recovery
from Metal Hdroxide Sludges, Report for
EPA Projects R-80930501 and R-80173601,
Montana College of Mineral Science and
Techology, November 1984, p. 290.
4. Dahnke, D. R., Twi dwell, L. G., Robins,
R. G. Selective Iron Removal from
Process Solutions by Phosphate
Precipitaton, CIM 16th Annual
Hydrometallurgical Meeting, Toronto,
Canada, October 1986.
5. Arthur, B. Recovery of Metal Values
from Iron, Chromium, Nickel Solutions,
M.S. Thesis, Montana College of Mineral
Science and Technology, August 1986.
6. Nordwiek, S. Conversion of Phosphate
Solids Into Hydroxides, M.S. Thesis,
Montana College of Mineral Science and
Technology, August 1986.
7. Chemical Market Reporter, January 6,
1986, Schnell Publishing Company, Inc.,
New York, NY.
343
-------
TABLE 1. ANALYSIS OF SLUDGE
Compound
NaOH
A1(OH)3
S1(OH)4
P04
so4
C1
Ca(OH)2
Cr(OH)3
Fe(OH)3
Zn(OH)2
% Composition
8.43
4.63
2.69
5.04
8.24
0.36
20.49
16.07
5.13
18.00
TABLE 2. ANALYSIS OF LEACH SOLUTION
Compound Concentration g/liter)
Fe 0.660
Cu 0.024
Zn 5.304
Cr 2.959
Al 1.187
Ca 0.692
P 0.883
344
-------
TABLE 3. SOLUTION ANALYSES FOR CEMENTATION OF COPPER
Starting
Final
Concentrations (g/liter)
Fe Cu Zn Cr Ni Al Ca Na Cd
0.847 3.596 3.365 5.606 6.848 0.038 0.232 0.576 3.604
11.141 0.032 3.257 5.397 6.943 0.037 0.236 0.652 3.518
TABLE 4. CEMENT COPPER COMPOSITION
% Composition
Fe Cu Zn Cr
75.1 22.4 0.70 1.31
Cd
0.54
345
-------
-TABLE 5. , SUMMARY OF CONTINUOUS COPPER EXTRACTION: ELEVEN DAY LONG TERM ORGANIC
EXPOSURE TEST RESULTS.
Sample No.
Condition
Copper
Extraction From
Copper Concentration (gpl)
3458
3474
3482
3493
3501 -B
3509
3519
3533
3542
3548
3552
3567
3605
3613
3619
3631
3639
3643
3657
3664
3670
3703
Starting Solution
First Day Raffinate
Starting Solution
Second Day Raff.
Starting Solution
Third Day Raff.
Starting Solution
Fourth Day Raff.
Starting Solution
Fifth Day Raff.
Starting Solution
Sixth Day Raff.
Starting Solution
Seventh Day Raff.
Starting Solution
Eight Day Raff.
Starting Solution
Ninth Day Raff.
Starting Solution
Tenth Day Raff.
Starting Solution
Eleventh Day Raff.
Initial
2.750
3.130
2.697
3.332
2.600
0.835
1.035
2.045
1.812
2.026
2.225
Final
0.054
0.062
0.106
0.088
0.039
0.056
0.033
0.027
0.073
0.043
0.049
Leach Solution
Copper Extracted (%)
98.0
98.0
96.1
97.4
98.5
93.3
96.9
98.7
97.0
96.0
97.8
346
-------
TABLE 6. SOLUTION ANALYSIS FOR PRECIPITATION OF Fe4"3, Cr+3» AND Al+3 PHOSPHATE
Concentrations g/liter
Fe Cr AT Zn P
Starting Solution 0.118 0.694 0.182 3.735 1.151
pH = 2.07
Ending Solution 0.048 0.073 0.016 3.760 0,684
pH = 2.48
347
-------
PH
9 10 n
OJ
u.
o.
SPECIES CONSiPBREP
I t. r«*04-2H20(8)
| 2.,F*COH)S{S)
| 4. F«OH*2
I 5, F«{CH)J
I 6. FB{OH)S
7. F*(OH)J
8. FB2(OH)f4
. F«SO4*
I*. HPO4-2
17. 'P04'3
18. H2304 -
IB. HSO4-
20. 8O4"2
21. FK2(S04)3(S)
22. N»*
124. N*HPO4-
|2S.
NOTE:
15
-LOG a.
FE _
FP = -LOO a
i I I
1 - 1 - L
FIGURE I . FERRIC PHOSPHATE STABILITY DIAGRAM FOR THE
FERRIC-PHOSPHATE-SULFATE-WATER SYSTEM.
348
-------
100
PH = 11, SO°C
pH = 12, 2S"C
CONDITIONS
FBP04- 2H2O( S ) SLURRIED IN PH
CONTROLLED SYSTEM
20 WT. % SOLIDS
PH = 8. 50 C
I
I
60
120 180
Time, Min.
240
300
FIGURE 2. THE RATE OF CONVERSION OF FERRIC PHOSPHATE TO FERRIC
HYDROXIDE AS A FUNCTION OF PH AND TEMPERATURE.
349
-------
PH
Q.
O.
5-
O
CL
10 11
I
I
SPECIES CONSIDERED
I . O»PO4 ( S )
2. Cn+3
3. CH(CH)+2
5. CR(CH)3
6. CH(CH)4-
7. Cl»2(OH)2+«
8. CR3(OH)4+5
9. C»(OH),(S)
H3P04
H2P04-
HP04-2
FIGUFJE 3.
CHROMIUM PHOSPHATE STABILITY DIAGRAM FOR THE
CHROMIUM-PHOSPHATE-WATER SYSTEM.
350
-------
e»
^
E
Cn
o
INITIAL CONDITIONS
FC P Nl Cn (GPL).
O 25°C 4.20 6.04 6.4S 6.26
O 50°C 4.14 5.78 6.44 6.27
•
• 80°C 3.59 5.09 6.48 5.93
Solution pH
FIGURE 4. INFLUENCE OF TEMPERATURE ON CHROMIUM PHOSPHATE PRECIPITATION
FROM A MIXED METAL SOLUTION.
351
-------
TREATMENT TECHNOLOGIES FOR CORROSIVE HAZARDOUS WASTES
H. Paul Warner
Robert A. Olexsey
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The 1984 amendments to the Resource Conservation and Recovery Act (RCRA), called the
Hazardous and Solid Haste Act (HSWA) amendments, direct EPA to determine if corrosive
wastes should be prohibited from landfills. This determination is to be made by July
1987. Barring a determination by the Agency that the land disposal of corrosive wastes
does not represent a threat to the environment, or that appropriate alternative treatment
technology is not available, a land disposal prohibition would go into effect at that
time. The purpose of this presentation is to review the state-of-the-art for treatment
processes for corrosive wastes that have been proposed to be used Instead of land disposal
for disposing of solid and liquid corrosive wastes.
INTRODUCTION
A hazardous waste is characterized as
a corrosive waste if it meets the following
criteria:
1, It is aqueous and has a pH less
than or equal to 2 or greater than
or equal to 12.5.
2. It is a liquid and corrodes SAE
1020 steel at a rate greater than
6.35 mm (0.25 inch) per year at a
test temperature of 55°C (13Q°F)
(1).
Corrosive wastes, along with wastes
containing metals and cyanides and halogen-
ated organics comprise the "California
List" wastes in that they are cited in
regulations developed by the California
Department of Health Services. The Hazard-
ous and Solid Waste Amendments (HSWA) of
1984 require that the U.S. Environmental
Protection Agency (EPA) decide by July 8,
1987 what restrictions to impose on the
land disposal of the California List
wastes. If EPA has not established treat-
ment standards for corrosive wastes by that
date, those wastes are banned from land
disposal (2).
The purpose of this paper is to pro-
vide a summary review of commercially
available treatment process technologies
for various types of corrosive hazardous
wastes. Applicability and limitations of
each process for specific wastes are
outlined with discussion.
WASTE CHARACTERISTICS AND GENERATION
The wastes which are listed as corro-
sive wastes under the stipulations of the
Resource Conservation and Recovery Act of
1976 (RCRA) have the following RCRA identi-
fication codes:
D002 A solid waste that exhibits the
characteristic of corrosivity
but is not listed as a hazardous
waste in Subpart D of the RCRA
regulations (waste is not an F,
K, U, or P waste)
K062 Spent pickle liquor from steel
finishing operations
The predominant source of generation
of 0002 wastes are the primary metals and
metals finishing industries, with about 44
percent of the total. The chemicals indus-
try is responsible for 27 percent followed
352
-------
by the utilities (8 percent) and petroleum
(4 percent).
The statutory descriptions of corro-
sive wastes are specific to liquids.
However, many of the wastes which are
classified as D002 wastes contain solids
and sludges.
yhile only two RCRA waste codes are
listed as corrosive, these wastes are
generated at substantial rates in the
United States. Data on waste generation
for 1981 are as follows (3).
Waste
I.D.
D002
K062
Waste Generation
(Millions Gallons)
19,916
849
Number Of
Facilities
513
64
In addition, many of the wastes that are
listed as metals or cyanide bearing wastes
also exhibit the characteristic of corrosi-
vity. For reference, total U.S. hazardous
waste generation in 1981 was 240 billion
gallons.
In 1981, about 3.4 billion gallons of
hazardous corrosive wastes were disposed of
to the land, with the largest amount, 3.0
billion gallons, going to deep well injec-
tions. Thirteen billion gallons were
treated in tanks, surface impoundments, and
incinerators. Simple storage accounted for
about 8.4 billion gallons (4). Treatment
is conducted mainly onsite, that is, at the
point of generation. Treatment, storage,
and disposal quantities may exceed genera-
tion rates due to some double accounting;
that is, some wastes which are both stored
and treated are accounted for in both cate-
gories.
Treatment Objectives
The goal of any alternative hazardous
waste treatment process is to render the
waste harmless, non-toxic, and non-leach-
able into the groundwater. Liquid dis-
charges must meet either National Pollutant
Discharge Elimination System (NPDES) or
Publicly Owned Treatment Works (POTW) pre-
treatment requirements.
A principal method for determining
residue quality has been the Extraction
Procedure Toxicity Characteristic (EPTC).
The Toxicity Characteristic entails use of
a leaching test to measure the tendency of
a waste to leach, coupled with extract
concentrations above which the waste is de-
fined to be a regulated, or hazardous
waste. The current EPTC addresses only 8
elements and 6 organic compounds. The
Agency is currently developing an improved
leaching test, the Toxicity Characteristic
Leaching Procedure (TCLP). The TCLP will
addresss over 50 compounds and is expected
to be more rugged and precise than the
EPTC. The expanded TCLP will most likely
be promulgated by EPA late in 1986 (5).
When the TCLP goes into effect, any
waste, including a corrosive waste, must
pass this test before it can be disposed of
to the land. Therefore a key benchmark for
any treatment process is the ability of the
residue to pass the TCLP. If the residue
does not pass the TCLP it must be subjected
to further treatment until it passes the
TCLP.
TREATMENT TECHNOLOGIES
The treatment technologies discussed
here were extracted from a draft report to
USEPA by Camp, Dresser, and McKee, Inc.
(4).
There are essentially three technolo-
gies employed in the treatment of corrosive
wastes. They are neutralization, resource
recovery, and incineration. Neutralization
is the most commonly applied treatment
technique and will be discussed more exten-
sively than the other technologies.
Simply, the process of neutralization
is the interaction of an acid with a base.
The typical properties exhibited by acids
in solution are a result of the hydrogen
ion concentration, (H+). Similarly, alka-
line (or basic) properties are a result of
the hydroxyl ion concentration, (OH~). In
aqueous solutions, acidity and alkalinity
are defined with respect to pH, where
pH = - log (H+), and pH = 14 - log (OH-)
(at room temperature), respectively. In
the strict sense, neutralization is the
adjustment of pH to 7, the level at which
the concentrations of hydroxyl ion and
hydrogen ion are equal. Solutions with
excessive hydroxyl ion concentration (pH>7)
are said to be basic; solutions with excess
hydrogen ions (pH<7) are acidic. Since
353 -
-------
adjustment of pH to 7 is not often practi-
cal or even desirable in waste treatment,
the term "neutralization" is sometimes used
to describe adjustment of pH to values near
neutrality (6).
The first step to be taken in the
treatment of corrosive wastes is treat-
ment process selection. The appropriate
applied treatment technology is dictated by
the results of a corrosive waste character-
ization. This characterization will answer
questions such as;
(a) Should the waste be neutralized?
(b) Does the waste contain chelated
metals, cyanides or chromium?
(c) If the waste is neutralized, at
what point in the treatment
scheme should it be neutralized
and with what neutralization
agent?
(d) Are the wastes compatable with
the neutralizing agent?
(e) What is the effect of the forma-
tion, if any, of soluble or in-
soluble salts?
(f) What are the downstream process
effects?
Following an evaluation of the results
of the physical and chemical characteriza-
tion of the corrosive waste, a generic
treatment proces is selected and, when
necessary, modified to incorporate specific
techniques which must be applied to treat
specific wastes.
There is very little information re-
lating to the concentration of waste acids
and bases in the National Survey Data Base
that describes the wastes in definite
qualitative terms. In this discussion,
dilute acidic wastes were assumed to be
1 percent sulfuric acid, and concentrated
acidic wastes were assumed to be 8 percent.
Similarly, dilute and concentrated alkaline
wastes were assumed to be 1 and 8 percent
sodium hydroxide, respectively.
In the treatment of corrosive wastes,
the wastes are segregated to facilitate
treatment as well as to avoid the mixing of
incompatable wastes and the possibility of
violent or dangerous reactions. Almost all
generic treatment of corrosive waste incor-
porates neutralization prior to other
treatment processes because highly acidic
or basic wastes interfere with or cause
failure of other downstream treatment
processes such as biological treatment or
carbon adsorption, and can cause severe
corrosion problems if the waste comes in
contact with downstream equipment.
Neutralization
Specific methods of neutralization
which can be incorporated in the treatment
scheme are: (1) mixing of acid and alkali
wastes, (2) limestone treatment of acid
wastes, (3) lime slurry treatment of acid
wastes, (4) caustic soda treatment of acid
wastes, (5) use of alkaline industrial
wastes such as carbide lime and cement kiln
dust to treat acid wastes, (6) blowing
boiler flue gas through alkaline wastes,
(7) adding compressed COg to alkaline
wastes, (8) submerged combustion to produce
COg to neutralize alkaline wastes, and (9)
adding sulfuric acid to alkaline wastes.
All of the processes have been used exten-
sively and are well-documented in the lit-
erature.
The variability of heats of neutraliz-
ation requires that careful study of the
components of the waste stream be performed
before choosing a neutralization system.
For instance, the increase in temperature
caused by neutralization reactions can re-
sult in decomposition of any nitric acid
present to toxic nitrogen dioxide gas. In
this case, measures must be taken to col-
lect and treat the gas if it is present in
harmful quantities. In addition, if any of
the wastes are strong corrosive diacid or
dibasic substances which will form insolu-
ble salts with the neutralization reagent,
the resultant heat may be high enough to
increase the corrosivity of the solution.
This should be taken into account during
system design, to prevent a materials fail-
ure. Thus, thorough analysis of the types
and quantities of wastes to be neutralized
must be made in order to avoid corrosion
of the treatment system or other adverse
consequences.
After a thorough evaluation of the
characteristics of a specific waste or
wastes, selected unit processes are incor-
porated into a treatment scheme to safely
and effectively "neutralize" the hazardous
characteristics of the waste. As an
354
-------
example of treatment for what may be called
a worst case aqueous waste, the following
generic treatment scheme is suggested.
This process is applicable to dilute and
concentrated acids and bases containing
metals involution up to several thousand
parts per million (ppm) and trace organics
at concentrations up to about 500 ppm.
This process is also applicable to cya-
nides, chelated metals, and chromic acid
(Figure 1). The process consists of two-
stage neutralization followed by a solid/
liquid separator process. The first neu-
tralization stage would control the pH
within a rather broad range and the sec-
ond would adjust the pH to a value between
6 and 9. The pH in the first stage of
neutralization may have to be raised to
between 9 and .11 to precipitate soluble
metals. The resultant sludge would probab-
ly not pass the TCLP test and would there-
fore require further treatment, such as
stabilization/solidification prior to being
disposed of in a secure landfill. The
effluent from the sedimentation tank, due
to organic and low level metal concentra-
tion, would, require further treatment.
Prior to carbon adsorption for organic
removal, the effluent from sedimentation
must have the pH adjusted to about seven
and be filtered to remove remaining sus-
pended solids. The filter backwash would
normally be recycled to the feed side of
stage one neutralization, and the spent
carbon could be thermally regenerated.
The carbon effluent could be discharged
to a municipal sewer or to surface
water.
If the corrosive feed contained cya-
nides, chromic acid, or chelated wastes,
or if these wastes were blended with the
feed, they would have to be treated sepa-
rately and prior to stage one neutraliza-
tion.
Cyanide wastes should never be mixed
with acid due to hydrogen cyanide forma-
tion, and should be treated by alkaline
chlorination. Alkaline chlorination is an
aqueous process suitable for destroying
free dissolved hydrogen cyanide and for
oxidizing all simple and most complex
inorganic cyanides. In this process, a
cyanide-bearing solution is first made
alkaline (pH above 9) and then treated
with either free chlorine or another
source of hypochlorite ion. Hypochlorite
reacts with cyanide via the following
steps:
(1) GIT H20 + OC1-
(2) Cl-CN + 20H-
(3) 20CN- + 30C1-
Cl-CN +' 20H-
OCN- + CL- + H20
2C0
N
3C1
The first step, conversion of cyanide ion
to dissolved cyanogen chloride, is very
rapid. The second reaction, hydrolysis of
cyanogen chloride, is slow at lower pH's
but rapid above pH 10 if transition metal
Applicability: Dilute and Concentrated Acids and Bases, Inorganic
•With trace organicsfcheleted waste*, oils, grease, so vents)
•Metals can bo present
• Handle chelated waste, cyanides, chromic acid soparalely
2 stage neutralization
f t
Chelated^_^ Mjx unk _
Floni.it.on _„ "odimontation S'ud98
tank ""
| Overflow
2 Stage neutralization
1 1
(with acids, Recflh/infl _ Equalization J Mix tank and c.rfw.«,.»i™ J - SIud9« _~ Secure
bases, ~~"* tank
metals)
NaOH CI2
Cyanide^ Cyanide 1
basic destruction |
Acid SO?
Chromtc__«. Chromium
acid reduction
— & mix tank 1 pHadjujtment •""""•'"•""" ^T^ dmvaterlng landfill
H,S04 H,SO« Overflow Filtrate
L*.pHadjustment|— * Filtration -* A^l'!"*d — *• mSSftumr
i carDon or surface water
1 t
Carbon
regeneration
Figure 1 (8).
355
-------
Ions are absent. In the presence of tran-
sition metal Ions, excess hypochlorite or
addition of other strong oxidants Is neces-
sary to drive the reaction at reasonable
speed. Without such additives, the reac-
tion can take several hours to go to com-
pletion. The third reaction is slow at
high pH values but rapid at lower pH values
(7). The effluent from alkaline chlorina-
tlon 1s then piped to the main treatment
stream prior to stage-one neutralization.
Chromic acid waste contains hexavalent
chromium which must be reduced to the tri-
valent form prior to stage one neutraliza-
tion precipitation. One commerically prac-
tical process is sulfur dioxide reduction.
In this process, gaseous sulfur dioxide is
added under pressure to the chromate con-
taining waste. The pH is maintained at
approximately two, and after the reaction
is complete, the effluent is fed to stage
one neutralization.
Chelated wastes contain metals which
are difficult to remove due to the presence
of chelatlng agents which, through chemical
bonding keep the metals in solution. These
wastes would probably require lime treat-
ment at a higher pH and it may be necessary
to reduce the pH and attempt to precipitate
the metals with aluminum sulfate or poly-
mers or a combination of aluminum sulfate
and polymers. The sludge from this proc-
ess, following dewatering, could be dis-
posed of in a secure landfill, assuming
passage of the TCLP, or stored for addi-
tional treatment. The effluent from this
process should be pH adjusted (^7), filter-
ed and fed to carbon adsorption for trace
organic removal.
The severity of the treatment required
is based on the specific characteristics of
each individual waste. For example, had
the previously discussed worst case aqueous
waste contained no organics, metals, chro-
mic acid, cyanide, or chelated waste, the
treatment process would be simplified to
include only two-stage neutralization,
sedimentation, and dewatering of the
sedimentation sludge. The sludge would
probably pass the TCLP test and could be
disposed of without further treatment in a
secure landfill.
Similarly, if the waste waste composed
of organic acids and bases containing met-
als, the treatment process could be modi-
fied to incorporate two-stage neutraliza-
tion, sedimentation, aerobic digestion with
powdered activated carbon, and sludge de-
watering (Figure 2). The sludge from
neutralization would probably require
additional treatment such as stabilization/
solidification prior to landfill disposal.
The sludge from biological treatment could
be incinerated and the incineration ash
disposed of in a secure landfill. The
scrubber liquor could be recycled through
the process or fed to the'aqueous treatment
of trace organic waste contained metals
(Figure 1). The overflow from biological
sedimentation, following pH adjustment (7),
could be filtered, treated with activated
carbon, and discharged to a sanitary sewer
or surface water.
For the treatment of heavy metal
sludges containing organic acids and bases,
the process would incorporate two-stage
neutralization and solids separation. In
this case, solids separation could be ac-
complished in a filter press and the filter
cake, if found nonhazardous as prescribed
by the TCLP tests, could be disposed of in
a secure landfill. The filtrate from the
press would be piped to the process for the
aqueous treatment of trace organics con-
containing metals (Figure 1).
Resource Recovery
For certain waste streams, recovery/
reuse processes are attractive alternatives
to neutralization. A suggested process for
the recovery of halogenated and nonhalogen-
ated solvents is presented in Figure 3.
The process is applicable to dilute and
concentrated solvents containing acid or
bases. Halogenated solvents would be
handled separately from nonhalogenated
solvents. The dilute or aqueous streams
would first go to an emulsion breaking
process if required. The emulsion would be
broken down by adding a highly ionized sol-
uble salt or an acid, and once the emulsion
is broken, the waste is separated into oil
and aqueous phase by air floatation.
The aqueous phase would be piped to
the aqueous treatment of trace organics
(Figure 1) for additional treatment. The
solvent phase would be combined with the
concentrated solvent and then neutral-
ized.
The choice of the neutralization agent
must be carefully considered in this in-
stance. Lime is not recommended due to
356
-------
Applicability: Dilute Organic* With Mineral Acids And Base*
•Dilute organic acidi and bases
* Metals can be present
•No cyanide and no chromic acid
2 stage neutralization
Dilute | |
organic— «•—•»• Weeewlng :L:::L^ Equalization „, Mix tank and __!
adds & tank & mix tank pH adjustment
t t t t
HaSO4 Lime HjSO, Lbm
1 PAC 1— »
pH adjuslm
JD irocnution d
t
Biological reactor
w/powdered activated —
carbon (PACI
Overflow
t
n * nitrat on
Sludne ,
ewstering
„ Recurs -,._
landfill
1
dmiurhig —
• Activated
cartion
, !
Carbon
regeneration
Ash
1
Scrubber liquor
t=|ti Discharge to
sanitary tewer
or surf see water
To aqueous
treatment
Figure 2 (8).
Nslogenated sofvemj^—
(aqueous)
Naiogenated solvents
1
1
t
f
1
Emulsion
breaking
if required
Acid or
aEum .
Emulsion
breaking - ;
if required
»
solvents
(aqutous) .
Nort-halogenated solvents
(concentrated) ~
Applicability: Ha ogenated And N on- Halogen a ted Solvent!
•Two phase systems (solvent and aquoous phase}
•Concentrated mixed so vents
•With aeidsand ba*e*
Filtration
Air
flotation
j Aqueous
_ J phase
To aqueous
treatment
~"1 Aqueous
- t phase
» 1
Air
flotation
filtration
phase
"*i$* . HJ
. -
R /
reu$&
1 1
1 Bottoms
NH4OH
[A* Seeura 1
' 1 landfill 1
fjH.OH Bottom! Scrubber To«qMtoui
I
I
Figure 3 (8).
357
-------
scaling problems in the distillation unit.
Sodium based alkalis are not recommended
either due to the fact that they form
eutetic solids in incineration and can
cause serious ash fusion and clinkering
problems. For this application ammonium
hydroxide is recommended. Following neu-
tralization, the stream would be fed to
distillation for resource recovery. The
bottoms from the distillation unit could be
corrosive and may require incineration. If
incineration of the bottoms is required,
the ash from incineration would probably
pass the TCLP and could be disposed of in a
secure landfill. The scrubber liquor would
be piped for further treatment to the
process presented in Figure 1.
A similar process is applicable to the
reuse/recovery of concentrated and dilute
oils and organics containing acids and
bases. If emulsions are present, they
would be broken down and then separated
into oil and aqueous phases in an air
flotation unit. The aqueous phase would
require additional treatment and could be
piped to the aqueous treatment of trace
organics (Figure 1). The oil phase would
be filtered and distilled. Concentrated
oils and organic streams would be filtered,
neutralized with ammonium hydroxide, and
then recovered by distillation. Similarly,
if the bottoms from distillation require
incineration, the ash from incineration
could probably be disposed of in a secure
landfill and the scrubber liquor would
require additional treatment as presented
in Figure 1.
Incineration
Incineration is an effective method of
treating wastes which are combustible as
well as corrosive. Wastes which can be
considered for incineration include organic
acids and bases or organic liquids, solids,
and sludges contaminated with strong min-
eral acids (or bases). The combustible
component may consist of oils, solvents,
hydrocarbons, tars, greases, or petroleum
residuals.
A suggested treatment scheme for the
incineration of corrosive/combustible
sludges, solids, and liquids is presented
in Figure 4. Metals can be present in the
the waste stream. Neutralization prior to
incineration is optional in this case pro-
vided the incinerator combustion chambers
are lined with an acid resistant refractory
or firebrick. The national survey data
revealed six facilities which burn D002
Corrosive Waste.
Combustible solids would probably not
be neutralized, since this would require
the addition of water to the waste stream,
reducing its heating value. Combustible
sludges would probably be dewatered in an
acid resistant (if not neutralized prior
to dewatering) filter press to 40 to 50
percent solids. The filtrate from the
A|iplicJtiilitv:CombuttibilBOrBin)cShjElB«i.So[i ind bun
• Orginic liquid* with minirtl Kidi & bttti
NjOH or „
limt iotution ~*
Indni
ration
I Scmbbw
Figure 4 (8).
358
-------
dewatering unit would be piped for addition-
al treatment to the process presented in
Figure 1.
A rotary kiln or fluidized bed incin-
erator would be suitable for handling most
of this type waste. The fluid bed is an
ideal combustor for liquids and pumpable
solids, and would be acceptable for solids
if they were reduced to a size of about 1
to 2 inches.
Flue gas- from incineration would re-
quire scrubbing. The scrubber liquor would
require additional treatment and could be
piped to the process for the treatment of
aqueous trace organics containing metals
(Figure 1). The incineration ash may be
hazardous, as defined by the TCLP, and, if
so, would require further treatment such as
stabilization/solidification prior to dis-
posal in a secure landfill.
CONCLUSIONS
Corrosive hazardous wastes are prima-
rily liquid wastes and account for almost
12 percent of total annual hazardous waste
generation in the United States. Treatment
technologies that can be used to treat
corrosives are not complex but process
selection is critical due to the impact of
the waste material on treatment hardware.
Post-neutralization residues may require
further treatment such as fixation or
incineration. The description of generic
treatment processes presented here can be
employed as screening mechanisms in treat-
ment selection for specific corrosive
wastes.
REFERENCES
1. U.S. Federal Register, 40 CFR Part
261, May 20, 1981.
2. Bromm, S.E., "USEPA's Land Disposal
Restrictions Regulatory Programs,"
in Proceedings, Hazardous Wastes and
Environmental Engineering, Cincinnati,
Ohio, May 1985.
3. Westat, Inc., "National Survey of Haz-
ardous Waste Generators and Treatment,
Storage, and Disposal Facilities Regu-
lated under RCRA in 1981," Report to
USEPA.
4. Camp Dresser and McKee, Inc., "Techni-
cal Assessment of Treatment Alterna-
tives for Wastes Containing Corrosives,"
Draft Report to USEPA, September 1984.
Friedman, D., "Development of an Organ-
ic Toxicity Characteristic for Identi-
fication of Hazardous Waste," September
10, 1985.
Treatability Manual - Volume III, Tech-
nologies for Control/Removal of Pollu-
tants, USEPA, Office of Research and
Development, September 1981.
Versar, Inc., "Technical Assessment of
Treatment Alternatives for Wastes Con-
taining Metals and/or Cyanides," Draft
Report to U.S. EPA, October 1984.
Warner, H.P. and R.A. Olexsey, "Treat-
ment Technologies for Corrosive Hazard-
ous Wastes," JAPCA 46:403 (1986).
359
-------
ENGINEERING ANALYSIS OF HAZARDOUS WASTE INCINERATION;
ENERGY AND MASS BALANCE
W. D. Clark
W. R. Seeker
Energy and Environmental Research Corporation
18 Mason
Irvine, California 92718-2798
and
C. C. Lee
United States Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
26 West Saint Clair Street
Cincinnati, Ohio 45268
ABSTRACT
A procedure exists for performing fast and accurate energy and mass balance calcu-
lations on the basis of limited input data using sound engineering principles. 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, development and demonstration (RD&D3 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 operating permits of
hazardous waste incinerators. Principles and applications of this procedure are
discussed including the evaluation of data from two full-scale field tests and one
hypothetical RD3D permit application.
INTRODUCTION
The Environmental Protection Agency
(EPA) is charged with regulating the in-
cineration of hazardous waste under the
Resources Conservation and Recovery Act
(RCRA) (7). Under current procedures the
regional permit writer must make a number
of engineering assessments during the
course of the permit application concerning
the adequacy of the design data, the con-
sistency 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.
Currently these engineering judgments are
made by the individual permit writers upon
considering the design data and trial burn
performance' data submitted by the appli-
cant. In many cases, the permit writers
must employ manual calculations on the
incinerator system design and performance
data in order to check the data consistency
and to aid in the decision making process.
These manual calculations are tedious and
cannot be done with sufficient accuracy.
Engineering analysis procedures exist,
as described by Clark et al. (1,2), to
allow these calculations to be made in
great detail; however, these procedures
require a significant amount of input data
which is often not available and the pro-
cedures are too complex to be used by any-
one other than experienced personnel. A
critical need for the EPA is a standardized
procedure which allows the permit writer to
rapidly and accurately perform energy and
mass balance calculations based on limited
permit data. Such an analytical tool would
greatly aid the permit writers in making
the regulatory decisions for all phases of
the permit procedures: research develop-
ment and demonstration (RD&D), construc-
tion, and full permit.
360
-------
The energy and mass balance must be
able to perform the following types of
analysis:
• RD&D and Construction Permits:
Ensure that all necessary
engineering data is provided
in the permit application.
Determine whether the design
specification and operating
parameters are consistent
with energy and mass
balances.
Determine whether the design
is capable of attaining the
design temperature.
• Full Permit
Determine if data in lieu of
trial burn data is
sufficient.
Determine the appropriate
operating conditions from
the trial burn conditions
through evaluation of
parameter effects.
Determine the sensitive
parameters that should be
monitored for ensuring
continued compliance.
Under contract to the EPA, Energy and
Environmental Research (EER) is in the
process of developing an energy and mass
balance analytical procedure that can be
used for regulatory decision making. The
approach to the development of the energy
and mass balance procedure is shown in
Figure 1. Development will occur in stages
of increasing complexity with each stage
being verified against and if necessary
modified in response to actual test
results. User friendly software will be
developed and documentation and training
sessions will make the procedure useful to
the permit writers. The first stage
applies to conventional fuel-lean systems
including most liquid injection incinera-
tors, rotary kilns and afterburners.
Updates will be provided in subsequent
stages to include fuel-rich pyrolysis sys-
tems, alternative heat sources, and more
detailed analysis of air pollution control
devices (APCD). At this time, stage 1 of
the program is nearing completion.
1, Develop Energy and Mass Balance
Software for Conventional Systems
2. Test Case Runs for Actual Permits
3. Develop User-Friendly Software
and Documentation
4. Training Sessions for Permit
Wr1ters
5. Pyrolysis Systems Updates
6. Alternative Heat Sources Updates
7. Detailed APCD Analysis Updates
Figure 1. Technical approach.
This paper reviews the energy and mass.
balance procedures and discusses the appli-
cation of those procedures to two inciner-
ation field tests and to a hypothetical
incineration problem.
ENERGY AND MASS BALANCE PROCEDURE
The energy and mass balance procedure
is based on standard chemical engineering
practices. Energy and mass balances are
performed, based on incinerator design
specifications and operating conditions
with the primary objectives of calculating
excess air levels,
temperatures,
residence times, and
total volumetric flows
for each unit of the incineration system.
A unit consists of a separate combustion or
thermal chamber. For example, a rotary
361
-------
kiln/afterburner system consists of two
units 1n series with the output from the
rotary kiln serving as input to the
afterburner.
As shown in Table 1, inputs to the
procedure include feed rate, temperature,
heating value, and composition of all input
streams to each unit including wastes,
fuels, water, air, and oxygen; incinerator
design specifications including the thick-
ness and conductivity of the refractory,
the volume of the unit, the area of the
refractory and any cooled surfaces, and the
outer shell temperature; and the APCD
design specifications including gas volu-
metric capacity, acid capacity, quench
water capacity and temperature, and the
temperature to which the gas must be
quenched. If unknown, many of these quan-
tities can be estimated based on common
Incineration practices and sound
engineering Judgement.
The mass balance is based on simple
stolchiometric calculations. In the cur-
rent version of the procedure, complete
combustion is assumed with the only prod-
ucts being Cf»2, ^0, HC1, S02, ash, N2, and
excess Og. The mass balance determines if
sufficient oxygen is available for complete
combustion, calculates the composition of
the combustion products, and calculates the
total mass flow through the incinerator.
The energy balance solves two
equations simultaneously: (1) balancing
sensible heat, heat of vaporization,
chemical heat, radiation, and convection;
and (2) balancing radiation, convection and
conduction. Sensible heat is calculated
from the mean heat capacities, tempera-
tures, and mass flows of all input streams
and the combustion products; heat of
vaporization is calculated from the
moisture content and mass flow of any
aqueous streams and the heat of vapori-
zation of water; chemical heat is
calculated using the mass flow and heating
value of fuel and waste streams; radiation
is discussed in detail in the paragraph
that follows; convection is based on
TABLE 1. SAMPLE INPUT DATA SHEET FOR PLANT F PRIMARY INCINERATOR
I PLACE OF UNIT IH SERIES' I 1 I
AUXILIARY
FUEL
WASTE A
WASTE B
WATER
AIR
OXYGEH
ENRICHMENT,
HAZARDOUS
COMPOUNDS
PHENOL
MUKtl OHIt l«tt*t
CMIOH Ht««l«.«l«
niCWWDCTKlKllt
STREAM
NUMBER
-
A
B
-
C
-
-
A/B
A/B
A/8
A/B
A/B
FEED
RATE
(Q/SEC)
0
SB. 4
89.8
0
m
0
-
3 flfl
0.67
0.42
0,31
0.26
PREHEAT
-
(=2985!
iilpll
Hill
-
-
PROXIMATE ANALYSIS
(AS RECEIVED)
w
-
11
it!
liil
i°l
VK tines
-
Hi
111
mum
ASH
-
2.13
0.78
MOISTURE
-
2.18
93,89
HEATING
VALUE
(HIGhER AS
RECEIVED)
(3/0)
-
34,390
1,696
ELEMENTAL ANALYSIS
(DRY)
C
-
72.20
•SSB-fe
Sup
H
-
8.29
mm
ilUO
N
(X)
-
0.67
mm
fliflO
sste
s
-
ai4
aoo
%8®
fffKi
ASH
(1)
-
2.20
12.7<
0
(%)
-
15.20
>'QM
Cl
(X)
-
1.30
5.89
INCINERATOR DESIGN SPECIFICATIONS
REFRACTORY THICKNESS
^6'l8%
9.03
24.15
0,00
APCD DESIGN SPECIFICATIONS
APCD GAS CAPACITY (sm3/see>
APCD HCL CAPACITY <9/sec)
QUENCH WATER AVAILABLE (liters/sec)
QUENCH WATER TEMPERATURE (°K3
QUENCHED GAS TEMPERATURE ' (°K)
_
«
_
-
- ESTIMATED FROM COMMON INCINERATION PRACTICE
- INFERRED FROM INPUT DATA
362
-------
Kroll's correlation for tubular combustors
(4); and conduction is based on the
refractory conductivity and thickness and
the temperature difference between the
inner and outer wall.
Radiation is the dominant mode of heat
transfer at typical incineration tempera-
tures. The equation for gas radiation is
pg = eg
AT
where P is power, e is emissivity, a is the
Stef an-Bol tztnann constant, T is tempera-
ture, g denotes gas, and the total area,
AT, is the sum of the refractory, cooled,
and open areas. In the energy balance a '
speckled wall approach, as described by
Hottel and Sarofim (5) is used in which the
various surfaces are assumed to be evenly
distributed throughout the incinerator with
incident radiation distributed accordingly.
In the present version of the model open
surfaces are assumed to be cold and totally
absorbing. Refractory and cooled surfaces
are assumed to absorb a certain fraction
{typically 80 percent) of input radiation,
reflecting the rest to be reabsorbed or
transmitted by the gas. Reflected radia-
tion is followed until it is reduced to
less than one millionth of the initial
value. The absorptivity of the gas is
assumed to be the same as its emissivity
which is calculated by Johnson's gray gas
approach (6) as described by Richter (8)
based on the temperature and composition of
the gas and on the mean beam length of the
incinerator. A certain fraction (typically
two percent) of the volatile carbon is
assumed to form soot which contributes to
the gas emissivity as described by Sarofim
(9). The refractory is assumed to radiate
according to the equation
PR =
AR
where R denotes refractory and where again
ER is typically 80 percent. Radiation from
cooled surfaces is assumed to be negligi-
ble. Net radiative heat transfer to each
surface is calculated as the difference
between absorbed and emitted radiation.
The two energy balance equations are
solved simultaneously by a Newton Raphson
technique to calculate the mean gas and
refractory temperatures. The gas tempera-
ture and composition determines its density
which, along with the total mass flow and
the volume of the incinerator, determine
the mean residence time.
Conditions at the exit of the last
unit are used to define the flow conditions
to the APCD. Mass flow, temperature, and
gas composition determine the quench water
requirement, the total volumetric flow, and
the acid flow to the APCD.
The Program is written in Fortran,
stored on a 360 K byte diskette, and can be
run on an IBM PC with 256 K bytes of memory
in less than 30 seconds.
APPLICATIONS
In order to demonstrate how the energy
and mass balance procedure can be applied,
three cases which may face a permit writer
will be presented. These test data come
from both operating incinerators and a
hypothetical example.
Field tests were conducted by MRI (10)
on a liquid injection/secondary incinera-
tion system (designated as Plant F) which
is a 2.5 MW (8.5 MMBtu/hr) unit burning
organic and aqueous liquid wastes. Table 1
shows input data for the primary for one
particular test condition. Portions of the
table shaded with dots indicate inputs
which were not reported or directly implied
from information available in the MRI
report. These values were estimated from
typical incineration practices. Portions
of the table shaded with lines indicate
inputs which were not reported but were
inferred from reported conditions. The air
feed rate was chosen to match the reported
oxygen concentration, the refractory
thickness (hence the unit volume and the
refractory surface area) was chosen to
match the reported residence time; and the
refractory conductivity was chosen to match
the reported temperature. All of these
inferred input values were reasonable and
were with in the bounds of common
incineration practice.
Table 2 shows that it is possible to
achieve excellent agreement between energy
and mass balance predictions and measure-
ments for the Plant F incinerator. This
cannot be considered a rigorous verifica-
tion of the model because of the incomplete
input specifications; however, it is an
indication that, with reasonable assump-
tions for the unknowns, the model and the
measurements are consistent. ORE predic-
tions were made using first-order post-
flame decomposition rates of Dellinger et
al. (3) assuming the waste remained at the
mean primary gas temperature for the mean
363
-------
TABLE 2. ENERGY AND MASS BALANCE RESULTS
FOR PLANT F
3==*=====================:================
PREDICTED MEASURED
TEMPERATURE
Primary
Secondary
CONCENTRATION
02
COg
RESIDENCE
TIME
13670K
1348°K
10.5%
8.5%
2.4 sec
'
1367QK
13390K
10.5%
7.6%
2.4 sec
percent of its input combustion heat
through the walls of each unit. Thus, the
wall temperatures are only slightly below
the gas temperatures.
Trial burn measurements are not always
consistent with energy and mass balances.
/-
^H
0.04MW
\ZJ
1328
1348QK
912$ Excess Air
1.0 sec
ORE FOR CC14
>99.99%
99.998%
primary residence time and at the mean
secondary gas temperature for the mean
secondary residence time. Results of the
energy and mass balance are shown in more
detail in Figure 2. The incinerator
operates with almost twice the amount of
air required for complete combustion and it
is well -insulated, losing only about two
liquid
Wail.
5lo.uo.
1355°K-
w—--
1367°K
912S Excess Air
1.4 sec
0.05MW //
.7:' .J
2.5
MW
Figure 2. Detailed predictions for
Plant F.
.... Si at 09* Storog* —
nh=i r.__,
Addod
icdcd
Solid!
Figure 3. Plant H incineration system, reprinted from Reference (10).
364
-------
TABLE 3. ENERGY AND MASS BALANCE RESULTS FOR PLANT H
MEASURED
PREDICTED
AS REPORTED
PREDICTED WITH
ASSUMED LEAK
TEMPERATURE
Primary
Secondary
STACK FLOW
CONCENTRATIONS
Stack D£
Stack C0£
Secondary Og
Secondary C02
ORE FOR CC14
13830K 10010K
1367°K 1040°K
13830K
1367OK
3.0 sm3/sec
13.0*
6.2%
>99.99%
Field tests were conducted by MRI (10) on a
solid and liquid incinerator and after-
burner system (designated as Plant H and
shown in Figure 3) for a 3.2 MW (11.1 MM
Btu/hr) trial burn of aqueous, liquid
organic, and solid wastes supplemented with
fuel oil. As with the Plant F burn, not
all of the necessary input data were re-
ported; however, in this case no reasonable
assumptions could make the measurements
consistent with a simple energy and mass
balance. The reported oxygen concentration
of 13 percent is inconsistent with the
reported secondary temperature of 1367°K
which is much higher than the adiabatic
flame temperature at these conditions. In
an attempt to explain this discrepancy, it
was assumed that a substantial fraction
(more than half) of the air entering the
incinerator was due to unreported leakage
of air into the system which occurred
between the secondary and the stack. As
shown in Table 3, all of the data co.uld be
rationalized with this assumption. MI
(10) also noted the temperature/oxygen-
concentration discrepancy and formed simi-
lar conclusions concerning air leakage.
If indeed such a leak existed, stack
measurements would inaccurately reflect
primary and secondary temperatures and
excess air levels and would be insensitive
indicators of incinerator performance. The
effect of excess air was investigated with Figure 4.
the energy and mass balance model by vary-
ing the primary air flow about the baseline
2.8 sm3/sec 2.8 sm3/sec
13.0%
5.7%
13.0*
,5.7%
97%
13. OS
5.7%
7.7%
9.6%
>99.99«
conditions with the assumed leak listed in
Table 3. Figure 4 shows that the safe '
operating range of the Plant H incinerator
is predicted to be between secondary excess
air levels of 20 and 78 percent. At excess
air levels lower than 20 percent the possi-
bility exists for incomplete combustion and
waste destruction due to the escape of
Stack Oxygen Concentration, % Dry
11 12
Secondary Excess Air, %
at Constant Load
Effect of excess air on incin-
erator temperature and stack
oxygen concentration.
365
-------
Total Heat Load, MW
at Constant Excess Air
Figure 5. Effect of load on incinerator
temperature and APCD loading.
fuel-rich packets, and at excess air levels
above 78 percent the mean temperature is
below Tgg gg i sec for CCU, the tempera--
ture required to destroy 99.99 percent of
the compound within one second based on the
kinetic rates of Dellinger et al. (3).
Stack oxygen concentration for this range
changes only from 11.7 to 13.6 percent due
to the leakage air. Thus, it would be
difficult to detect a failure in this
incinerator due to too much or too little
excess air simply by monitoring the stack
oxygen concentration.
The energy and mass balance model was
used to estimate the safe operating range
of the Plant H incinerator with respect to
thermal input and waste water content. As
the thermal input to the incinerator de-
creases, a larger proportion of the heat is
lost through the walls causing the mean
temperature to drop. Figure 5 shows that
at loads below 0.6 MW, the mean gas temper-
ature is predicted to be too low to destroy
CC14 and at loads above 4.1 MW, the total
volumetric throughput is predicted to be
greater than the assumed limit of the APCD
(Actual APCD limits were not reported).
Thus, the operating envelope indicated by
energy and mass balance would be between
0.6 and 4.1 MW.
The second operating condition inves-
tigated was the water content of the waste.
As the water content of the waste in-
creases, the heating value decreases
causing the temperature to drop. Figure 6
shows that for the composite waste fed to
the Plant H incinerator, if the waste water
content is greater than 57 percent (corres-
ponding to a waste heating value below 7500
J/g) the mean gas temperature is predicted
to be too low to destroy CC14- If the
Waste Heating Value, J/g
5000 10,000 15,000
Water Content of Waste
at Constant Mass Loading
Figure 6. Effect of waste heating value
on .incinerator temperature and
APCD quench water requirement.
366
-------
1.26 Kg
Contamln;
Combustion!
Air 1
kj
4.5 HW Ji
Natural Gas I--J
/sec
ted Soil
t—- --^
n
•/)
11259°K (Wall) ^ /
1 2.3m . 11
1268°K
I ^ i ««-. If
Burner
100K Excess Air
4.6m
Figure 7. Baseline incinerator for oxygen
enrichment example.
water content is below 14 percent (corres-
ponding to a waste heating value above
14,000 J/g) the gas temperature entering
the APCD is predicted to be too high to be
quenched to the expected limit temperature
with the assumed quench water capacity
(again, actual values were unreported).
The appropriate operating range for water
content of the waste stream would therefore
be between 14 and 57 percent. A safer
operating range could be selected within
this full range.
The previous two examples have used
actual field test data to demonstrate that
the energy and mass balance model can be
used to assess the consistency and com-
pleteness of data, to identify sensitive or
insensitive monitoring parameters, and to
estimate safe operating ranges. This can
be useful to the permit writer in evaluat-
ing operating permits and in establishing
operating limits based on trial burn data.
In the next example, a hypothetical incin-
erator will be used to show how the energy
and mass balance program can be used to
help assess the feasibility of a new or
innovative concept, aiding the permit
writer in evaluating RD&D permit
applications.
The hypothetical incinerator, shown in
Figure 7, has been designed to process soil
contaminated with 0014. The soil has
little heating value, so 4.5 MW natural gas
is required as auxiliary fuel. Tests have
shown that the incinerator can achieve its
design temperature of 1268°K, the tempera-
ture necessary to achieve 99.99 percent
CC14 destruction in one second (3); how-
ever, the incinerator fails to achieve
99.99 percent CC14 destruction due to the
unanticipated entrainment of contaminated
fine particles. Based on the size distri-
bution of the entrained fines, it has been
estimated that entrainment can be virtually
eliminated if the bulk gas velocity can be
cut in half from 3.88 to 1.94 m/sec.
Oxygen enrichment has been proposed as
an innovative solution to this problem. By
replacing a portion of the combustion air
with oxygen, the same excess oxygen level
can be maintained with reduced total
oxidant flow, and because of the decrease
in nitrogen diluent, less auxiliary fuel is
required to maintain the same temperature.
This reduces the bulk gas velocity and
decreases the auxiliary fuel requirement to
help offset the costs of oxygen and system
modifications. Figure 8 shows the energy
and.mass balance prediction of the effect
of oxygen enrichment. Bulk gas velocity
80
Auxi1i ary Fuel, g/sec
60 40
20
u
0)
O o
O i
ISl
10
CD
1/2
Baseline
Velocity
\
Temperature - 1268 K
0 50 100 150
Oxygen Enrichment, g/sec
Figure 8. Effect of oxygen enrichment
on bulk gas velocity.
367
-------
can be cut "In half with the addition of 140
g/sec oxygen while maintaining the same
temperature and excess oxygen level. This
is accompanied by a reduction in auxiliary
fuel from 82 to 32 g/sec and by enrichment
of the combustion air to 40 percent oxygen.
Thus, on the basis of energy and mass
balance, oxygen enrichment is a feasible
concept for reducing fine particle entrain-
ment in this system. The permit writer can
then decide that further investigation
through the issuance of an RD&D permit is
safe and warranted.
SUMMARY
The energy and mass balance procedure
is based on sound engineering principles.
It has been demonstrated to be useful in
evaluating the feasibility of incinerator
designs and concepts and in evaluating the
consistency of trial burn measurements.
The procedure has been used to interpolate
and extrapolate data to estimate safe
operating ranges and to identify sensitive
monitoring parameters. Field test data are
often insufficiently reported to allow
rigorous energy and mass balance calcula-
tions to be performed; nevertheless sound
engineering judgement can generally be used
to fill in the blanks to allow assessment
of the consistency of the data.
REFERENCES
1. Clark, W. D., M. P. Heap, W. Richter,
and W. R. Seeker, 1984. The
Prediction of Liquid Injection
Hazardous Waste Incinerator
Performance, ASME Paper 84-HT-13.
2. Clark, W. D., J. F. La Fond, D. K.
Hoyeda, W. F. Richter, and C. C. Lee,
1985. Engineering Analysis of
Hazardous Waste Incineration, Failure
Mode Analysis for Two Pilot Scale
Incinerators, Presented at the llth
Annual Research Symposium on Inciner-
ation and Treatment of Hazardous
Waste, Cincinnati, Ohio.
3. Dellinger, B., J. Torres, W. Rubey, D.
Hall, and J. Graham, 1984. Determi-
nation of the Thermal Decomposition
Properties of 20 Selected Hazardous
Organic Compounds. Proceedingsof the
Tenth Annual Research Symposium on
Incineration and Treatment' of
Hazardous Waste, EPA Report to.
600.9-84-022, pp. 116-123.
4. Field, M. A., D. W. Gill, B. B.
Morgan, and P. G. W. Hawksley, 1967.
The Combusti on of PulveHzed Coal, The
British Coal Utilization Research
Association, Leatherhead, p. 119.
5, Hottel, H. C. and A. F. Sarofim, 1967.
Radiative Transfer, McGraw-Hill, New
York, p. 298.
Johnson, T. R. and J. M. Beer, 1973.
The Zone Method Analysis of Radiant
Heat Transfer: a Model for Luminous
Radiation. J. Inst. Fuel, 46_, p. 388.
Resources Conservation and Recovery
Act, 1976, Public Law 94-580.
Richter, W., 1981, Assessment of
Pulverized Coal Fired Combustor Per-
formance - Models for Coal Combustor
Performance, Analytical Tool Verifi-
cation, Topical Report, DOE Contract
No. DE-AC-80PC30297.
Sarofim, A. F. and H. C. Hottel, 1978.
Radiative Transfer in Combustion
Chambers: Influence of Alternative
Fuels, Proceedings of the 6th
International Heat Transfer
6.
7.
9.
10.
Conference.
Pub!i shers,
Vol. 6, Hemisphere
Washington, pp. 199-217.
Trenholm, A., P. Gorman, and G.
Jungclaus, 1984. Performance
Evaluation of Full-Scale Hazardous
Maste Incinerators - Volumes I-V,
Final Report for EPA Contract No.
68-02-3177, Midwest Research
Institute, Kansas City, Missouri.
368
-------
THE FORMATION OF PRODUCTS OF INCOMPLETE COMBUSTION
IN RESEARCH COMBUSTORS
George L. Huffman and Laurel J. Staley
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Experiments were conducted in which various organic chemicals were burned, generally
one at a time, along with a diluent fuel, heptane, in a pilot-scale combustor (the Tur-
bulent Flame Reactor, TFR) at EPA's Center Hill Facility. Excess air levels were varied
in the TFR and Tenax trap samples were taken to determine the accompanying Destruction
and Removal Efficiencies (DREs) of the organic chemicals and the Products of Incomplete
Combustion (PICs).
Results showed that PICs did not always occur predictably, but that they were in-
fluenced by the Principal Organic Hazardous Constituent (POHC) being burned. When appre-
ciable PICs were formed however, their emission levels generally increased with increas-
ing levels of excess air, over the range of excess air levels investigated. Furthermore,
the highest levels of PICs and unburned POHCs generally occurred at high excess air
values.
INTRODUCTION
The research described in this paper
was done to identify certain sets of
operating conditions where Products of
Incomplete Combustion (PICs) form [re-
sulting from the burning of selected
chemical compounds representing typical
Principal Organic Hazardous Constituents
(POHCs) of simulated hazardous wastes]
and to find the regimes where these PICs
are more prevalent. There is general
concern on the part of the public regard-
ing the nature and emission levels of
PICs that might emanate from any hazard-
ous waste incinerator that Is to be loca-
ted in their neighborhood. The research
described herein is the result of the
initial efforts of EPA's in-house staff
to conduct experiments In their Thermal
Destruction Laboratory located at EPA's
Center Hill Facility in Cincinnati, Ohio.
EXPERIMENTAL EQUIPMENT
The Turbulent Flame Reactor (TFR)
shown in Figure 1 was the experimental
combustor used. A two mole percent
mixture of either 1,1,2-trichloroethylene
(TCE), carbon tetrachloride, 1,1,2,2-
tetrachloroethane, chlorobenzene or
Freon-113 (1,1,2-trichloro, 1,2,2-tri-
fluoroethane) in heptane was delivered to
the TFR via three nitrogen-pressurized,
one-gallon fuel tanks. The fuel/waste
mixture enters the TFR at the bottom of
the reactor through a pressure-atomizing
Delavan nozzle having a flowrate of 0.5
gph at 120 psi nitrogen pressure. Com-
bustion air supplied by an air compressor
enters the TFR through an International
Flame Research Foundation adjustable
swirl windbox. Gears on the bottom of
the windbox allow the Swirl Number to be
adjusted from 0 to 2.7.6. A water jacket
using cold city-water removes much of the
heat generated by the TFR flame and, in
so doing, quenches most of the post-
369
-------
CJ
"M
O
FLOW CONTROL
BAFFLE
AIR BLOWER
EXHAUST GAS TO SCRUBBER
.-WATER INLET
-WATER OUTLET
STAINLESS STEEL
WATER-COOLED CHAHBER
- TURBULENT SPRAY FLAME
|-«- QUARTZ WINDOW
QUARL
-SPRAY NOZZLE
PRESSURIZED
SWIRL VANE
FUEL/WASTE
RESERVOIR
Figure 1. Turbulent flame reactor (TFR)
-------
flame reactions. Because of this, the
Products of Incomplete Combustion (PICs)
are expected to be maximized (1).
Exhaust gases are sampled at the exit
port atop the TFR. Three sampling probes
are used. One is used for ORE determina-
tions via sorbent trap sampling. Another
is used for the continuous monitoring of
CO, carbon dioxide, and oxygen. The
third is used to collect sample for
continuous total hydrocarbon measurement.
For these tests, volatile POHCs and
PICs were collected on sorbent tubes
filled with Tenax-SC resin. Exhaust
gases were removed from the exit port- of
the TFR and passed through a heated
length of 1/4 inch (outside) diameter
teflon tubing kept at approximately
130°C. The heated portion of the sample
line ended six inches upstream of the
tenax-filled sorbent tube in order to
allow the sample gases to air-cool from
130°C to about 24°C (or less) while
passing through a particulate filter. .
The exhaust gases then passed through 'the
sorbent tube packed with 1.5 grams of
Tenax-GC resin, a diaphragm pump, a rota-
meter and a dry gas meter and were ,
finally exhausted to the atmosphere.-
Exhaust gas temperature was measured at
the exit port of the TFR (near where the
samples were taken) using a Type K
(chrome!-alumel) thermocouple (2).
The sorbent trap samples collected
in this way were analyzed by a gas chro-
matograph (GC) equipped with a Hall
Detector. The GC was set up to detect
only a limited number of PICs; they were:
TCE (C2HC13), carbon tetrachloride (CC14),
1,1,2,2-tetrachloroethane (Cg^C^h
chlorobenzene (CgHsCl), chloromethane
(CH3C1), methylene chloride (CHgClg)
and chloroform (CHC13).
DESCRIPTION OF TESTS
For these experiments, a two mole
percent mixture of either 11,1,2-trichloro-
ethylene (TCE), 1,1,2,2-tetrachloroethane,
carbon tetrachloride, chlorobenzene, or
Freon-113 (1,1,2-trichlo'ro, 1,2,2-tri-
fluoroethane) in heptane was burned in
the TFR. During each of these five
tests, the excess air level was varied
from 120% of theoretical air to 240% of
theoretical air. At five, or sometimes
six, sets of operating conditions, Tenax
trap samples were taken to determine
levels of POHC destruction and PIC for-
mation. To test mixture effects, a sixth
test was conducted in which both chloro-
benzene and Freon-113 were burned in
heptane under conditions similar to those
used in the first five tests.
- At one set of operating conditions
during most of the tests, the flame was
deliberately extinguished by shutting off
the flow of fuel to the TFR two minutes
into a five-minute sampling run while an
actual exhaust sample was being taken.
Sampling continued for the remaining
three minutes in order to determine what
levels 'of PICs and unburned POHCs were
emitted during this simulated "upset"
condition (2).
RESULTS
The highest levels of volatile POHC
and PIC emissions generally occurred
under high excess air conditions.
Emissions of the,POHCs .1,1,2-trichloro-
ethylene, carbon tetrachloride and
1,1,2,2-tetrachloroethane increased with
increasing excess air levels as shown in
Figures 2, 3 and 4. This observed in-
crease in emissions could be because the
temperature inside the TFR dropped rough-
ly 100°C in changing from very low excess
air conditions to very high excess air
conditions. As shown in the tetrachloro-
ethane burn of Figure 4, emissions of
both POHC and PICs (especially the pre-
dominant PIC, TCE), increased with in-
creasing excess air levels. In fact,
POHC and PIC emissions at 192% of theo-
retical air exceeded those measured
during a flameout (at 182% of theoretical
air). The emissions of the PIC 1,1,2-tri-
chloroethylene were consistently higher
than those of the POHC, 1,1,2,2-tetra-
chloroethane. Even though the ORE for
1,1,2,2-tetrachloroethane was above
99.99% at all but the highest excess air
levels, PIC emissions were relatively
high. This result is significant because
it indicates that achieving 99.99% ORE
does not necessarily guarantee that
emissions are minimized (2).
The POHC and PIC emissions from the
chlorobenzene burn are shown in Figure 5,
and from the chlorobenzene/freon burn in
Figure 6. Chlorobenzene emissions tended
to be high when the flame was turned off.
However, it should be pointed out that,
371
-------
2.8-
2.6-
2.4-
2.2-
°o 2.0-
X
§ 1'8'
i 1.4-
Bl
§ 1.2-
Q.
S 1.
| 0.6
10 Q,4 "
c
o
S d.2.
I
ixaxB,
m « 1,1,4- T«tchioto«t*ji«»« (ti» 21 row)
i • 1
« - OS3.C1 I \
1
, -cacij 1
•^ • CC14 J
* * Chlorobtniene 1
• - Conflucoe* of M*ny Data Point* I
P.O. - riJEM Oat 1
. *
1 I
i
0
• °
1
u
125 13S t45 1SS
% Theoretical Air
: Figure 2. POHC & PIC emissions for the TCE burn.
165
372
-------
1.S-
1.4.
1,3
1,2-
1.1 •
1 .
* 0.9
01
£ 0.8
8
£0.7
O)
O 0.6
O 0.5
O
jg 0.4
10.3
g 0.2
1 0.1
I
ui o
LECIBJ!
• - Carbon T«tr«:Uorlde (th« 21 POHC)
» - CH3C1
* - CHClj
» - ICE
• * Confluence of Many C*ta Fota
120
140 160
X Theoretical Air
180
Figure 3. POHC & PIC emissions for the CC14 burn.
373
-------
I
D>
O
0.
t.
o
o
o
a.
i
n
c
o
Ul
120
CO level 500 ppm
UOBTOi
m - l,l,i,Z-T«««c»iloroeth»oe (th* 21 POHC)
* - cicij
, . CBjCll
• - CC14
— - chlorobenzene
. Confluence ol M»ny D.t« Point"
F.O* H FlJMfl Out
CO level 1000 npm
emission levels
lor 99,99% ORE
140 160
% Theoretical Air
Figure 4. Emissions for the tetrachloroethane burn.
200
2
1.91
.1.6
J1.4
gl,3
E1.2
o>
| OJ:
50.8
gO.6
10.5
J?Q.4
eO.3-
I0'2'
lo.ij
LU 0
i (th* :
I - CHjCl
« - C32C12
^ - Confluence id H«ny Data !
F.O. " Flaa* Out
KO.
130
150
170 190 210
% Theoretical Air
230
Figure 5. POHC & PIC emissions for the chlorobenzene burn.
374
-------
• Chlorobeoz«Dc (the 2S POUC)
N~
s
M
tfi
(0
•g
O
E
O!
*.,
U
Z
^
o
U
X
o
o.
IA
•5"
»
e
.2
w
I
in
32
30.
28
26-
24
22-
20
18-
16 •
14 •
12-
10-
8-
6 -
4 -
2-
» T» " t**o* (» 21 rose)
h I -
-------
ANALYSIS OF PIC AND TOTAL MASS EMISSIONS
FROM AN INCINERATOR
Andrew R. Trenholm
MIDWEST RESEARCH INSTITUTE
Kansas City, Missouri
C. C. Lee
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio
ABSTRACT
The destruction of Resource Recovery and Conservation Act (RCRA), Appendix VIII com-
pounds in hazardous waste incinerators and their residual stack emissions have been
studied extensively. Complete characterization of organic compounds emitted in all ef-
fluents from an incinerator has not been conducted. This paper presents study plans and
preliminary results of such a characterization, to the extent that the emitted compounds
can be identified and quantified. Measurements will be made of both Appendix VIII and
non-Appendix VIII compounds in all effluents (stack, ash, water, etc.) from an inciner-
ator operating under both steady state conditions and during typical transient condi-
tions. A broad array of sampling and analysis techniques will be used. Sampling
methods include Modified Method 5, volatile organic sampling train (VOST), and specific
techniques for compounds such as formaldehyde, and analysis techniques include gas
chromatography (GC), gas chromatography/mass spectrometry (GC/MS), and high performance
liquid chromatography (HPLC). Continuous measurements will also be made for a variety of
compounds including total hydrocarbons by flame ionization detector (FID).
INTRODUCTION
Proper handling and disposal of haz-
ardous wastes continues to be a major con-
cern of agencies charged with protection of
public health and the environment. As
greater restrictions are placed on the land
disposal of these hazardous wastes under
the Resource Conservation and Recovery Act
(RCRA), more attention will be focused on
hazardous waste disposal via combustion
mechanisms. The Environmental Protection
Agency's (EPA's) Hazardous Waste Engineer-
ing Research Laboratory (HWERL) has the
responsibility to provide information on
the ability of these combustion systems to
dispose of hazardous wastes in a manner
that provides adequate protection of the
public health and welfare. Past HWERL
studies in this area have primarily exam-
ined the performance of combustion systems
relative to the destruction and removal
efficiency (DRE) for RCRA Appendix VIII
compounds in the waste feed.
These earlier studies demonstrated
that in general most facilities performed
quite well relative to the DRE. However,
the studies themselves and subsequent re-
view of these studies by EPA's Science
Advisory Board raised additional questions
about the performance of hazardous waste
combustion systems. One major issue that
surfaces was the question of what addi-
tional Appendix VIII or non-Appendix VIII
constituents that were not identified in
the .earlier tests might be emitted from
hazardous waste combustion. The issue
376
-------
addresses not only stack emissions, but all
other possible effluents such as organics,
trace metals, and other chemicals associ-
ated with incinerator ash, spent water, and
particulates. Because these effluents may
be equally or more hazardous than the spe-
cific Appendix VIII compounds in wastes
that are regulated by RCRA, research is
needed to qualitatively and quantitatively
study the characteristics of all possible
effluents and to provide engineering data
for regulatory support. To address this
issue, EPA has initiated a project to
quantify total mass effluents (TME) from
a commercial incinerator.
The TME project will apply a broader
array of sampling and analysis techniques
than used in previous tests to characterize
to the extent possible all effluents from
an incinerator. The difference in emis-
sions between steady-state operation and
operation with typical transient conditions
(e.g., burner adjustment) will also be
studied. Selection of the site for the
test and preparation of the test protocol
are presently being finalized. However, a
preliminary assessment was made of com-
pounds emitted from stacks on combustion
devices using currently available data.
This assessment is the subject of the re-
mainder of this paper.
Approach to Preliminary Assessment
The goal of this assessment was to re-
view organic emissions from hazardous waste
combustion device stacks to the extent pos-
sible with available data from full-scale
tests. The devices included incinerators,
industrial boilers, and mineral process
kilns. A limited comparison was also made
with emissions data for coal-fired power
plants and municipal incinerators. Most of
the data available are from samples col-
lected using either a Modified Method 5 or
volatile organic sampling train and ana-
lyzed by gas chromatography/mass spectrom-
etry (GC/MS). Practically all of the data
are for RCRA Appendix VIII compounds only.
Thus, the study was constrained by these
data limitations.
The primary data source was HWERL-
sponsored tests at eight hazardous waste
incinerators, nine industrial boilers that
cofired hazardous wastes, and five mineral
processing kilns that fired hazardous
wastes as fuel. In addition, semivolatile
organic emissions data for two municipal
solid waste (MSW) incinerators and seven
coal-fired power plants were also reviewed
for comparative purposes.
All of the data referred to above were
obtained from published reports. Limited
additional data were obtained on non-
Appendix VIII compounds by an intensive re-
view of the GC/MS data from five specific
test runs. These tests included runs from
three incinerator tests, one MSW inciner-
ator test, and one coal-fired power plant
test. This review allowed estimation of
some organic constituents not on the
Appendix VIII list, which are detected by
GC/MS.
All of the emission rate data were re-
duced to common units to allow comparison
across the data. One of the units chosen,
nanograms of emissions per kilojoule of
heat input (ng/kJ), is used to present re-
sults in this paper. This unit was se-
lected because it normalizes emissions
relative to size of the facility and allows
comparison among different types of combus-
tion devices. For perspective, values of
0.1 to 1.0 ng/kJ were usually near the de-
tection limit of the sampling and analysis
methods used.
The assessment of these data included
a search for general trends and several
specific comparisons. The comparisons in-
cluded: (a) incinerators versus boilers
versus kilns; (b) total quantity of spe-
cific compounds identified versus total
hydrocarbon; and (c) hazardous waste com-
bustion versus nonhazardous waste combus-
tion.
RESULTS
A review of the available data showed
that 62 compounds (32 volatile and 30 semi-
volatile) were detected in the stack emis-
sions from combustion of hazardous waste.
These compounds were emitted at rates that
span over five orders of magnitude (0.09 to
13,000 ng/kJ). The volatile compounds
tended to be detected more often and in
significantly higher concentrations than
the semivolatile compounds. The compounds
that occurred most frequently and in the
highest concentrations, nine volatile and
six semivolatile, are listed below.
377
-------
Volatiles
Benzene
Toluene
Carbon tetrachloride
Chloroform
Methylene chloride
Trichloroethylene
Tetrachloroethylene
1,1,1-Tri chloroethane
Chlorobenzene
Semivolatiles
Naphthalene
Phenol
Bis(2-ethylhexyl)-
phthalate
Diethylphthalate
Butyl benzylphthalate
Dibutylphthalate
All of the compounds listed above fre-
quently occurred in stack emissions even
when they were not present in the waste
that was burned. In most cases, they oc-
curred in emissions from incinerators,
boilers, and kilns, and over a wide range
of process conditions. For example, tem-
peratures ranged from 700°C to 1500°C,
residence times from 0.2 to 6 seconds, and
oxygen concentrations from 2 to 15 percent.
The total quantities of organic com-
pounds emitted were also compared to mea-
surements of total hydrocarbon emissions.
This comparison showed that on average only
a small percentage of the total hydrocarbon
emissions have been identified. Table 1
shows mean values for the percent of total
hydrocarbons that the total quantity of de-
tected organic compounds accounted for of
12 and 14 for incinerators and kilns, re-
spectively, and 0.4 for boilers. The range
of values for individual test runs is very
wide, with values ranging from less than 1
up to 50 percent. In general, the quantity
of total hydrocarbons varied between runs
more than the quantity of detected organics.
Identifying the other components of incin-
erator effluents is a primary objective of
the THE project.
Reviews of GC/MS data files for three
incinerator test runs were also conducted
to search for non-Appendix VIII compounds
that were not identified in the original
data analysis. This search showed the
quantity of volatile non-Appendix VIII com-
pounds detected by GC/MS varied from only a
fraction of the total quantity of Appendix
VIII compounds in one case to twice the
quantity in another case. For semivola-
tiles, the search showed that the total of
non-Appendix VIII compounds was 3 to 30
times the quantity of Appendix VIII com-
pounds. Some of the non-Appendix VIII
compounds identified in the highest concen-
trations were:
Acetone
Acetophenone
Benzaldehyde
Benzenedi carboxalde-
hyde
Benzoic acid
Chlorocyclohexanol
Cyclohexane
Cyclohexanol
Cyclohexene
Dioctyl adipate
Ethenylethyl benzene
Ethylbenzaldehyde
Ethylbenzene
Ethylbenzoic acid
Ethyl phenol
(EthylphenylJethanone
Ethynylbenzene
Phenylacetylene
Phenylbutenone
l,l'-(l,4-Phenylene)-
bisethanone
Phenylpropenol
Propenylmethyl benzene
Tetramethyloxi rane
Trimethylhexane
A subset of the available data was de-
fined which consisted of only those data
where specific compounds were identified in
stack emissions but not in the waste
burned. These data were used to compare
emissions from incinerators, boilers, and
kilns as shown in Table 2. Three of the
semivolatile compounds listed earlier were
not included because there was minimal data
for comparison. The data in Table 2 show
that values from test run to test run
varied considerably, thus these data do not
allow prediction of results for any given
test. Benzene was the only compound with
high mean levels for all three combustion
devices. Many of the volatile compounds
showed higher levels for boilers, and in-
cinerators tended to show higher levels for
semivolatiles.
Data were also available from several
baseline tests on boilers and kilns. The
fuel for most of the baseline tests was oil
or natural gas, with some coal or coke
burned in a few kiln tests. These data
were used to compare emissions from hazard-
ous waste combustion with combustion of
fuel without hazardous waste. Table 3 pre-
sents this comparison for volatile com-
pounds. Sufficient data were not available
for a similar comparison for semivolatile
compounds. Considering the wide range in
values from test to test, the data in
Table 3 suggest there is little inherent
difference between waste and fuel combus-
tion in emissions of the volatile compounds
shown in the table.
Sufficient data for five semivolatile
compounds were available to compare their
emissions when burning hazardous waste ver-
sus their emissions from municipal inciner-
ators and coal-fired power plants. Similar
data were not available for volatile com-
pounds. Table 4 presents this comparison.
The four phthalate compounds in the table
378
-------
all show very similar emission rates from
all three sources. Naphthalene emissions
were lower for power plants than the other
two sources. Again, the data suggests that
for some compounds there is little inherent
difference between combustion sources.
CONCLUSIONS
Based on a brief review of available
data on organic emissions from combustion
devices, the following preliminary conclu-
sions were reached.
1. Numerous Appendix VIII and non-
Appendix VIII compounds are emitted from
hazardous waste combustion devices. A few
specific compounds are frequently detected.
2. Significant data gaps remain to
allow a full characterization of total
emissions from hazardous waste combustion
stacks or other effluents.
3. Limited comparisons for a few com-
pounds suggest that the emissions of those
compounds are not inherently different for
hazardous waste and nonhazardous waste com-
bustion.
379
-------
TABLE 1. TOTAL APPENDIX VIII
COMPOUNDS MEASURED
AS PERCENT OF TOTAL
HYDROCARBON
Mean
Incinerators 12
Boilers 0.4
Kilns 14
TABLE 2. EMISSION RATES OF SPECIFIC
BOILERS, AND KILNS, ng/kJ
Range
0.3-44
0.01-1.1
0.1-50
COMPOUNDS FROM
INCINERATORS,
Benzene
Toluene
Carbon tetrachloride
Chloroform
Hethylene chloride
Trichloroethylene
Tetrachl oroethyl ene
1,1,1-Trichloroethane
Chlorobenzene
Naphthalene
Phenol
Di ethyl phthal ate
Incinerators
Mean Range Mean
87 2-980 30
1.6 1.5-4.1 280
0.8 0.3-1.5 1.
3.8 0.5-8.4 120
2.2 0-9.6 180
5.2 2.3-9.1 1.
0.3 0-1.3 63
0.3 0-1.3 7.
1.2 0-6.0 63
44 0.7-150 0.
7.8 0-16 0.
3.7 2.8-4.8 0.
Boilers
Range
0-300
0-1,200
8 0-7.2
0-1,700
0-5,800
2 0-13
0-780
5 0-66
0-1,100
6 0.3-2.1
3 0-0.8
4 0.04-1.6
Kilns
Mean Range
580 290-1,000
No data
No data
No data
No data
1.3 0.7-2.8
No data
2.4 (One value)
152 33-270
No data
0.02 0-0.05
No data
380
-------
TABLE 3. VOLATILE COMPOUND EMISSION RATES FROM WASTE AND
FUEL COMBUSTION, ng/kJ
Waste Fuel
Benzene
Toluene
Chloroform
Methyl ene chloride
Trichloroethylene
Tetrachloroethylene
1,1,1-Trichloroethane
Mean
120
210
90
130
1.8
49
5.6
Range
0-1,000
0-1,200
0-1,700
0-5,800
0-13
0-780
0-66
Mean
320
190
14
160
1.5
7.7
1.7
Range
0-1,100
0-1,500
0-180
0-3,400
0-14
0-52
0-14
TABLE 4. SEMIVOLATILE COMPOUND EMISSION RATES FROM WASTE COMBUSTION,
MUNICIPAL INCINERATORS, AND COAL POWER PLANTS, ng/kJ
Hazardous
waste
Mean Range
Municipal
waste
Mean Range
Coal power
plant
Mean Range
Naphthalene 17 0.3-150 71 0.4-400 0.5 0.06-1.8
Bis(2-ethylhexyl)phthalate 4.6 0-21 4.6 0.4-12 7.6 0.2-24
Diethylphthalate 1.2 0.04-4.8 0.5 0-0.9 2.8 0.4-5.7
Butyl benzylphthalate 3.7 0.7-23 — No data —- 0.5 0.3-1.0
Dibutylphthalate 0.3 0-1.1 3.9 1.5-7.6 3.0 0.09-8.7
381
-------
CONCENTRATION AND PURIFICATION OF DILUTE HAZARDOUS WASTES
BY LOW PRESSURE COMPOSITE MEMBRANES
D. Bhattacharyya, T. Barranger,
M. Jevtitch and S. Greenleaf
University of Kentucky
Department of Chemical Engineering
Lexington, Kentucky 40506-00^6
ABSTRACT
The use of membrane processes for waste purification and volume reduction is
gaining considerable attention in many industries. For hazardous wastes containing
priority organics and salts, reverse osmosis membranes can provide simultaneous
separation of both organics and inorganics. The industrial development of non-
celluloslc (aromatic polyamide, sulfonated polysulfone, etc.) thin-film composite
membranes has provided a means for reverse osmosis treatment with high solute
separations and minimal compaction problems. The low pressure reverse osmosis membranes
have definite advantages in terms of energy savings, capital cost, and broad pH (2 to
12) operating ranges. Synthetic wastes utilized in this study include: PAH compounds,
phenol, chloro- and nitrophenols and phthalates. The polyamide (FT3Q-BW) membranes used
in the study had 97-98? standard NaCl rejections and 2U to 30 gal/ft -day (gfd) pure
water flux at 300 psi. For ionizable organics, such as phenol, chlorophenols, and
nitrophenols, the rejections and flux drops were highly dependent on operating pH
values. Membrane experimental results showed 99.5 - 99.8? rejections (at pH 11) of
phenol, 2-chlorophenol, 2,i(-dichlorophenol, 2,Jt,6-trichlorophenol, etc.
INTRODUCTION
Industries engaged in organic
chemicals and allied products
manufacturing operations use contact
process water, and the spent aqueous
wastes often contain various hazardous
priority pollutants (1-1). Of the 300 to
JJOO million tons of industrial wastes
generated in the United States each year,
it is estimated that *J2 million tons are
hazardous (5). Sixty-two percent of this
quantity is generated by the chemical
products industry. Another source of
aqueous hazardous wastes is from
unsecured industrial waste storage and
from disposal sites leachate (6,7).
These contamination streams are diverse
in terms of composition and
concentration, and contain a broad
spectrum of priority organics and heavy
metals. Many contaminant streams are
often relatively dilute, thus a
concentration step prior to
detoxification or disposal may be
necessary.
Shuckrow et al. (1) have reported a
detailed compilation of data on the
performance of unit processes for
concentrating the hazardous constituents
of aqueous waste streams. Various
laboratory scale treatability
(physicochemical) techniques for priority
pollutants have also been reported in the
literature (8-10). These techniques
included carbon adsorption, steam
stripping, and ozone oxidation. Strier
(11) reported the relationship of
estimated theoretical treatability of
organic priority pollutants with water
solubility, partition coefficient
(octanol-water), bioconcentration and
aquatic life toxicity. Other novel
382
-------
destruction techniques for hazardous
wastes include supercritical extraction
and supercritical water oxidation. For
dilute wastes a combination of membrane
processes (to concentrate the aqueous
wastes) followed by incineration or wet
air oxidation will substantially reduce
the energy consumption of these units
because of 20 to 50 fold reduction of
waste volumes.
MEMBRANE CONCEPTS AND SELECTED
APPLICATIONS
The use of membrane processes for
waste purification and concentrate volume
reduction is gaining considerable
attention in many industries. For
hazardous wastes containing heavy metals
and priority organics, reverse osmosis
membranes provide simultaneous separation
of both inorganics and organics. The
recent industrial development of non-
cellulosic (aromatic polyamide,
sulfonated polysulfone, etc.) thin-film
composite membranes yields a membrane
system for high solute (metal salts and
organics) separation at pressures of 200-
300 psi. Compaction problems with these
membranes are minimal. The low pressure
reverse osmosis membranes have definite
advantages over high pressure systems in
terms of energy savings, and capital
cost. The composite membranes perform
better than cellulose acetate membranes
in most aspects, including water flux,
rejection, temperature and pH stability,
and pressure requirements.
Membrane processes are generally
evaluated in terms of three parameters!
membrane rejection (R), permeate-water
flux (J ), and extent of water recovery
(r). The membrane rejection parameter,
R, is a measure of the extent of solute
separation, _
R
1 -
c.
(1)
in which C and C. are the permeate and
feed-stream solute concentrations,
respectively. Primary separation of
solutes occurs at the thin-film (skin)
barrier layer. Thin-film membranes
result in a higher flux at pressures
considerably less than asymmetric
cellulose-acetate membranes,
A number of models have been
developed to describe the transport of
solute and solvent through membranes.
Recently, Soltaneih and Gill [1] provided
an excellent review of various models.
The most commonly used model is the
solution-diffusion model. The water and
solute fluxes (under a chemical-potential
driving force) are given by
J - A(AP - Air) (2)
and solute Flux,
J - B(AC) (3)
in which (AP - Air) is the net trans-
membrane pressure, and "A" is the
membrane permeability (function of
temperature) constant. "B" and "AC" in
Equation 3 are the solute permeability
(function of solute-distribution
coefficient between solution phase and
membrane phase) and the concentration
gradient between the membrane surface and
the permeate, respectively. In the case
of negligible concentration polarization,
AC and Air become the concentration and
osmotic pressure difference between the
bulk solution and the permeate,
respectively.
The solution-diffusion model does
not.take into account any specific
solute-polymer interactions or pore flow.
The surface force-pore flow (SFPF) model
proposed by Matsuura and Sourirajan
offers a new approach to the analysis of
experimental reverse osmosis data-
(12,13). Unlike the solution diffusion
mechanism, this model considers the
surface of a membrane to be microporous.
Fluid transport under pressure takes
place through capillary pores. The
mechanism of solute separation is a
result of surface phenomena involving
interactions between solute, solvent and
membrane pore wall. Thus the pore volume
distribution and the chemical nature of a
given membrane constitute critical data
needed for the description of the SFPF
model.
Review of the development of
membrane technology includes works by
Lonsdale (11!), Sourirajan (13), Scott
(15), Strathman (16), Lloyd (17), and
Belfort (18). Reverse osmosis for the
treatment of wastewaters can be applied
to six major industrial categories.
These, categories include landfill
leachate, textiles industry,
electroplating and metal finishing,
petroleum and petrochemical industry,
pulp and paper industry, and the food and
beverage industry. Non-industrial
applications include municipal wastewater
383
-------
treatments, desalination, and water
reuse.
Although many of these applications
refer to the use of cellulose acetate
membranes, more recent work has produced
very favorable results with the newer,
more efficient thin-film composite
membranes (aromatic polyamide). Reverse
osmosis systems reported by Chian (19)
were capable of removing more than 99? of
fifteen major pesticides including seven
chlorinated hydrocarbons. Carcinogenic
substances can successfully be removed by
reverse osmosis as shown in a study done
by Light (20) where a reverse osmosis
feed containing 70 mg/1 TOG was reduced
by 92.5? to 3.8 mg/1 with both spiral
wound and hollow fiber polyamide
membranes. Individual carcinogenic
species in synthetic wastewater were
studied with overall rejection
efficiencies of 85 and 82? for spiral
wound and hollow fiber modules. Shrem
and Lawson (21) employed reverse osmosis
units to treat wastewater from an organic
chemical manufacturing plant. Spiral
wound polyamide units were installed
which could be operated over a pH range
of 3 to 9 and at temperatures of 1 to
57°C. During seven weeks of continuous
operation conductivity rejection averaged
about 92? at 80? recovery and organic
matter rejection remained over 90?.
Bhattacharyya et al. (22) have done
extensive work with biotreated coal
liquefaction wastewater with thin-film
aromatic polyamide (FT-30) spiral wound
modules. This highly colored wastewater
contained NaCl (1600 mg/1), TOO (80
rag/1), trace heavy metals, etc., and the
membrane permeate results showed NaCl at
120 rag/1, TOG < 2 mg/1, and Cu and Zn at
< 0.01 mg/1. Individual HPLC results of
TOC showed 92-95? rejections of organic
acids (raaleic, formic, etc.) and 100?
rejection of color organics
(valerolactam, methylpiperidinone, etc.).
Standard salt rejection (even after one
year of operation) remained constant at
97.1?, demonstrating excellent membrane
stability.
OBJECTIVES
The development of a low pressure
membrane (noncellulosic composite
membranes) process to concentrate
selected priority pollutants from
hazardous wastes, will substantially
improve conventional destruction
techniques. This work deals with the use
of thin-film, composite membranes (at
250-350 psi) for concentration and
separation of pollutants from aqueous
waste streams. Spiked waste streams
utilized will include: PAH's, phenolics,
chloro- and nitro-phenols, chlorinated
benzenes and phthalates. As the work
progresses, actual waste streams will
also be collected for treatment in the
membrane system. Separation of
pollutants and operation of the membrane
system will be evaluated as a function of
system pressure, flow rate, input waste
concentration.
EXPERIMENTAL
Membrane studies were conducted in
batch, continuous thin channel, and
spiral-wound modules. The batch
operating conditions were 1400-1800 ml of
feed solution, a system pressure (AP) of
200-300 psi and pH - 1.5-11.8. The
continuous operating conditions were: AP
= 100-300 psi, Reynold's number (Re) =
1000-9000, and pH = 3.3 - 11.8. For both
the batch and continuous systems (Figure
1), standard distilled water flux and
salt (NaCl) rejection were obtained prior
to experiments with any hazardous organic
compounds. The membranes used in this
study were made of aromatic polyamide
(FT30-BW). Future membrane studies will
also include sulfonated polysulfone
membranes. For each experimental run,
samples of feed, concentrate, and
permeate were collected, properly stored,
and analyzed.
Membrane feed, concentrate, and
permeate samples were analyzed in terms
of TOC (direct injection), HPLC (direct
injection or after solvent extraction if
below detection limit), and GC (after
solvent extraction-concentration). The
reproducibility and recovery of the
solvent extraction-concentration step
were checked with known synthetic
solutions and spiked samples. The
objectives of the HPLC analysis by
reverse phase columns are two-fold:
establishment of membrane output
concentrations and to correlate membrane
rejection behavior with HPLC elution
times. Previous studies have indicated
an increase in rejection with HPLC
elution times.
384
-------
RESULTS AND DISCUSSIONS
Membrane separation (batch cell and
continuous unit) of selected classes of
priority pollutant mixtures was studied
at TOO to 300 psi. The batch cell study
focused on the mixtures of selected
sparingly soluble PAH compounds
(naphthalene, anthracene, phenanthrene),
nitrophenols, and phthalates. The
continuous unit study focused on the
mixtures of selected chlorophenols and
nitrophenols. One run involving phenol
and a salt mixture was also conducted.
For ohloro- and nitro-phenols, a wide
range of pH values was selected in order
to establish the rejection behavior of
nonionized and ionized species.
Membrane stability was continuously
checked with standard NaCl runs and with
distilled water. Figure 2 shows the
distilled water flux behavior of FT30-BW
membranes (AP = 300 psi) over an
operating time of 25 days. The membrane
shows a 15? drop in distilled water flux
with standard NaCl rejections remaining
constant at 97~98?, indicating good
membrane stability.
Membrane Studies in Batch Cell
Membrane separation of naphthalene
(solubility 20-22 mg/1), anthracene
(solubility 0,12 mg/1), phenanthrene
(solubility 0.58 mg/1), and
dimethylphthalate (278 mg/1) was carried
out in the batch reactor. Rejection of
dimethylphthalate by FT30-BW was about
97$, and rejection of naphthalene was
98.0?. For the higher molecular weight
anthracene and phenanthrene, which are
chemically similar to naphthalene, the
rejections were 98-99?. The flux drops
with these compounds were only 3~5?. The
material balance analysis of PAH
compounds showed significant loss of
these compounds through the test,
probably due to adsorption on the
membranes.
Additional experimental runs were
performed on a mixture of chlorophenols
(phenol, 2-chlorophenol (2-CP), 2ti|-
dichlorophenol (2,i)-DCP), 2,1,6-
trichlorophenol (2,4,6-TCP), and H~
chlorocresol (1-CCR)), and nitrophenols
(2-nitrophenol (2-NP) and 4-nitrophenol
(if-NP)). It is important to emphasize at
this point that phenolics are ionizable
species. As such, the pH values of
various mixtures were varied from pH =
4.5 to pH » 11.8 to observe the
separation characteristics at various pH
levels. At solution pH > 11 essentially
all species are 100? ionized.
Experimental runs involving the
chlorophenol mixture included feed
concentrations (each ) of 4-8 ppm, 22
ppra, and 102 ppm at pH values of 5, 9,
and 11. Tables 1 and 2 illustrate the
flux behavior of the chlorophenol
mixtures during the experimental runs.
In each chlorophenol experimental run
involving high pH, there was a smaller
drop in permeate flux than in low-pH runs
(Table 1). Table 2 illustrates the
permeate flux behavior for each
experimental run compared to that of
distilled water. In each experimental
run involving high pH, there was a much
smaller drop in permeate flux compared to
distilled water than for those runs at
low pH. Figure 3 shows the effect of
ehloro substitution on a phenol molecule
in terms of rejection by the FT30-BW
membrane at increasing pH. As predicted
the rejection of each species increases
with pH. The rejection of the individual
species also increases with each ehloro
substitution on the phenol molecule.
Similar runs with nitrophenol mixtures
(46 mg/1 phenol •*• 65 mg/1 2-NP + 56 mg/1
4-NP) at high pH (pH » 11.5) showed very
high rejections of phenol (98.3?), 2-NP
(99.3?), and 4-NP (99.1?).
As a means of studying the
possibility of fouling or adsorptive
behavior of the solutes on the FT30-BW
membrane, the concentrate volume from
each experiment was withdrawn and
analyzed. Based on the volumes and
concentrations of the feed and permeate
from each experimental run, mass balance
calculations were performed to predict
the concentrate pollutant concentration.
Figure H shows the relationship between
the calculated concentrations and actual
concentrations as measured in the
laboratory. The chlorophenol mixture
runs exhibited overpredict!on of the
concentrate concentration at low pH
values. The substituted phenolics
displacing the water layer in the
membrane pore (occurring at low pH under
non-ionizing conditions) would result in
a lower concentrate concentration than
expected. For higher pH runs, the closer
385
-------
material balance indicates insignificant
adsorption of phenolics on membranes.
Membrane Studies in ContinuousUnit
Continuous runs were made with
chloro- and nitro-phenol mixtures in the
turbulent flow regime. High flux drop
was observed under non-ionization
conditions (pH » 3-3) of the various
solutes. These results were obtained
even under a high Reynolds number (9000).
The behavior was similar to that observed
in the batch cell. With a chlorophenol
mixture of an equivalent concentration_pf
1,9 mM the flux dropped from 14.2 x 10
cm/a to 7.'I * 10 cm/s. These results
indicate that reversible interactions
(flux recovered by water flushing)
between the membrane polymer and the
solute itself are taking place and that a
phenomenon of polarization is not the
primary cause of flux decline. As it can
be observed in Table 4 at high pH, when
the chlorophenol derivatives are ionized,
that the flux drop is negligible (from
111,8 x 10 cm/s to 11.2 x 10 cm/s),
and the rejections are all above 99?.
To establish the flux drop and
rejection phenomena, various types of
mixtures were run using thin channel
cells. Runs were conducted with mixtures
of Phenol, 2-CP, 2,4-DCP and 4-CCR
containing a total molar concentration of
1.8 niM. The effect of pressure on the
rejection of this mixture is shown in
Figure 5. At 300 psi the flux drop in
this case was 31*. With 2,4,6-TCP in the
mixture the flux drop was 48?. A single
solution of 2,4,6-TCP (1.6 mM) dropped
the water flux by 35?. The flux behavior
of the chlorophenol mixture (Phenol,
20CP, 2,1-DCP, 4-CCR) over the range 100
psi - 300 psi was linear with AP, thus
indicating the absence of surface
polarization phenomena.
To understand the effect of
multicomponent systems on the flux
behavior, the mixtures of phenol-2 CP
(16,6 ppm - 51.8 ppm), phenol-2,4-DCP
(15.5 ppm - 62.8 ppm) and 2 CP - 2,1-DCP
(55.0 ppm - 71.9 ppm) were used. The
order of flux drops compared to the
double distilled water (DDW) flux was
13?» 24?, and 25?, respectively. It can
be concluded that 2,4,6-TCP and 4-CCR are
the two compounds causing the most flux
decrease. It should be noted that for a
single phenol a flux drop of only 5.5?
occurred. For-any type of mixture under
non-ionization conditions, the rejections
of the various chlorophenols were always
R ^ R <* R <* R
phenol "-> "~ N "~ " """ N "'
Tnis sequ«
HPLC elution time pattern in a reverse
phase (C,0) column.
2,4.6rTeP-
same as the
-o
The rejection of nltrophenols was
first studied with a phenol, 2-WP, 4-NP
mixture. The run was performed under
non-ionization conditions and at a
Reynolds number of 9000. The effect of
pressure on the rejection is shown in
Figure 6. It can be observed that
although the solution flux was higher for
the case of the nitrophenols compared to
the chlorophenols, their rejections were
lower than that of the chlorophenols
under similar experimental conditions.
For instance, since 2-CP and 2-NP or 4-NP
are molecules of similar Stoke's radius,
the difference in rejection can be caused
by various interactions between membrane
pore wall and solute. It should be noted
that phenol rejection was similar in the
chlorophenol and nitrophenol mixture.
The preferential sorption capillary
pore flow model developed by Sourirajan
gives a better understanding of the
rejection and flux phenomena. The
negative and positive adsorption of
solute at the membrane-solution interface
arises from net repulsive or attractive
forces acting on the solute from the
adjacent membrane surface. The model
admits that there is a layer of water
preferentially adsorbed at the pore wall.
In some cases this layer of water can be
displaced by some molecules of solute
exerting stronger adsorption forces
toward the pore wall. For example, the
Stoke's radius of water and phenol are
0.87 A and 2.1 A, respectively. A layer of
phenolics displacing the water layer will
definitely cause a water flux drop by
reducing the available path of the fluid.
This phenomena occurs at low pH with no
ionization of the solutes.
In order to predict the separation
of organ!cs by a membrane, one needs to
know the pore distribution of the
membrane. The skin pore distribution of
the FT30-BW membrane was measured by C0_
(217°K) and by N (77°K) gas adsorption
technique (23). The pore distribution is
shown in Figure 7. Utilizing the pore
386
-------
distribution, and a new calculation
technique for the simultaneous solution
of the radial velocity profile of the
solvent through the pores, the solute
concentration in the product water and
the interaction parameters were applied
to compute rejection of phenolies.
Figure 8 shows excellent agreement
between calculated and experimental
results,
CONCLUSIONS
This study of thin-film composite
membranes for the separation of selected
classes of hazardous organic compounds
has proven quite effectively the benefits
of such a process. This particular waste
treatment technique offers definite
advantages in terms of high solute
separations at low pressures (100-300
psi) insignificant compaction problems,
and broad pH operating ranges (pH 2 to
12). The membrane stability was
established over an extended operating
time (25 days) with standard NaCl
rejections of .97-98$ with 2*1 to 30 gfd
pure water flux at 300 psi. Also,
membrane rejection of solutes actually
improves with membrane usage and permeate
flux can be maintained with minimal
cleaning procedures (NaOH washes). For
ionizable organics such as phenol,
chlorophenols and nitrophenols the
rejections and flux drops were highly
dependant on operating.pH. Membrane runs
(feed - 100 ppm each of chlorophenols)
showed a permeate flux drop of 35.7? at
low pH (4.6) and 3.1? at high pH (11.5)
at 80$ recovery. Similarly the permeate
flux was 16.2? lower than distilled water
at high pH and 75.3? lower than distilled
water at low pH. Experimental results
showed 99.5-99.8$ rejections (at pH 11)
of phenol, 2-chlorophenol, 2,4-
dichlorophenol, 2,4,6-trichlorophenol,
etc. Similarly, 2-nitrophenol and 4-
nitrophenol were rejected 99.1-99.3? at
pH 11. The sparingly soluble compounds
of naphthalene, anthracene, and
phenanthrene showed rejections of 98-99?
with negligible permeate flux drop.
Solute adsorption on the membrane was
observed in all pAH compound membrane
runs, and in ehlorophenol mixture runs at
low pH. Evidently, the ionization of the
solute species not only improves
rejection and reduces permeate flux drop,
but effectively reduces the adsorption of
species on the membrane. The effect of
Reynold's number and trans-membrane
pressure were studied in the continuous
membrane unit. Rejection of all species
increases with increasing pressure and
Reynold's Number. Permeate flux also
increases with increasing pressure.
However, there is an optimum point at
which the increased pumping and pressure
costs would overcome the benefits of
higher permeate flux and better solute
separation. It is also important to note
that solute rejection increases with
chloro-substitution on the phenol
molecule and decreases with nitro-
substitution. In addition to
experimental results, a computer model
was developed utilizing the surface
force-pore flow model developed by
Sourirajan. Using the pore distribution
of the membrane, the solute concentration
in the product water and the interaction
parameters were applied to compute
rejection of phenolics. Preliminary
results show excellent agreement between
calculated and experimental values.
Future work will include membrane
experiments involving mixtures of
chlorobenzene mixtures and actual
wastewater.
REFERENCES
1. Shuckrow, A.J., Pajak, A.P., and
Osheka, J.W., "Concentration
Technologies for Hazardous Aqueous
Waste Treatment", EPA Report, EPA-
600/2-81-019 (1981).
2. Hackman, E.E., "Toxic Organic
Chemicals Destruction and Waste
Treatment", Noyes Data Corporation,
Park Ridge, New Jersey (1978).
3. "Industrial Hazardous Waste
Management", Industry and
Environment Special Issue - No. 4
(1983).
U. McCarty, P.L., "Organics in Water -
An Engineering Challenge", J. Env.
Eng. (ASCE), EEj, 1 (1980).
5. "Hazardous Waste Generation and
Commercial, Hazardous Waste
Management Capacity", EPA Report SW-
891 (1980).
6. Shuckrow, A.J., Pajak, A.P., and
Touhill, C.J., "Management of
Hazardous Waste Leachate", EPA
307
-------
Report, Contract No. 68-03-2766
(1980).
?. "Land Disposal of Hazardous Wastes",
EPA Report, SPA-60Q/9-8JJ-007 (1984).
8. Pochtman, E.G. and Eisenberg, W.,
"Treatability of Carcinogenic and
Other Hazardous Organic Compounds",
EPA Report, EPA-600/2-79-097 (1979).
9, Dobbs, R.A. and Cohen, J.M», "Carbon
Adsorption Isotherms for Toxic
Organics", EPA Report, EPA-600/8-80-
023 (1980).
10. Smith, J.K., "Laboratory Studies of
Priority Pollutant Treatability",
EPA Report. EPA-600/2-81-129 (1981).
11. Strier, M.P., "Estimated Theoretical
Treatability of Organic Priority
Pollutants", EPA Report, EPA-4W1-
79-100 (1979).
12. Matsuura, T., Taketani, Y., and
Sourirajan, S., Journal of Colloid
and Interface Science, 95, No. 1, 10
(1983).
13- Matsuura, T. and Sourirajan, S.,
"Reverse Osmosis/Ultrafiltration
Process Principles", National
Research Council, Canada, Ottawa
(1985).
14. Lonsdale, H.K., "The Growth of
Membrane Technology", Journal of
Membrane Science, W, 81 (1982).
15. Scott, J., "Membrane and
Ultrafiltration Technology! iecent
Advances", Noyes Data Corp., Park
Ridge, New Jersey (1980).
16. Strathman, H.f "Membrane Separation
Processes", Journal of Membrane
Science, 9_, 121 (1981).
17. Lloyd, D.R., "Material Science of
Synthetic Membranes", ACS Symposium'
Series 269, Washington, D.C. (1985).
18. Belfort, G., "Synthetic Membrane
Processes: Fundamentals and Water
Application", Academic Press Inc.,
New York, New York (198H).
19. Chlan, E.S.K., Bruce, W.N., and
Pang, H.H.P., "Removal of Pesticides
by Reverse Osmosis", Environmental
Science and Technology, £, 52
(1975).
20. Light, W.G., "Chemistry and Water
Reuse", 1, Ed. W.J. Cooper, Ann
Arbor Science, Ann Arbor, Michigan,
352 (1981).
21. Shrem, M. and Lawson, T., Industrial
Water Engineering, 1_4, 16 (1977).
22. Bhattacharyya, D., Jevtitch, M.,
Ghosal, J.K., and Kozminski, J.,
"Reverse Osmosis Membrane for
Treating Coal-Liquefaction
Wastewater", Environmental Progress,
3, No. 2, 95 (1981).
23. Bhattacharyya, D., Jevtitch, M.,
Fairweather, G., and Schrodt, J.T.,
"Prediction of Membrane Separation
Characteristics by Pore Distribution
Measurements and Surface Force-Pore
Flow Model", Chemical Engineering
Communications, In Press (1986).
ACKNOWLEDGEMENTS
This project is supported by a
Cooperative Agreement Grant with the
U.S. EPA-HWERL-Cincinnati (CR 811976).
The project officer is Mr. John F. Martin
of HWERL, Cincinnati.
308
-------
Table 1
Flux Behavior of FT30-BW at High Recovery
System
Saturated mixture of PAH's
4.88 ppm each of ehlorophenols
-4.80 ppm each of ehlorophenols
21.6 ppm each of ehlorophenols
22.4 ppm each of ehlorophenols
21.5 ppm each of ohlorophenols
101. 4 ppm each of ehlorophenols
102.3 ppm each of ehlorophenols
(c)
nitrophenols mixture
51.1 ppm each of chloro and
nitrophenols
pHFeed
5.5
5.5
11.8
5.5
8.9
10.8
it. 6
11.5
11.5
11. J»
% Flux Drop*
+ 7.09
11.3
6.49
12.6
'"6.52
10.7
35.7
3.08
1.49
8.69
(a) 21.1 ppm naphthalene, 0.012 ppm anthracene, 0.57 ppm phenanthrene
(b) phenol, 4-chlorocresol, 2-chlorophenol, 2,4-dichlorophenol, 2,4,6-
trichlorophenol
(c) 45.5 ppm phenol, 65.2 ppm 2-nitrophenol, 55.6 ppm 4-nitrophenol
(d) phenol, 4-chlorocresol, 2-chlorophenol, 2,4-diehlorophenol,
2,4,6-trichlorophenol, 2-nitrophenol, 4-nitrophenol
Jw e
389
-------
Table 2
Flux Drop During Experimental Runs Compared to That of Distilled Water
JwPermeate JwDist. HJ3
System
4.88 ppm each of ehlorophenols
4.80 ppm each of chlorophenols
21.6 ppm each of chlorophenols
22.4 ppm each of chlorophenols
21.5 ppm each of chlorophenols
101.4 ppm each of chlorophenols
102.3 ppm each of chlorophenols
nltrophenols mixture
51.1 ppm each of chloro- and
(c)
nltrophenols
" Feed x 10 cm/a
5.4 8.34
11.8 9.11
5.5 6.25
8.9 7.94
10.8 10.2
4.6 2.25
11.5 8.27
11.5 7.25
11.4 6.57
4
x 10 cm/s
10.1
10.6
11.0
10.7
10.5
9.12
9.94
10.3
9.20
(a) phenol, 4-chlorocresol, 2-chlorophenol, 2,4-dIchlorophenol,
2,4,6-triehlorophenol
(b) 45,5 ppn phenol, 65.2 ppm 2-nitrophenol, 55.6 ppin i|-nitrophenol
(o) phenol, 4-chlorocresol, 2-chlorophenol, 2,4-dIchlorophenol,
2,4,6-trichlorophenol, 2-nitrophenol, 4-nitrophenol
390
-------
Table 3
Rejection and Flux Behavior of Ionized Chlorophenols at High pH
Membrane: FT30-BW
AP = 300 psi; pH - 11.8
Avg. distilled water flux = 14.8 x 10
cm/s
Mixture of phenol (53 mg/1)
+2 chlorophenol (52 mg/1)
+2, it-dichlorophenol (18 mg/1) •
+2, 4, 6-trichlorophenol (41 mg/1)
+4 chloro 3 methyl phenol (46 mg/1)
(Total mM = 1.8)
Mixture of phenol (45 mg/1)
+2 chlorophenol (43 mg/1)
+2, 4-dichlorophenol (41 mg/1)
+2, 4, 6-triehlorophenol (43 mg/1)
+4 chloro 3 nethyl phenol (42 mg/1)
(Total mM - 1.6)
Reynold's No.
4490
PH
perm
10.7
Water Flux
ii
em/s x 10
13-5
9000
10.6
14.2
% Rejection
99.6 (0.22 mg/1)*
99.5 (0.25 mg/1)
99.2 (0.38 mg/1)
99.4 (0.24 mg/1)
99.5 (0.20 mg/1)
99.7 (0.15 mg/1)
99.7 (0.13 mg/1)
99.7 (0.12 mg/1)
99.7 (0.12 mg/1)
99.8 (0.093 mg/1)
* permeate concentration
-------
Figure 1. Continuous Membrane Unit with Two Parallel Modules
20-
w
u
X
X
u.
01
•o
OJ
-4-)
w
18-:
16-
14-
a NaCl rejection at 30Qpsi
Distilled tfotor Flux ot 300psi
2 V b b lo is 14 IB ie io is 14
Operating Time (days)
100
- 95
- 90
o
n
ID
to
O
DC ^ I"
O3 !_,.
o
^ 80
r 75
70
Figure 2. Membrane Stability of FT30-BW Membrane
392
-------
*•*
•*"1
o_
I
«w*
B
CO
01
•*"*
o
at
a_
to
o
o
01
01
.5-
Feed : 22ppm each chloronhenols
s.s
PHp = 3.9
* Phenol
* 2-chlorophanol
• 2,4-dichlorophanol
• 2.4,B-trichlorophenol
pH of permeate water
8.5
Figure 3. Effect of pH on Rejections of Selected
Chlorophenols•at 300 psi
9.5
T3
Ql
4J
O
*— •<
U
*— i
a
^o
c
o
*>
(_
c
QJ
U
C
o
o
0)
a
t_
c
a
c
o
14U-
130-
120-
110-
100-
A Phenol
a 2-chlcrophenol
» 2, 4, 6-trichlorophanol
A.
90- n°L . a
80-
70-
60-
50-
40-
*
d A
o * x-
* X
X
X
X
X
X
X
X
X
X
X
0 50 ^0 70
• 4-chlorocresol ^ .
o 2, 4-dichlorophaiol id" x
*• ^ xx
o x
b X
• X
0 xx
^^
X
n x
x
/ a- J l°w pH
fe : high pH
Feed: 22 ppm each of chlorophenols
io k I'dO 110 120 1*30 14
Concentrate Concentration (actual)ppm from HPLC
Figure 4. Calculated Concentrate Concentration of
Selected Chlorophenols
393
-------
o
100
35-
3Q-
85-
60-
75-
o
QJ
Ql 7(1-
Oi '"'
65-
60
A phenol
• 24-OCP
PH=5.5
Re? =9000
» 2-CP
o 4-CCR
50
100 150 200
Operating Pressure, P in psi
250
300
Figure 5. Effect of Transmembrane Pressure on
Rejections of Selected Chiorophenols
100
95-
* phenol
o 4-Nitrophsnol
D 2-Nitrophanol
§
•*-1
•8
01
V
B5-
80-
75-
70-
65-
60-
55-
50-
PH= 4.8
RE// =9000
100 ISO 200 250 300
Operating Pressure, P in psi
Figure 6. Effect of Transmembrane Pressure on Rejections
of Selected Nonionized Nltrophenols
394
-------
nnoius , NH
4.
-------
TRIAL BURNS - PLASMA ARC TECHNOLOGY
Nicholas P. Kolak, Ph.D.
New York State Department of Environmental Conservation
Division of Solid and Hazardous Waste
Albany, New York 12233
Thomas G. Barton, Ph.D.
Pyrolysis Systems, Incorporated
Wei land, Ontario, Canada L3B 5P1
Chun C. Lee, Ph.D.
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Edward F. Peduto
GCA Corporation
Bedford, Massachusetts 01730
ABSTRACT
A mobile plasma arc system for the destruction of liquid hazardous wastes has been
designed and constructed, under a NYSDEC contract, by Pyrolysis Systems Inc. (Weiland,
Ontario, Canada), The (pyrolytlc) destruction process produces acid gas, a fuel gas
and finely divided carbon. The acid gas Is removed in a caustic scrubber, which also
captures most of the carbon; the fuel gas is flared. The system has a design feed-
rate of 4 kg/minute.
The operation of the system has been tested in a series of short-term runs at
Kingston, Ontario using MEK/MeOH, carbon tetrachloride, and PCBs as feed materials.
DREs of greater than six "nines" were obtained for the chlorinated hydrocarbons, and
the emissions from the system met environmental discharge criteria established by
Canadian regulatory authorities.
INTRODUCTION
The New York State Department of
Environmental Conservation (NYSDEC)
Division of Solid and Hazardous Waste
and the U.S. Environmental Protection
Agency (USEPA) Hazardous Waste
Engineering Research Laboratory
established a cooperative agreement in
1982 for the construction and testing of
a mobile plasma arc system for the high
efficiency destruction of hazardous
wastes. The project consists of four
phases:
Phase I: Design and construction of
the mobile plasma arc system by the
contractor, Pyrolysis Systems Inc.
(PSI).
Phase II: Performance testing of the
plasma arc system at the Kingston,
Canada test site.
Phase III: Installation of the plasma
arc system and additional performance
testing at a site in New York State.
396
-------
Phase IV; Demonstration testing as
designated by NYSDEC at a New York
State hazardous waste site.
Implementation of each phase is subject
to USEPA/NYSDEC approval of the results
of the preceding phase.
The first two phases have been per-
formed in Canada with the cooperation
of Canadian federal, provincial, and
local authorities. The third and fourth
phases will be performed in New York
State. Short duration tests using
methyl ethyl ketone, carbon tetrachlor-
ide, and polychlorinated biphenyls
(Askarel) have been completed.
PROCESS DESCRIPTION
The plasma process is based upon the
concept of pyrolyzing waste molecules
using a thermal plasma field (1). A
colinear electrode assembly, built and
supplied by Westinghouse Corporation, is
used to produce the electric arc. Dry
low pressure air is used as the medium
through which the electric current
passes. As a result of the intense
energy, the air stream is ionized to
form a plasma at temperatures approach-
ing lO.OOOoc. Liquid hazardous wastes
are injected into the plasma and
broken down to the atomic state in an
oxygen deficient atmosphere. The major
compounds formed are hydrogen, carbon
monoxide, hydrogen chloride, and
particulate carbon; minor components
include carbon dioxide, acetylene and
ethene. The product gases are sub-
sequently scrubbed to remove acid gas
(HC1) and then flared to remove com-
bustibles.
Figure 1 shows a process diagram of
the plasma pyrolysis system which
consists of a liquid waste feed system,
plasma torch, reactor, caustic scrubber,
on-line analytical equipment and flare
(2,3). The system is rated at 4 kg/min
or approximately 55 gallons per hour
of waste feed. Product gas production
rates are about 5-6 m3/min prior to
flaring. For the purposes of this
program, a flare containment chamber and
10 meter stack were constructed to
facilitate testing. After combustion,
the stack gas flow rate is approximately
30-40 standard m3/min.
A Hewlett-Packard Model 5880A gas
OFF GASES TO RARE
EMERGENCY CARBON FILTER
GASCHROMATOGRAPH-
MASSSELECTMTYUNtT
LABORATORY
ANALYSE EQUIPMENT
GASCHROMATOGRAFH
SALT VOTER TO DRAIN
Figure 1, Process Schematic of Plasma Pyrolysis System.
397
-------
chromatograph is Installed in the
mobile trailer to provide information on
bulk gas composition. Pre-flare gas
samples are analyzed for hydrogen,
carbon monoxide, carbon dioxide, water,
nitrogen, methane, ethylene, ethane,
acetylene, propane, propylene and
1-butene. A Hewlett-Packard Model 5792A
gas chromatograph is coupled to a
Hewlett-Packard Model S970A mass selec-
tive detector and used to analyze
pre-flare gas samples for waste feed
residuals,
RESULTS
Methyl Ethyl Ketone (HElQ/Hethanol
The first objective of these tests
was to verify the mechanical operation
of the plasma pyrolysis system with a
non-chlorinated feed material. The
feed was a blend (1:1 vol) of MEK and
methanol (MeOH). Because polynuclear
aromatic hydrocarbons (PAH) had been
observed in earlier tests utilizing this
feed, the second objective was to deter-
mine whether PAH concentrations
(measured as benzo-a-pyrene, BaP) in the
scrubber water were low enough to permit
the discharge of the scrubber water to a
wastewater treatment plant.
The scrubber water from each of the
1-, 5- and 6-hour tests during October
and November, 1985 was stored in a 9000
L tank pending the results of laboratory
analyses. Operating data for the 1-hour
test are presented in Table 1. Scrubber
samples were collected during warm-up
and cool-down phases with four discrete
Table 1. OPERATING DATA FOR 1 HOUR TEST
Feed Rate 2.7 L/min
Total Volume Fed 162 L
Total Mass Fed 129 kg
Plasma Torch Operating
Power 326 kW
Scrubbed Product Gas
Temperature 60°C
Spent Scrubber Water
Temperature 50°C
Scrubber Water Flow
Rate 25 L/min
Product Gas Temperature
Exiting Reactor
samples collected at 15 minute inter-
vals during the 1-hour test. Concen-
trations for BaP in the aqueous phase
and on the carbon phase (Table 2)
were determined by Zenon Environmental
(Burlington, Ontario). The average BaP
concentration in the scrubber water was
3.1 mg/L. Based upon these data and
data from additional samples collected
at the wastewater plant, Canadian
authorities permitted discharge of the
stored scrubber water from these tests.
TABLE 2. ANALYSIS OF SPENT SCRUBBER
WATER FOR BaP
Sample BaP (ppb)l BaP (ppm) BaP (ppm)
Water Phase Carbon Phase Total
Field
Blank
Startup
1
2
3
4
Cool Down
5.6
12
123
656
359
57
18
NA2
0.367
0.250
3.44
7.22
0.272
NA2
0.0056
0.166
0.221
2.89
7.29
0.31
0.018
(1) Water samples were only decanted.
Residual particulates resulted in
aqueous BaP concentrations exceed-
ing the saturation value of 3-4
ug/L.
(2) Not applicable.
Carbon Tetrachloride
The primary purpose of the 1-hour
carbon tetrachloride (CCl/j) tests was to
demonstrate the destruction of a simple
chlorinated compound which is difficult
to destroy and to demonstrate effective
HC1 removal by the scrubber. The CC14
was introduced into the system in a
blend of MEK, ethanol, and water at a
rate of 1 kg CC14 per minute. The stack
testing was conducted by GCA Corporation.
Results of the three tests are presented
in Table 3.
The results indicate that the system
398
-------
TABLE 3. CARBON TETRACHLORIDE TEST RESULTS
Date
Sample Time, min.
Feed Rate, L/min.
CC14
MEK mixture
Total Mass Fed, CC14, kg
Chlorine Loading, mass %
Reactor Operating Temperature, °C
Plasma Torch Power, kW
Stack Gas Parameters
Average flow rate, dscfm 1
dscmm
Average temperature, °C
NOx cone., ppm (v/v)
Emission Rate, kg/hr
CO cone. , ppm (v/v)
Emission Rate, kg/hr
Oo, percent
C&2, percent
HC1 , mg/dscm
Emission Rate, kg/hr
CC14 cone., ppb
Emission Rate, mg/hr
Scrubber Effluent Parameters
Scrubber Effluent Flowrate, L/min
CC14 cone., ppb
Discharge Rate, mg/hr
Destruction Removal Efficiency(3)
CC14, percent ORE
Run 1
2/18/85
60
0.63
2
60.0
35
974
280
,346.3
38.13
893.3
106
0.46
48
0.13
12.7
6.0
(1)
(1)
<2 (2)
<29.3
30.0
1.3
2.3
99.99995
Run 2
2/26/85
60
0.63
1.6
60.6
40
1008
298
1,048.2
29.69
807.1
92
0.31
57
0.12
14.4
5.7
137.7
0.25
<2 (2)
<22.8
30.0
5.5
9.9
99.99996
Run 3
2/26/85
60
0.63
2
60.6
35
1025
300
1,052.7
29.81
677.3
81
0.28
81
0.17
15.1
4.9
247.2
0.44
<2 (2)
<22.9
30.0
3.3
5.9
99.99996
(1) Invalid data.
(2) The detection limit of 2 ppb CC14 in the stack gas was used to calculate the CC14
mass emission rate for each run.
(3) The ORE is based on stack emissions and excludes scrubber effluent.
399
-------
1s capable of achieving destruction and
removal efficiencies much greater than
the 99.99% required under RCRA for this
listed waste.
Polychlorinated Biphenyls (PCBs)
The purpose of conducting three
1-hour tests using PCBs was to determine
the ORE for PCBs and to acquire operating
data for this environmental contaminant.
Upon reaching a product gas temperature
of 1100°C at the reactor exit while using
HEK/MeOH as the feed material, the feed
was then switched to a blend of PCBs,
MEK, and MeOH. The scrubber water was
retained before discharge to sewer to
ensure that the concentration of BaP,
PCBs, dioxlns and furans did not exceed
criteria established by Canadian
regulatory authorities. Operating data
are presented in Table 4.
TABLE 4. OPERATING DATA FOR PCB RUN #1
Elapsed operating time: 70 min. at
operating temperature
Feed rate
Total feed-
TABLE 5. PCB 1-HOUR TEST RESULTS
PCB feed -
Feed Composition (mass)
Reactor Operating
Temperature
Plasma Torch Power
*Trichlorobenzene
3.09 L/min
2.83 kg/min
0.40 kg/min
14.1% PCB
11.0% TCB*
74.9% MEK/MeOH
11360C
327 kW
The stack gases were monitored for
participates, PCBs, dioxins, furans, NOx,
HC1, flow rate, and temperature. For
these runs the stack monitoring was
conducted by IMET, Inc. {Markham,
Ontario). The sampling and analysis
program Included the necessary quality
assurance/quality control protocols.
The procedures and equipment were stan-
dard and may be found in the IMET test
report. A summary of the results is
presented in Table 5.
Run 1
Run 2
Run 3
Date 12/5/85 12/17/85 1/16/86
Sample Time,
min. 50 60 60
Stack Gas Parameters
Flow rate,
dscmm 37.9 45.0 38.1
Temperature, °C 836 678 962
NOX, ppm 117 N/A 139
HCf, mg/dscm N/A 43 68
Oo, percent 14 14.5 16.5
C02, percent 5.5 5.0 3.0
CO, percent 0.01 0.01 0.01
Total PCB, (1) <0.013 0.46 3.0
ug/dscm (2) <0.013 0.32 <0.011
Total Dioxins,
ug/dscm
Total Furans,
ug/dscm 0.26 1.66 <0.30
Total BaP,
ug/dscm 0.18 0.45
0.076 (3) <0.43 <0.13
2Q
•o
Scrubber Effluent Parameters
Effluent Flowrate,
L/min 41 36
Total PCB, ppb(l) 1.56 2.15
(2) 0.06 4.7
Total Dioxins.ppt 5.8 <259
Total Furans, ppt 1.5 399
Total BaP, mg/L 0.04 0.92
Destruction Removal Efficiency
PCB, percent ORE
(1) 99.99999 99.99994
(2) 99.999999 99.99997
33
9.4
<0.01
<1.05
<1.05
2.0
99.9999
99.999999
(1) These values are based upon mono-
decachlorobiphenyl.
(2) These values are based upon tri-
decachlorobi phenyl
(3) No tetra or penta dioxins were
detected at 0.05 ng on a gc column,
except for Run #1 where 0.06 ng
tetra dioxin was reported.
CONCLUSIONS
From the onset of this project, there
was some concern for the production of
400
-------
compounds as a result of recombination
downstream of the point in the process at
which the liquid waste is injected.
Although PAH compounds have been found
adsorbed onto the particulate carbon
which is formed under the pyrolytie con-
ditions, the PCB tests have demonstrated
that the concentration of the PAH com-
pounds in the scrubber effluent can be
maintained within acceptable discharge
criteria.
These tests have demonstrated the
ability of the plasma arc system to
effectively destroy liquid wastes con-
taining high concentrations of CC14 or
PCBs. A ORE of 99.99998 percent was
achieved for a liquid feed containing 43
percent (mass) CC14. A ORE of 99.9999992
percent resulted from a feed containing
14 percent (mass) PCBs (Arochlor 1254 and
1260). These destruction efficiencies
are based upon the residual levels of
these compounds in both effluent streams.
This test program has also demon-
strate that a waste feed containing 40
percent (mass) chlorine can be success-
fully treated as indicated by the CC14
tests.
ACKNOWLEDGEMENTS
The valuable assistance of Dr. Hugh
Dibbs and Rob Booth of Environment Canada,
Dave Edwards of the Ontario Ministry of
Environment and Ken Kavanagh of the
Canadian Department of National Defense
and of the many individuals in their
agencies is gratefully acknowledged. The
many helpful technical discussions with
Jim Yezzi, John Brugger, and Larry Bernson
of USEPA Region II and individuals of GCA
Corp.'s stack monitoring team, particu-
larly Mark Gollands, is appreciated.
REFERENCES
1. Barton, Thomas G., 1984a.
Mobile Plasma Pyrolysls. Hazardous
Waste. Vol. 1 No. 2, pp 237-247.
2. Barton, Thomas 6., Nicholas P. Kolak,
and Chun C. Lee, 1984b. TheNew
York State Haste Disposal Project.
Waste Technology Journal, Second
Quarter, Westlnghouse Corporation,
pp 14-15.
Barton, Thomas G., Nicholas P. Kolak,
and Chun C. Lee, 1984c. Hazardous
HasteDestruction Using Plasma Arc
Technology. EPA Tenth Annual
ResearchSymposium -Land Disposal,
Incineration, and Treatment of
Hazardous Waste.
Kentucky.
Fort Mitchell,
401
-------
RADIO FREQUENCY ENHANCED IN-SITU DECONTAMINATION OF SOILS
CONTAMINATED WITH HALOGENATED HYDROCARBONS
Harsh Dev
IIT Research Institute
Chicago, Illinois 60616
ABSTRACT
The techniques of radio frequency in-situ heating can be used for the treatment of
soils containing hazardous chlorinated hydrocarbons. RF heating is used to rapidly heat
the soil to the treatment temperature to facilitate decontamination. The feasibility of
soil decontamination by thermal recovery of low boiling contaminants and chemical treat-
ment of high boiling materials such as PCBs with dehalogenating agents, has been estab-
lished. In-situ RF heating is accomplished by inserting tubular electrodes into boreholes
or by laying horizontal electrodes over the surface of the soil and by energizing them
with an RF power source. The electrode pattern is selected based on the depth of penetra-
tion of the contaminant and the required temperature. The operating frequency is selected
on the basis of the electromagnetic characteristics of the soil. Vapor and gas contain-
ment and collection systems are used to enclose the treated module.
The estimated cost for the treatment by the thermal recovery option is $29.40 to
$57.40 per ton of soil, for a temperature range of 100° to 250°C and a soil moisture
content of 5 to 20 %.
INTRODUCTION
Radio Frequency (RF) in-situ heating
is an electromagnetic (EM) technique
originally developed and demonstrated by
Krstansky et al. (7) for in-situ thermal
processing of hydrocarbonaceous earth
formations for resource recovery. This
technique uses EM energy in the radio
frequency band for rapid in-situ heating
of earth and mineral formations. The
mechanism of EM energy absorption and
conversion to heat 1s similar to that of a
microwave oven, except that the frequency
is lower and the scale of operation is
much larger. Field tests have demonstrat-
ed 1n-situ RF heating to 200°-400°C of
large blocks of earth from 35 cu ft (1.0
cu m) to 880 cu. ft (25 cm m) in size.
Sustained average heating rate of 0,8°-
1.0°C/hr was achieved in these tests.
The purpose of this paper is to
present the results of laboratory studies
performed to determine the feasibility of
using RF heating techniques for the
decontamination of soils containing
hazardous chemicals such as chlorinated
hydrocarbons, PCBs, benzene, toluene,
etc. The occurrence of numerous large
uncontrolled sites of contaminated soil
containing the above mentioned chemicals
is well documented and previously reported
in literature (2,9,12,15,16).
The radio frequency EM energy is
applied to earth and mineral formations by
a system of exciter and ground electrodes.
These electrodes may be placed over the
surface of the contaminated site, or in
vertical or horizontal bore-holes drilled
through the contaminated zone. When the
electrode array is supplied with EM ener-
gy, an electromagnetic wave is launched by
the exciter electrodes into the target
volume of soil. The energy in the EM wave
is attentuated due to absorption by the
molecules present in the heated volume.
The EM energy appears as heat due to
induced dipole rotation and molecular
402
-------
vibration. Rapid heating of the volume of
the soil bound by the array is accom-
plished because the heating is not depen-
dent upon the relatively slow process of
thermal conduction. Heating occurs
through-out the volume of the target
material.
In-situ radio frequency heating
offers two alternatives for the purpose of
soil decontamination. These alternatives
are: (1) thermal decontamination of
contaminated soil by vaporization and
recovery of contaminants, and (2) in-situ
treatment in conjunction with a chemical
dechlorinating agent for contaminants such
as PCB's. In alternative (1) above, the
contaminants will be vaporized, distilled
or steam-stripped from the soil and recov-
ered at the surface. In alternative (2)
above, RF heating of the soil will be used
to improve the rate of reaction between
the contaminant and the applied reagent.
In addition, the heating can be used to
condition the soil by removing the mois-
ture before or after reagent application.
Laboratory scale feasibility experi-
ments were performed to determine the
feasibility of each of the two alterna-
tives mentioned above. In addition, a new
design of an electrode structure was
tested through scale-model studies to
determine its operating characteristics
and key design parameters.
THERMAL DECONTAMINATION OF SOIL
Many hydrocarbons and organic chemi-
cals with limited solubility in water can
be steam distilled at a boiling tempera-
ture which is lower than the normal boil-
ing point of pure water or the organic
chemical contaminant. For example, a
benzene-water mixture boils at a tempera-
ture of 68.3°C as compared to 80.1°C, the
normal boiling point for benzene. The
only requirement for steam distillation is
the presence of two liquid phases in the
system. If, however, the concentration of
the organic material is very low, such
that the solubility limit is not exceeded,
then steam distillation cannot be used.
However, the gas sweep provided by steam
will provide improved vaporization rate by
continuously removing the contaminant
vapors and thus maintaining a concentra-
tion gradient between the liquid and gas
phase, which is necessary to maintain good
vaporization rate.
At various sites of interest to the
U.S. Air Force, chlorinated organic con-
taminants such as tetrachloroethylene,
dichloroethane, trichloroethylene, chloro-
benzene, etc., together with benzene,
toluene, and components of JP-4 jet fuel
are found. These materials have a boiling
range of 80° to 232°C, have substantial
vapor-pressure at 100°C, and can be steam
distilled if present in excess of their
solubility limit. For the purpose of
establishing the feasibility of thermal
recovery of such chemical contaminants,
tetrachloroethylene (120.8°C, nbp) was
selected as a representative contaminant.
Uncontaminated (clean) sandy soil obtained
from the vicinity of a waste site was
spiked with tetrachloroethylene for per-
forming the recovery experiments.
Experimental
Approach. After verifying that the soil
sample obtained from the field did not
contain chlorinated contaminants, the
moisture content of the soil was adjusted
to 5 or 10 wt%, and the soil was spiked
with 10 and 1000 ppm tetrachloroethylene.
The spiked soil was heated to a tempera-
ture range of 90° to 130°C and maintained
for a period of 4 hours. During the
treatment, the vapors were condensed and
recovered. A mass balance was established
by analyzing the condensate and the treat-
ed soil for tetrachloroethylene.
Apparatus. The apparatus for performing
the soil decontamination experiments is
illustrated in Figure 1. - The spiked soil
was placed in the round bottomed 500 ml
pyrex glass flask. The flask was heated
with a heating mantle. The vapors formed
upon heating the soil were condensed in a
water-cooled condenser and recovered
through a side leg. A Tenax® trap was
attached to the gas outlet of the conden-
ser to trap any condensed vapors.
Procedure ,
'A large batch of the sandy soil was
obtained from the field. Approximately
1000 gms of this soil was placed in a
large glass jar and tumbled overnight. A
soil sample was analyzed for the presence
of chlorinated solvents by extraction in a
Nielson-Kryger (10) steam distillation
apparatus followed by analysis of the
extract on a GC equipped with an ECD
detector. It was verified that the soil
403
-------
Vent
Tenax®
Trap
Cooli ng
Condensate
Receiver
Figure 1. Set-up for soil decontamination
experiments.
did not contain any chlorinated solvents.
The moisture content of the soil was then
adjusted to the desired level.
The clean moist soil was prepared for
experiments by placing it in the 500 ml
flask and spiking it with a known volume
of a spiking solution of known tetrachlo-
roethylene concentration. The flask was
sealed and tumbled in an ice bath for a
period of 3 hours to homogenize the soil
and the spike. At the end of the tumbling
period, the flask was attached to the
solvent recovery condenser shown in Figure
1. The side leg of the condenser was
placed in a chilled receiver containing
approximately 10 ml of pesticide grade
hexane.
The soil was heated to the desired
final temperature and maintained for a
period of 4 hours. At the end of this
time, the heating mantle was removed and
the flask cooled to room temperature. The
condenser was washed with hydrocarbon-free
water and hexane. These washings were
combined with the distillate, the Tenax®
trap was also washed with pesticide grade
hexane and the washings combined with the
distillate. The distillate was dried by
passing it through a bed of anhydrous
sodium sulfate. The dried distillate was
brought up to volume with hexane in a 100
ml volumetric flask. The distillate was
analyzed for tetrachloroethylene on a SC
equipped with an EC detector.
The glass flask containing the treat-
ed soil was attached to the Nielson-Kryger
distillation head to determine the quant-
ity of residual tetrachloroethylene pres-
ent in the treated soil. The amount of
tetrachloroethylene extracted by this
procedure was corrected by the extraction
efficiency. The extraction efficiency of
tetrachloroethylene from sandy soil was
determined by the extraction of soil
spiked at nine levels in the concentration
range of 0.1 to 1190 ppm. The average
extraction efficiency based on all nine
samples was found to be 97.11 with a
standard deviation of ±4.34.
Results
The decontamination experiments were
performed on sandy soils with tetrachloro-
ethylene concentration of 9.28 and 957.3
ppm. The results of decontamination
experiments are summarized in Tables 1 and
2. The results of all high concentration
experiments (Table 1} show that an average
recovery of 96.6% was achieved with a
standard deviation of ±1.96. It was found
that there was no significant effect of
varying the temperature in the range of
90° to 130°C. Similarly, there was no
effect on recdvery due to varying soil
moisture in the range of 5.7 to 9.85
wt%. Similar results were obtained when
soil containing 9.28 ppm tetrachloroethy-
lene and approximately 5.7% moisture was
treated at 90°and 100°C. The average
recovery of tetrachloroethylene in 5
experiments was 95% with a standard devia-
tion of ±1.5.
The results show that substantial
recovery of tetrachloroethylene from soil
is feasible in the temperature range of
90° to 130°C. The results further show
that there is no significant difference in
the amount of tetrachloroethylene recov-
ered whether the soil is heated to 90° or
404
-------
TABLE 1.
THERMAL RECOVERY OF TETRACHLOROETHYLENE
FROK SOIL (HIGH CONCENTRATION*)
TABLE 2.
THERMAL RECOVERY* OF TETRACHLOROETHYLENE
FROM SOIL (LOW CONCENTRATION)
Run
No.
918LT-6
920LT-8
913LT-3
95LT-2
92QLT-7
917LT-5
924LY-10
924LY-9
115LT-11
Ini-
tial
Ho1s-
ture,
%
5.7
8.7.
5.7
5.7
5.7
5.7
9.85
9.85
5.9
Temp.,
"C
90
90
101
101
101
131
89
102
21
Time,
hr
4.00
4.10
3.55
3.62
4.02
3.75
4.00
3.50
4.00
Residual
Concen-
tration,
ppm
58.4
52.0
29.1
18.4
56.3
20.6
1.6
25.0
1089t ppm
Recov-
ery,
X
94.2
94.9
97.1
98.2
94.5
98.0
99.8
97.6
0.4
•Initial concentration: 957.3 ppm.
^Initial concentration: 992 ppm.
130°C. The normal boiling point of tetra-
chloroethylene is 120.8°C. Thus, even at
temperatures which are 30°C below the
normal boiling point, 95% of the tetra-
chloroethylene can be recovered in a
period of 4 hours.
CHEHICAL DECONTAMINATION OF SOIL
Various researchers (3,11) have shown
that the reaction between alkali metals or
alkali metal hydroxides with polyglycols
or polyglycol monoalkyl ethers provides
reaction products that are capable of
decontaminating a host of chlorinated
hydrocarbons such as PCBs, DDT, hexachlo-
robenzene, trichlorobenzene, etc. In
particular, it has been reported (3) that
when the alkoxides produced by the above
reactions are prepared from polyethylene
glycol (PEG) or polyethylene glycol mono-
methyl ether (PEGM) of average molecular-
weight range of 350-400, good decontami-
nation is observed.
It was the purpose of this study to
determine the feasibility of in-situ
decontamination of soils containing PCB
1260 by the application of potassium poly-
ethylene glycolate (KPEG) in conjunction
with RF heating. The use of treatment
temperatures in the 90° to 140°C range was
considered to improve the reaction rate,
facilitate flow distribution of KPEG
through the soil, and to help remove the
native moisture present in the soil. Use
of high temperature for improving the
distribution and flow of KPEG through soil
Run
No.
109LTL-5
107LTL-4
107LTL-3
103LTL-1
103LTL-2
115LTL-9
Ini-
tial
Mois-
ture,
J
5.7
5.7
5.7
5.7
5.7
6.9
Temp.,
°C
89
90
99
100
101
21
Time.
hr
4.07
3.73
3.58
3.72
3.45
4.0
Residual
Concen-
tration,
ppb
242
499
461
635
525
9690*
Recov-
ery,
97.5
94.9
95.3
93.5
94.7
0.3
*In1t1al concentration: 9.28 ppm
tln1t1al concentration: 9.92 ppm
was considered because of the high viscos-
ity of KPEG at room temperature. This
problem, however, can be overcome by using
a solution of KPEfi in an appropriate sol-
vent. The use of water as a solvent is
cheaper and environmentally safer than
organic solvents. However, it has been
reported (11,12) that the decontamination
efficiency of KPEG is higher if the water
concentration is low. Thus, RF heating
can be used to boll out the native water
as well as any water added to the KPEG to
reduce its viscosity prior to application.
Laboratory scale experiments were
performed to determine the feasibility of
decontaminating soil with KPEG. In all
the experiments, 25 gms of soil spiked
with PCB 1260 was thoroughly mixed with a
KPEG solution in water. The KPEG/soil
mixture was placed in a preheated oven for
various periods of time. After treatment
in the oven, the soil was mixed with
chilled hydrochloric acid to stop the
reaction. The pH of the resultant slurry
was adjusted to a range of 6 to 7. The
slurry was dried in an oven at 100°C and
then analyzed for PCB 1260 by extracting, a
sample of the dried soil in a Soxhlet
extractor. The extract was analyzed on a
GR equipped with an EC!) detector.
Results
The results of the experiments per-
formed with 25 gms of soil premixed with
KPEG are shown in Table 3. The results
show that depending upon the treatment
conditions, 58 to 99% of the PCB present
in the soil can be decontaminated.
405
-------
TABLE 3. RESULTS Of PCB 1260 DECONTAMINATION EXPERIMENTS
Run
No.
TR7a
TR11
TR8a
TR9a
TR13
TRIOa
TR14
TR14a
TR5
TR4
TR3
Wt. of
KPEG
Solu-
tion.
gn
37.3
37.5
37.5
7.5
7.5
7.5
3.75
3.75
21
19
17
Cone.
of KPEG
Solu-
tion.
%
10
10
10
50
50
50
100
100
81.0
89.5
100
KPEG
Soil
Ratio
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.6B
0.58
0.68
Temp.,
°C
140
140
140
140
140
140
140
140
140
140
140
Tine.
hr
1.0
2.5
8.0
1.0
2.5
8.0
2.5
2.5
2.5
2.5
2.5
Destruc-
tion,
I
65
79
100
63
58
100
84
99
97
99
96
The effect of heating time was stud-
led by running the experiments for a
period of 1, 2.5, and 8 hours. These
experiments used approximately 10 times
the stoichiometric amount of KPEG neces-
sary to react with all the chlorine atoms
on PCB 1260. It was observed that on in-
creasing residence time from 1 to 8 hours
(runs TR7a, TR8a, TR9a, TR11, TR13, and
TRIOa) the percentage destruction of PCB
1260 increased from 64 to 100%. In these
runs, the ratio of KPEG/to soil was 0-15.
The results of these experiments show
that decontamination of soil containing
PCB 1260 is feasible by treatment with
KPEG, in reasonable time and temperature
ranges.
PRINCIPLES OF RADIO FREQUENCY HEATING
The term radio frequency (RF) gener-
ally refers to the frequencies used in
wireless communications. These frequen-
cies can be as low as 45 Hz or extend well
above 10 GHz. The frequencies primarily
used for radio frequency, dielectric, or
microwave heating range from 6.78 MHz to
2.45 GHz.
The principles of radio frequency
heating are similar to those of a micro-
wave oven, except that the frequency of
operation is different and the size of the
application is much larger. In these
systems, the temperature rise occurs due
to ohmic or dielectric heating mechanisms.
Ohmic heating arises from an ionic current
or conduction current that flows in the
material in response to the applied elec-
tric field. This is similar to the cur-
rent that flows in a lightbulb or in
resistive heating elements.
Dielectric heating results from the
physical distortion of the atomic or
molecular structure of polar materials in
response to an applied electric field.
Since the applied AC electric field chang-
es rapidly, the alternating physical
distortion dissipates mechanical energy
which is translated into thermal energy in
the material.
The dielectric properties of soil
determine the amount of RF power that can
be dissipated in soil. These properties
are relative dielectric constant (er) and
the loss-tangent. The loss-tangent, tan 6,
is defined as a/
-------
breakdown, corona discharge, or other
undesirable effects can result. Thus,
there is an optimum range wherein a suit-
able frequency can be chosen for a given
volume of material and set of dielectric
parameters.
IN-SITU RF HEATING SYSTEMS
A fully operational in-situ RF heat-
Ing system for decontamination requires
the development and testing of at least
four major sub-systems. These are: (1) RF
energy deposition electrode array; (2) RF
power generation, transmission, monitoring
and control system; (3) vapor barrier and
containment system; and (4) gas and liquid
condensate handling and treatment system.
Among the sub-systems mentioned
above, the electrode array (also called
the exciter array) design is the critical
item which will drive the design require-
ments and constraints for the other three
sub-systems.
EM Exciter Array Designs
Previous attempts (1,6,8,13) to use
electrical energy for heating of earth
formations were aimed at resource recovery
from hydrocarbonaceous deposits. Simple
techniques such as burying electrical
heating elements or a pair of electrodes
to which 60 cycles AC power is applied
were not successful due to two main
reasons: (1) nonuniform heating leading
to unacceptable levels of energy ineffi-
ciency; and (2) inability to heat beyond
the boiling point of free water. To
overcome the temperature limitation im-
posed by 60 Hz heating, antennas radiating
very high frequency or microwaves have
also been considered (1). Though these
methods provide rapid volumetric heating
even above the boiling point of free
moisture, they suffer from inefficient use
of the applied energy.
To overcome these limitations, it is
necessary to use bound-wave exciters as
opposed to the radiated wave horns or
antennas previously used. The bound-wave
exciters are designed to fully contain the
EM radiation within a defined volume of
soil. There are two basic types of bound-
wave exciter arrays. These are the tri-
plate line and the fringing-field trans-
mission line.
Trip!ate Line
The triplate transmission line is the
rectangular analogue of the more familiar
cyclindrical coaxial cables. The triplate
line is formed by a fully enclosed rec-
tangular cavity in which a central planar
conductor, parallel to the large sides of
the cavity has been inserted. Clearly
emplacement of solid metal plates that
enclose a rectangular cavity below the
soil surface is impractical. This problem
has been resolved (14) by simulating the
fully contained rectangular cavity by
inserting an array of electrodes in bore-
holes drilled through the soil. The elec-
trodes are Inserted in three parallel rows
which represent the-two outer walls and
the central conductor of the fully con-
tained rectangular triplate. It has been
demonstrated that through appropriate
selection of the row-spacing and the
spacing of the electrodes within each row,
it is possible to fully contain the ap-
plied electromagnetic field within the two
outer rows of electrodes.
The triplate line is suited for those
soil decontamination applications where
the contaminants have penetrated more than
3 feet below the surface and for contami-
nants that require treatment temperatures
in excess of 130° to 150°C.
Figure 2 illustrates a conceptual
design of such an array which is covered
by a vapor barrier to contain and.collect
gases and vapors as they rise from the
surface of the soil. A triplate line was
tested by Krstansky et al. (7) by heating
a tar sand formation to 200°C.
Fringlng-Fleld Line
This concept makes use of the fring-
ing fields or leakage fields that exist
near a transmission line. Figure 3 illus-
trates a schematic diagram showing rows of
horizontal electrodes placed over.the
surface of the soil. Fringing fields are
formed around the electrodes when they are
energized. Thus EM energy is first ab-
sorbed by the moist layers of soil nearest
the electrodes. As the moisture is driven
out of the.soil, the energy is selectively
absorbed at greater depths. The heating
zone is eventually restricted because of
the exponential fall-off of the fields in
to the soil.
407
-------
RF Power
Source
Barrier
Exeiter Hectrodes
Cas And Vapor
Treatment System
Figure 2. In-sltu treatment module for area 10,000 ft2, depth 8 ft.
Electrodes
120 in.
Contaminated
Soil
Figure 3. Schematic of a field demonstration test
with fringing field line (size 10 x 10 ft).
Figure 4 illustrates a 1:3 scale
model of a single electrode from the
fringing-field line shown in Figure 3.
This model was tested by packing soil to a
depth of 14 in. below the electrode.
Thermowells were installed on the Inner
wall of the electrode and at various
depths below the surface. Figure 5 shows
the temperature rise in the soil as a
function of time and depth. After 27.5
hours of heating, the average temperature
at depths of 4 in. and Sin. was 110° and
88°C, respectively. At this time, 27.5
Ibs of 90% KPEG/H 0 solution was poured
over the soil surface. The heating was
resumed for 2.5 hrs to verify that it is
still feasible to match into the load and
to obtain impedance data on soil contain-
ing KPEG.
408
-------
14 In.
METAL Box
ELECTRODE
SOIL
Figure 4. Scale model of a single-electrode
module. (Based, on the 10 x 10 ft field
demonstration -scale.)
The results of this test show that
the fringing field line is suitable for
the treatment of soils in which the depth
of contaminant penetration is less than 3
feet and where the required treatment
temperatures do not exceed 130°C. Addi-
tional development and testing is, how-
ever, necessary to fully explore the
limitations of the fringing field lines,
and, in particular, to determine ways to
improve the penetration depth.
COST ESTIMATE
The cost of treating a waste site
containing low boiling contaminants such
as tetrachloroethylene, chlorobenzene,
etc., was developed.
The following assumptions were made
to develop the cost estimate,
« typical treatment site is 3 acres in
area, with contamination to a depth of
8 ft.
• costs were developed for fully tested
and mature technology. There is no
allowance for the recovery of research,
development, promotional, or management
costs.
The site will be treated by the pro-
gressive treatment of 96 x 96 ft mod-
ules.
The base case is for a treatment tem-
perature of 170°C at a site containing
12% moisture level.
The cost was developed by separately
estimating the fixed and variable operat-
ing costs, the capital cost, and finance
charge on capital. The capital cost was
depreciated over a period of 20 years, and
it was amortized over the number of 96 ft
x 96 ft modules treated over the life of
the equipment.
It is estimated that each module can
be treated in a period of 45 days which
includes time for set-up, heating, cool-
down, and removal of equipment. A double-
stream treatment system was designed which
was based on one 1.0 MW RF power source
and a duplicate set of all other equipment
items. This would allow one module to be
under RF heating cycle while a second
module is being prepared for subsequent
heating. With this scheme of selective
duplication of key capital equipment, it
should be possible to treat 14 modules per
year.
The capital cost estimate for the
treatment system is shown in Table 4. The
TABLE 4. CAPITAL COST FOR A RF SOIL TREATMENT SYSTEM
(THERMAL TREATMENT OPTION)
1.0
Sub-System
MH RF power source.
Quantity
1
Total Cost,
$
1,000,000
dummy load, transmission
and control system
Vapor barrier and gas 2
collection system
Gas, vapor, and condensate 2
treatment system
Analytical Instrumentation 2
and safety equipment
96,000
300.000
100,000
Tools, etc.
Back-hoe
Utility vehicles
Electrode system for
2 modules at 2650 ft/module
1
1
2
2
10,000
40,000
40,000
10,600
TOTAL CAPITAL COST
1,596,000
409
-------
130
110
90
70
50
30
SOIL*. 290 Lis- OP SANDY Soil.
KPE6: 27-5 IBS OF 901 KPES/HjO SOLUTION ADDED
EOUIPMEHT
FAILURE
KPES ADDITION
CENTER CONDUCTOR
" 1 IN. DEPTH
* 8 IN. DEPTH
_L
J_
_L
_L
J_
8
12
28
32
16 20 21
TIHE (HOURS)
Figure 5. Soil temperature rise, fringing-field line {Test #2).
total cost of a treatment system is ap-
proximately $1.6 million. Based on an
average finance charge on capital of 20%
per year, $327,880 is the annual charge
for capital cost. Based on 14 modules
treated per year, the finance charge per
module 1s $23,420.
The fixed operating cost for the base
case was estimated at $30,640 which
includes the cost items which are indepen-
dent of the treatment duration. These
costs are: module site preparation,
equipment mobilization, installation, tear
down and decontamination, licenses, per-
mits, environmental monitoring, replace-
ment electrodes and spares, etc. The
variable operating cost for the system
consists of all cost incurred during the
heating period. The variable operating
costs include AC power for RF generation,
operating labor, and consummable supplies
necessary for the operation. The estimat-
ed AC power cost for heating.in the base
case is $35,500, which includes 25% allow-
ance for heat loss. The other variable
operating costs were estimated as $2788/
day of heating, or $66,900 for 24 days of
heating time in the base case.
The total cost of treatment of a
module for the base case is $188,000 which
was obtained by summing up the four cost
components discussed above, and by adding
a 20% contingency allowance. This is
equivalent to $42.40 per ton of treated
soil.
A sensitivity study was performed to
determine how cost varies with temperature
and moisture content. A temperature range
of 100° to 250°C and soil moisture range
of 5 to 20% was considered. It was found
that the treatment cost varies in the
range of $29.40 to $57.40 per ton.
CONCLUSIONS
The laboratory feasibility study has
shown that 95% of the tetrachloroethylene
present in the soil can be recovered by
heating to a temperature range of 90° to
130°C in a period of 4 hours. It was also
shown that decontamination of PCB 1260 is
feasible by reaction with KPEG at a tem-
perature of 140°C in a period of 2.5 to 8
hours. There is some evidence (not pre-
sented here) that PCB decontamination with
410
-------
KPEG is feasible at even lower tempera-
tures. This should be further investi-
gated. Soils containing these materials
can be heated to the required treatment
temperatures using the triplate or the
fringing field lines. The depth of pene-
tration of the soil and the required
temperature determine the line used for
any given application.
The successful development of a
large-scale in-situ RF heating method for
soil decontamination will have the fol-
lowing benefits and advantages:
• true in-situ treatment of most organic
substances on EPA list of priority
pollutants
• deli sting of the waste site will not be
necessary
• safe containment of removed hazardous
waste
• controlled rate of waste removal with
turn-down capability
• excavation or digging through the
contaminated volume is not required
• major capital equipment will be mobile
and acquisition costs may be amortized
through multiple usage at numerous
sites
• Proper design and equipment selection
will allow treatment of waste sites
from 10,000 ft2 to one acre in area,
having contaminant penetration of 1 to
20 ft.
» Preliminary studies (4) show that the
process has significant cost advantages
over the alternative treatment method
of excavation and incineration of soil
in an approved incinerator.
ACKNOWLEDGMENT
The work reported herein'was support-
ed by the U.S. EPA, USAF, and IIT Research
Institute through a U.S. EPA co-operative
agreement CR-811529-01-0. The author
acknowledges the technical direction and
support given by Mr. Charles Rogers, U.S.
EPA, Cincinnati, Ohio, Mr. Doug Downey,
USAF, HO AFESC, Tyndall AFB, and Messrs.
Jack Bridges, Gug Sresty, Joe Enk, and Jim
Krstansky of I IT Research Institute.
REFERENCES
1. Abernathy, E. R., 1974. Production
increase of heavy guide oils by
electromagnetic heating. Paper
374007 presented at the 1974 Annual
Technical Meeting ofthe' Petroleum
"Soci ety of CIM, Cal gary, Al berta.
2. Barnhart, Benjamin"J., 1979. TRe
Disposal of Hazardous Waste, Env.
Scien. Technol.
3. Brunella, D. J. and D. A. Single-
ton, 1983. Destruction/Removal of
Polychlorinated Biphenyls From
Nonpolar Media. Reaction of PCB
with Polyethylene 61ycol/Potassium
Hydroxide. Chemosphere, 12, No. 2,
pp. 183-196.
4. Dev, H., Jack E. Bridges, Guggilam
C. Sresty, 1984. Decontamination of
hazardous waste substances from
spills and uncontrolled waste sites
by radio frequency in situ heating.
Hazardous Materials Spills Confer-
ence,Tennessee,April 1984, pp. 57-
WT
5. Dev, H., 0. Bridges, and D. Clark,
- 1985* Rad1o Frequency Enhanced
Decontami nati on of Sol Is.Con'tan 1 nat-
ed^ wiW Hal ogenated Hydrocarbon^
Interim Report, IITRI Mo. 'COesOO,
U. S. EPA Grant No. CR-811529-01-0,
IIT Research Institute, Chicago,
Illinois, 60616.
6. Flock; K. L., 1975. Unconventional
methods of recovery of bitumen and
related research areas particular to
•' the oil sands of Alberta, Journal
of Canadian Petroleum Technology,
pp.17-27 (July-Sept.'1975).
7. Krstansky, J., et a!., 1982. RF
Heating of Utah Tar Sands, Final"
'Report, ;I ITRI E06482, DOE Contract
No. DE-AC01-79ER10181. IIT Research
Institute, Chicago, Illinois, 60616.
8. Ljungstrom, F., 1951. • Skifferolje-
•i fragen-Galma Och Nya Sgnpunker
Tekuisk. Tekm'sk Tidskrift, 81, pp.
33-40.
411
-------
9. Hurray, Chris, 1979. Chemical Waste 13.
Disposal, A Costly Problem, C&EN,
March 12.
10. Onuska, F. I. and K. A. Terry,
1985. Determinatin of chlorinated
benzenes In bottom sediment samples 14.
by WCOT column gas chromatography.
Analytical Chemistry, 57, No. 4, pp.
#01-805.
11. Pytlewskl, L. R. et al., 1980. The
Reaction of PCBs with Sodium Oxygen
and Polyethylene Glycols. Proceed- 15.
Ings of the 6th Annual Symposiurn on
Treatment of"Hazardous"Waste, EPA
Report No. EPA-60Q19-80-01I.
12. Rogers, C. J., 1983. Chemical
Treatment of PCBs in the Environ- 16.
ment. Proceedings of the 8th Annual
Symposium on the Treatment of
Hazardous Waste, EPA Report EPA-
600/9-83-003, pp. 197-201.
Salomonsson, G., 1953. Underjordisk
Skifferpyrloys Em'gt Lungstrom-
Metoden. Tidskrift For Tekriiskl-
Vetenskap'llg For SKninj 24:3, pp.
118-123.
Sresty, G. C., Harsh Dev, Richard H.
Snow, and Jack E. Bridges, 1986.
Recovery of bitumen from tar sand
deposits with the radio frequency
process. SPE Reserve1r Englneerln.g»
Vol J_, No. 1, pp. 85-94.
U. S. Environmental Protection
Agency, Office of Public Awareness,
1978. SolidWaste Facts - A Statis-
tical Handbook.OPA113/8, Washing-
ton, August.
U. S. Environmental Protection
Agency, Office of Water and Waste
Management, 1980. Everybody's
Prob!em: Hazardous Waste. Waste
AT ert Serl es SW-826", Wash 1 ngton.
412
-------
THE ROLE OF ROSUE DROPLET COMBUSTION
IN HAZARDOUS WASTE INCINERATION
James A, MulHolland
U. S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
and
Ravi K. Srivastava and Jeffrey V. Ryan
Acurex Corporation
Durham, NC 27713
ABSTRACT
The trajectories of combustible liquid droplets injected in single, monodisperse
streams into swirling, turbulent gas flames were measured in a 100 kilowatt combustor (50
centimeter internal diameter) as a function of the droplet injection parameters of size,
velocity, spacing and angle. Cold flow tests were performed to measure droplet drag as a
function of Reynolds number and droplet spacing. Using a general three-dimensional algo-
rithm for modeling drag during the initial heating up and evaporation of a burning droplet,
predictions of droplet trajectories are compared with the experimental observations. It is
found that physical processes can dominate chemical kinetic effects during large droplet
combustion, resulting in droplet bypassing and penetration of the flame zone. In the
incineration of liquid hazardous wastes, where a destruction removal efficiency (ORE) of
greater than 99.99 percent is required, lack of complete combustion of a very few individu-
al droplets can result in an incinerator failure. Thus, predicting the fate of these rogue
droplets is of paramount importance in understanding limits to ORE in the thermal destruc-
tion of hazardous wastes. Further work is necessary to incorporate turbulence effects into
the model and to correlate droplet penetration with droplet destruction efficiency.
INTRODUCTION
The combustion of liquid fuels in
most practical swirled combustors is
dominated largely by the overall dynamics
of the fuel spray and by the combustion
of particle clouds, rather than by the
trajectories of single droplets (1). How-
ever, in the incineration of liquid haz-
ardous wastes, the situation is different.
Here, a destruction removal efficiency
(ORE) in excess of 99.99 percent is
required, and the penetration or bypass-
ing of the flame zone by a very few
individual droplets can lead to a failure
mode. For example, the escape of one
rogue 300 micrometer (urn) droplet out of
10 million droplets, with a mean diameter
of 30 wm, will lead to a ORE of less
than 99.99 percent and, thus, to a failure
of the incinerator. Measurements of drop-
let size distributions in various fuel
sprays have shown that droplets are pres-
ent with diameters of 100 to 500 urn, which
are an order of magnitude larger than the
mean (2). These rogue droplets tend
to concentrate at the outer edge of the
spray co.re. The subsequent fate of these
individual drops is then of paramount
importance. This research, therefore,
focuses on the measurement and prediction
of trajectories of individual large drop-
lets in hot, swirling practical combustor
flow fields. The objective is, first, to
determine the extent to which the behavior
of a rogue droplet in a practical aero-
dynamic flow field can be predicted using
existing models and, second, to delineate
413
-------
the requirements necessary for a model
that can ultimately allow the relationship
between nozzle atomlzation quality and
ORE to be determined.
Droplet burning is fundamental to
many combustion systems and has received
much attention (3-5). Theoretical solu-
tions have been derived for both single
droplet and group combustion by assuming
idealized combustion environments (6-7).
Thus far, however, these models, with
their simplified (external) fluid mechan-
ics, have been unsuccessful in predicting
experimental results from practical com-
bustors (8). This work complements the
work of others, describing the application
of a simplified three-dimensional droplet
trajectory and evaporation model to predict
the experimental results of large single
droplets in complex swirled combustion
flow fields.
EXPERIMENTAL METHODS
Cqmbusto r and D roplet Generator
The bench-scale combustion facility
is shown in Figure 1. The 100 kilowatt
(kW) combustor is a horizontal cylinder
with a large quartz window for flame
visualization and multiple ports for
in-flame sampling. A movable-block swirl
burner (9) with interchangeable fuel
nozzles is mounted on the combustor and
is capable of providing near-burner zone
aerodynamic simulation of various flame
types. The front wall contains a slot to
allow for off-axis droplet injection, as
shown (Figure 1).
Natural gas was burned at an excess
air of 30 percent, ensuring that suffi-
cient oxygen was available for droplet
combustion. Two flame shapes with very
different mixing patterns were estab-
lished: Type C and Type A (10). The Type
C flame, produced with a high combustion
air swirl (swirl number=1.3) and a 45 de-
gree fuel nozzle, was 10 to 15 centimeters
(cm) long and was stabilized on the bur-
ner quarl by strong primary internal re-
circulation of the hot combustion gases.
Most droplet trajectories were measured
in the Type C flame. The Type A flame,
produced with a minimum combustion air
swirl for flame stabilization (swirl
number=0.65) and axial fuel injection,
was 50 to 60 cm long. The Type C flame
is most typical of practical, high
efficiency combustion systems; therefore,
this flame type was used as the nominal
test condition.
The droplet generator is a vibrating
orifice device (11), with ancillary elec-
tronics to facilitate droplet-to-droplet
spacing variation (12). A mixture of 80
percent (by volume) Shell fuel additive
EXHAUST
WATER-COOLED
CONVECT1VE
SECTION
HIGH VOLTAGE
POWER
SUPPLY
FREQUENCY
DIVIDER
•
DC
POWER
SUPPLY
SIGNAL
GENERATOR
AXIAL
SAMPLE
PORT
NEAR-BURNER
COMBUSTION
MODULE
Figure 1.
CHARGING
DEFLECTION PLATES
TRAP
NATURAL
GAS
VARIABLE
SWIRL
RADIAL BURNER
SAMPLE
PORTS
Schematic of bench-scale combustion facility. The 50 cm internal diameter, 90
cm long near-burner combustion module is water-cooled and insulated with a 4 cm
thick refractory lining. The 1FRF burner has a movable-block air swirl generator
and interchangeable fuel nozzles. The monodisperse droplet generator is,shown
with electronics for droplet separation.
.414
-------
ASA-3, comprised mostly of xylene, and 20
percent distillate fuel oil was used as a
nominal test liquid. This fluid was
selected on the basis of conductivity
(for electrostatic charging and deflec-
tion), viscosity (for droplet formation)
and luminosity 1n the flame (for visual-
ization). Droplet parameters varied were
spacing, size, velocity and injection
angle. Spacing was varied by charging,
then deflecting, every Nth droplet, with
the neutral droplets being trapped and
recirculated (12). Size was varied by
changing orifices, with droplet diameter
found to be approximately twice the
orifice diameter. Droplet diameter, D,
and droplet spacing in the stream, A
-------
consistent with the Input air velocity at
the burner throat of 28 m/s. The calcu-
lated bulk mean axial velocity, approxi-
mately 2.6 m/s, 1s approached beyond the
50 cm axial distance.
Mean radial velocities were found
never to exceed 8 m/s (Figure 3b). Very
little radial motion of the combustion
gases was seen beyond the 25 cm axial
location. Mean tangential velocities
were found to be negative everywhere,
consistent with the direction of swirl
(Figure 3c). A strong degree of swirl
was observed near the burner quarl exit,
with the swirl mostly dissipated by 40 cm
downstream. The peak tangential velocity
of approximately 12 m/s is consistent
with a swirl number of 1.3.
Droplet Trajectory Measurements
Fourteen test conditions were evalu-
ated, with droplet trajectories measured
for variations in droplet spacing, veloc-
ity, size and injection angle (Table 1).
The domain of droplet trajectories
observed for the nominal test condition
1s shown (Figure 4) in the two planes of
observation. Superimposed on the cross-
sectional view along the centerline axis
are lines of constant 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 combustion gas. This
point, labeled "IS," denotes the location
of the earliest ignition point observed.
The burning times were very short, and
the burning droplet "flecks" lasted only
over a few centimeters. The large shaded
area denotes the distribution of burning
droplets, igniting and burning out at
various locations in this area. Thus,
there is no single trajectory that de-
scribes this test condition. The average
burnout point of the droplets that pene-
trated axial1y the farthest is marked and
labeled "BO." The existence of a domain
of droplet stream burnout locations rather
than a single burnout point suggests that
either the droplet injection properties
were nonuniform or the gas flow field
fluctuated with time (i.e., turbulence).
Uniformity of droplet formation and
injection was carefully monitored during
testing. Therefore, the range of trajec-
tories observed in the experiment is at-
tributed to the high combustion gas
turbulence that was observed.
20
10
s
0
MEAN AXIAL VELOCITY, m/t
«.M«<
10 20 30 40 50 60
AXIAL DISTANCE, cm
(a) Axial velocity map.
u>
MEAN RACIAL VELOCITY, mil
-5
10 20 30 40 SO 60
AXIAL DISTANCE, cm
(b) Radial velocity map.
0 10 20 30 40
AXIAL DISTANCE, cm
(c) Tangential velocity map.
Figure 3. Near-burner field velocity
maps for Type C flame base-
line. The lines of constant
velocity are constructed from
three-dimensional axial and
radial velocity profiles.
Symmetry about the center-
line axis was verified and 1s
assumed for the entire field.
416
-------
TABLE 1. DROPLET CONDITIONS AND TRAJECTORY ENOPOINTS
Test
Point
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Test
Variable
nomi nal
spacing
velocity
size
angle
flame
Droplet
Vel.
(n/s).
7.0
7.0
4.9
9.1
8.5
8.5
7.0
7.0 x
Injection Parameters
Dia.
(vmi)
210
210
180
210
250
315
210
210
Spacing
(urn)
510 x 64
510 x 1
x 2
x 4
x 8
x 16
340 x 64
640 x 64
620 x 64
615 x 64
510 x 1
510 x 64
510 x 64
510 x 64
Angle
(deg.)
30
30
30
30
0
0
45
30
Flame
Type
C
C
C
C
C
A
Ijjniti on/Burnout Point
Axial
(cm)
15/45
35/75
30/65
10/55
10/45
25/40
15/25
30/60
35/60
45/70
20/75
20/45
25/50
10/35
Radial
(cm)
7/15
8/25
5/23
6/20
6/18
8/11
5/10
8/15
8/13
10/20
16/16
16/16
8/12
7/5
Angle
(deg.)
335/245
240/200
240/200
330/250
330/240
270/260
330/270
300/225
270/225
270/225
330/330
330/330
270/270
350/260
Axial droplet penetration distances
are shown as a function of the test param-
eters in Figure 5 for all 14 test condi-
tions. In Figure 5a, closely spaced drop-
lets are shown to penetrate farther than
isolated drops. Little effect of spacing
is observed for values of greater than 20
diameters. Axial penetration is shown to
roughly double for a doubling of initial
droplet velocity (Figure 5b). Axial pene-
tration increases as droplet diameter 1n-
Q
ac
creases, as well (Figure 5c). The effect
of injection angle is observed for both
the isolated droplet and the droplet
stream (Figure 5d). Little change in axi-
al penetration distance is observed
either for isolated droplet injection at
0, 30 and 45 degrees or for droplet
stream Injection at 0 and 30 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 the
penetration of the droplet (Figure 5e).
10
20
20 40 60
AXIAL DISTANCE, cm
(b) End view.
Figure 4.
Droplet trajectory for the nominal test condition. The domain of droplet burnout
locations is shown by the shaded area. An axial velocity profile is superimposed
in the top view drawing. The injection point (IN), the first ignition point (IS)
and the last burnout point (BO) are noted.
417
-------
80
§
. 60
Ul
g 40
s
£ 20
S
80
O-
°6
80
8
uJ 6°
I
g 40
o
I 20
Z
12345
' ln{t|/D°)
(a) Droplet spacing.
2 4 6 8
VELOCITY, m/s
(b) Droplet velocity.
10
80
~
S 40
0 100 200 300 400
DIAMETER, pin
(c) Droplet size,
THEORETICAL MODEL
A simple numerical model to predict
the trajectory of a single stream of drop-
lets injected Into swirled combusting
flow fields was used to provide comparison
with the experimental results and to
§uide future experimental efforts.
Description
The model is a direct adaptation of
the numerical procedure developed by Ayers
et^al.(8) and is used to solve the un-
coupTed equations of droplet motion in a
Lagrangian framework. A three-dimensional
grid structure is established for specify-
40
13
EXPERIMENT
15
75
30 45 60
ANGLE, deg.
(d) Droplet Injection angle.
90
80
,.? 60
£2 40
C A
FLAME TYPE
(e) Flame type.
Figure 5. Comparison of model prediction
and experimental results of
axial penetrations. A listing
of the input droplet conditions
is given in Table 1. Each
plot represents variation of a
single parameter. The droplet
trajectory for test condition
1 is shown in each plot because
it is the nominal test condi-
tion droplet path.
ing the (measured) background gas veloc-
ity, temperature and chemical speciation,
A single stream of spherical, homogeneous
droplets enter the computational domain
at a specified position with specified
(experimental) initial values for droplet
diameter, spacing, temperature and veloc-
ity. The process is modeled in two steps:
droplet heating to the liquid boiling
point followed by droplet evaporation.
Calculations are terminated when complete
evaporation occurs, or when the droplet
exits the computational domain. Drag,
calculated using an empirical coefficient,
and gravitational forces are assumed to be
the only external forces on the droplet.
418
-------
It was hypothesized that droplet-to-
droplet spacing would affect drag. Cold
flow tests were performed to evaluate
this effect in a nonevaporating, quiescent
environment. These data will be published
elsewhere. An empirical formulation of
the drag coefficient, CD, as a function of
Re and £d/D was derived that predicted
all of the cold flow test results.
During droplet heating to the liquid
boiling point, its diameter is assumed to
be constant (i.e., no evaporation). Here,
the time rate of change of the droplet
temperature is given by
dT
— = 6
dt
NuX
D2c
(Too - T).
(3)
The Musselt number (Nu) is calculated
using an empirical relation (13). After
the droplet temperature reaches the
boiling point, evaporation begins. In
this stage of pre-ignition, the time rate
of change of the droplet diameter is
given by
dD 1
— = -CB — (1 + 0.23 Re u-3), (4)
dt 20
0.5,
where the evaporation constant (CR) de-
pends on local gas properties (14).
Comparison with Experimental Results
It was observed that, in the initial
portion of a large droplet lifetime, the
droplet injection parameters dominate its
trajectory. For a given test condition,
the droplets tracked an approximately
common straight path until evaporation
began. Then, as the droplet size (and,
thus, momentum) decreased, the turbulent
gas flow field became increasingly domi-
nant, resulting in the break-up of the
droplet stream (Figure 4). Theoretical
model calculations predict these two
modes as well. The rate at which the
droplet approaches the gas velocity is a
function of droplet size. Initially, the
predicted droplet trajectory is insensi-
tive to combustion gas flow field veloci-
ties. As evaporation occurs, sensitivity
to gas velocity increases, and the droplet
is transported by the mean gas flow field.
However, since the current model does not
include turbulence effects, the distribu-
tion of trajectories for a given condition
is not predicted.
The predicted axial location at which
the droplet is completely evaporated is
compared to the experimentally measured
droplet penetration in Figure 5. While
the predicted radial and circumferential
distances travelled before complete drop-
let evaporation closely agree with exper-
imental results, the model predictions
are all much less than the axial penetra-
tions observed in the experiment. How-
ever, predicted trends match those meas-
ured experimentally for droplet injection
size, velocity and angle. The model pre-
dicts only a very slight increase in axial
penetration as initial droplet spacing de-
creases, due to reduced droplet drag. It
is hypothesized here that decreasing drop-
let spacing delays droplet ignition and
burning rate, resulting in a significant
experimental trend. No model prediction
was made for test condition 14 as the Type
A flame baseline data had not been reduced
at the time of this publication.
Model Predictions
In order to determine both the effect
of centerline injection and the signifi-
cance of rogue droplets, the model is
further exercised. Two trajectories were
calculated for each of two angles of
injection: 0 and 45 degrees.
A droplet with a size (50
and velocity (50 m/s) typical of mean
droplets in actual sprays is considered
first, with a Type C flame gas flow field
input. The axial penetration is predicted
to be 3.8 cm for the axially injected
droplet and 4.5 cm for the droplet in-
jected at 45 degrees. Therefore, both
predictions indicate that this droplet
would be evaporated in the flame zone.
Next, large rogue droplet (300 urn)
trajectories are considered. In actual
sprays, the large droplets would be ex-
pected to have velocities that are less
than the mean, since the pressure drop
across a nozzle is fixed. For a large
droplet velocity set at 10 m/s, axial
penetrations of 44.3 and 17.1 cm are pre-
dicted for the axially and 45 degree
injected droplets, respectively. Thus,
the large droplets with reduced veloci-
ties penetrate significantly farther than
mean droplets.
419
-------
DISCUSSION AND CONCLUSIONS
REFERENCES
Large droplet penetration and by-
passing have been observed in both experi-
mental tests and model calculations. The
short burning distances and relatively
long trajectories of large droplets ob-
served 1n the experiment suggest that
droplet trajectory pnior to ignition is
primary, and gas phase combustion second-
ary, 1n understanding and predicting ORE.
Future model development must address
two critical areas to more accurately pre-
dict the absolute penetration of droplets,
in addition to predicting trends, and to
predict droplet scattering. First, some
of the simplifying assumptions in the cal-
culation of droplet heat transfer rate need
to be relaxed. Droplet-to-droplet inter-
actions in the heat transfer to a droplet
can be approximated using group combustion
theory to make a better estimation of
flame standoff distance. Also, droplet
Ignition effects on droplet evaporation
can be estimated. The second critical
area to be addressed relates to the role
of turbulent fluctuations in practical
combustor flow fields. The model used in
this paper predicts mean droplet trajec-
tories using mean flow fields in swirling
combustion environments. These experiments
have shown that there is a need to model
turbulence as well. One possible approach
would be to use a normal distribution of
velocity fluctuations, based on measured
mean and standard deviation values for gas
velocity components, to randomly select
local velocities. Many droplets would be
tracked until evaporation, resulting in a
distribution of trajectories that could be
described statistically.
A method for measuring droplet in-
cineration and correlating it to droplet
penetration is needed. It is planned
that this work be extended to study mul-
tiple streams and sprays. The final
product may then allow for the prediction
of nozzle failure modes based on an under-
standing of rogue droplet combustion.
ACKNOWLEDGMENT
This research was conducted in the
U. S. Environmental Protection Agency's
Air and Energy Engineering Research Labo-
ratory at Research Triangle Park, NC. All
financial support was provided by EPA.
1. Chigier, N. and C. 6. McCreath,
1974. ActaAstronaut., j[, p687.
2. Kramlich, J. C., M. P. Heap, W. R.
Seeker and G. S. Samuel son, 1984.
Twentieth Symposium (International)
on Combust i on, (i n press), The
Combustion Institute.
3. Spalding, D. B., 1952. Fourth Sym-
posium (International) on Combustion,
The Combustion Institute, p847.
4. Williams, A., 1973. Combust. F1ame,
21, pi.
5. Faeth, G. M., 1977. Comb. SciI. Tech.,
21_, p!23.
6. Law, C. K., 1982. Prog I Eng. Comb.,
8, p!69.
7. Chiu, H. H., R. K. Ahluwalia, B. Koh
and E. J. Croke, January 1978. Spray
Group Combustion. Paper presented
at the Sixteenth Aerospace Sciences
Meeting, Huntsville, Alabama.
8. Ayers, W. H., F. Boysan and J.
Swithenbank, September 1980. Drop-
let Trajectories in Three-Pimensional
Gas Turbine Flow Fields,AirForce
Report AFOSR-TR-81-0543.
9. Beer, J. M. and N. A. Chigier, 1963.
Swirling Jet Flames fromanAnnular
Burner, International Flame Research
Foundation, Doc. No. K2Q/a91.
10. Leucke1,W., 1968. Swirl Intensi-
ties, Swirl Types and Energy Losses
of Diffgrent Swi rl Generating De-
vices.International F1ameResearch
Foundation, Doc. No. G02/al6.
11. Sangiovanni, J. J. and M. Labowsky,
1982. Combust. Flame, 47, p!5.
12. Russo, R., R. Withnell and G. Hieftje,
1981." Appl. Spectrosc., 35, p531.
13. Ranz, M. E. and W. R. Marshall, 1952.
Chem. Eng. Prog., 48_, 141, p!73.
14. Wise, H. and G. A. Agoston, 1958.
Literatureonthe Combustion of
Petroleum. American Chemical
Society, Washington, D.C.
420
-------
CHARACTERIZATION OF HAZARDOUS WASTE INCINERATION RESIDUALS
Don Van Buren, Gary Poe, Carlo Castaldlni
Acurex Corporation
Energy & Environmental Division
Mountain View, California 94039
ABSTRACT
The Office of Solid Waste and Emergency Response (OSWERi-EPA) is considering
establishing a criterion for disposal of waste or residue into the land. This criterion
is based on the achievement of residue quality equivalent to that from effective
incineration. The purpose of this study was to provide data on the quantities and
characteristics of solid and liquid discharges from hazardous waste incineration
facilities. A total of 10 facilities were sampled comprising major incineration designs
and flue gas treatment devices. All inlet and outlet liquid and solid streams were
sampled and subjected to extensive analyses for organic and inorganic pollutant
concentrations. Laboratory analyses for solid discharge streams also included leachate
evaluations using standard EPA toxicity tests for metals and new EP III toxicity
procedures for volatile and semi volatile organics and metals. Monitored data on
incinerator facility operation was then used to determine the discharge rates of detected
pollutants. This paper presents only the portion of the study associated with incinerator
ash residuals. .--.-.
INTRODUCTION
Under the amendments to the Resource
Conservation and Recovery Act (RCRA) that
were passed in 1985, the Environmental
Protection Agency (EPA) is required to ban
the land disposal of many hazardous wastes
unless it can be proved that such wastes
can be safely disposed of to the land.
Incineration has been proven to be an
effective method for the destruction of
many hazardous wastes. OSWER is
considering establishing the criterion
that the achievement of residue quality
equivalent to that from effective
incineration will be required before a
waste or residue will be allowed to be
disposed of into the land. EPA's Office
of Research and Development (ORD) has
characterized stack gas emissions from
hazardous waste incinerators under a
previously conducted field testing program
to support OSWER's regulation development
process. This testing, conducted at eight
full-scale operating incinerators, was
directed at assessing the incinerators'
ability to achieve the required destruction
and removal efficiency (ORE) of
99.99 percent*. Some analysis of bottom
ash, flyash, and scrubber discharge liquid
was conducted. However, in order to assist
OSWER in establishing a standard for
residue quality, this effort was undertaken
to conduct more comprehensive
characterization of incinerator bottom and
flyash at a greater number of hazardous
waste incineration facilities. In addition
to meeting a residue quality criterion to
be established by OSWER, facilities that
treat, store, or dispose of hazardous waste
(TSDs) will be subject to existing
pretreatment discharge standards
established by the Office of Water (OW) or
such standards as may be developed by OW in
Trenholm, A.» P. Gorman, and 6. Jungclaus. Performance Evaluation of full-Scale
Hazardous Waste Incinerators. Midwest Research institute, KansasCity, Missouri.
421
-------
the future. Therefore, there also exists
a need for comprehensive data on the
characteristics of any wastewater that may
be discharged from a hazardous waste
Incineration facility.
APPROACH
Candidate incinerator test sites were
Identified and screened based on site
availability, operational status, and
types of wastes being incinerated. During
the site selection process, emphasis was
placed on facilities that incinerate solid
waste, generate ash, use air pollution
control devices, and facilities which were
previously tested for air emissions and
thermal destruction performance. A total
of 10 sites were tested in this program
comprising a broad range of hazardous
waste incinerator design and current
operating practice.
The selected sites span a broad range
of incinerators, 6 with rotary kilns,
3 with fixed hearth, and 1 with a
fluidized bed incinerator. All sites with
rotary kilns also burned liquid wastes
downstream of the rotary combustor. Air
pollution control equipment ranged from
uncontrolled to primarily wet controls.
Excluding the two sites with no control
devices, all sites had a quench system and
a scrubber. A couple of sites with low
pressure wet scrubbers also employed wet
electrostatic precipitators. Table 1
summarizes the incinerator configurations
encountered.
The wastes fired during the sampling
period were selected by the host site and
were generally felt to be representative
of material normally incinerated. In a
couple of cases, solid hazardous wastes
were selected to provide a more uniform
feed to promote the gathering of a more
representative sample. The wastes were
not spiked as is frequently performed in
source testing operations. Table 2
summarizes the sampled input and output
streams and types of analyses performed.
The typical test involved sampling
nongaseous incinerator inlet and outlet
streams during a 2 to 4 hour period of
operation. The samples were collected and
analyzed in accordance with EPA/OSW "Test
Methods for Evaluating Solid
Waste-Physical/Chemical Methods," SW-846.
The analysis protocol for the collected
samples is listed in Table 3. At sites 2,
7, and 8, not all streams were sampled due
to safety and/or proprietary concerns.
Wastes not sampled included lab packs,
hospital wastes, nitriles, magnesium scrap,
and, at one site, all drummed wastes.
Concurrent with sampling, system operating
information was also obtained to
substantiate normal operation. Two
extraction procedure toxicity test methods
were used, namely Method 1310 in SW-846,
the EP Toxicity Test Procedure, and a draft
Toxicity Characteristic Leaching Procedure
(TCLP) using the EPA draft protocol EP III,
revised December 1985. Extracts from the
former were analyzed for priority pollutant
metals. Extracts from the latter were
analyzed for priority pollutant metals and
semi volatile organics, and for volatile
organics using a zero-head extraction
vessel (ZHEV).
Analyses are not presented in this
paper for Site 10 since no ash residual
stream was generated at that site during
the sampling activity. Also, EP III
volatile organics are not reported here,
since they were not scheduled for analysis
until after the preparation of this paper.
RESULTS
Analytical results are presented in
Tables 4 and 5 for all ash residual
samples. Site 5 Large Incinerator Bottom
Ash and Site 9 Kiln Ash are not included in
Table 4 since volatile organics were not
detected in the ash at nominal detection
levels of 100 and 0.5 mg/kg, respectively,
and semi volatile organics were not detected
in the ash or EP III leachate at nominal
detection levels of 0.1 mg/kg for ash and
2 yg/L for leachate, respectively.
VOLATILES AND SEMIVOLATILE ORGANICS
A total of 13 distinct volatile
organics and 24 distinct semivolatile
organics were detected in the ash residual
samples. Even the low volatiles
concentrations in the ash noted would
generally not be expected. However, these
levels might be due to the ash adsorbing
volatiles from quench water (Sites 1, 2, 3,
7, 8, and 9), flue gas, or air; products of
incomplete combustion (PICs) (especially
possible with Site 4) or early ash
quenching before complete ash burnout
(possible with Sites 3 and 8); or sample
contamination. Except for Site 4 where the
feed material was a relatively pure
422
-------
TABLE 1. HAZARDOUS WASTE INCINERATOR CONFIGURATIONS AND WASTE IDs
Site No.
Incinerator type
EPA Haste Identtf (cation No.
Incinerator ash quench
1
Rotary kiln
with secondary
combustor in
parallel with
a liquid
waste-fired
boiler
D001
F001
FQQ2
F003
F005
X
2
Rotary kiln
with secondary
coitibustor in
parallel with
a liquid
injection
combustor
D001
0008
X
3
Rotary
kiln with
secondary
corabustor
D001
F001
F002
F003
FQOS
X
4 5
Flu1d1zed Fixed hearth
bed (2 separate
Incinerator incineration
systems)
None D001
FQQ1
F002
F003
F005
6 • 7
F1xe,d Fixed
hearth hearth
with
secondary
combustor
D001 D001
F003 F001
F005 F002
F003
F005
X
8
Rotary kiln
with
(secondary)
liquid
injection
conbustor.
Drums also
conveyed
through
combustor
D001 F001
D002 FOOZ
D006 F003
D007 FOQS
D008 U002
D009
X
9
Rotary
kiln with
secondary
combustor
D001
F001
F002
F003
F005
X
(rotary kiln only}
10
Rotary
kiln with
secondary
combustor
D001
F001
F002
F003
F005
X
(But no ash
during
testing)
Secondary combustion
chamber with liquid
waste injection
Hot gas cyclones
Quench
Scrubber + demister
Acid absorbers
Waste heat recovery boiler
Wet ESP's
No control device
(Constraints on fuel and
firing rates)
Selective material reburning
X
X
X
X
(liquid-waste fired)
X X
(drums and residue)
-------
TABLE 2. SUMMARY OF SAMPLES COLLECTED AND ANALYSES PERFORMED FOR
10 HAZARDOUS HASTE INCINERATION FACILITIES
Analyses
Stream description
Site numbers
Priority EP Draft
pollutant toxicity EP III PCB
Volatiles Semivo1at11es metals procedure procedure identity3
Input Streams
APCD supplyb
Aqueous or low-Btu waste
Coating waste solids
Chloroprene catalyst sludge
CS tear gas powder
Oichlorobutene synthesis coke
Drum feed liquids
Drum feed solids
Lacquer chips
Lacquered cardboard waste
Latex coagulum solids
Liquid injected waste fuels
PCB-contaminated soil
PCB liquid waste
30 mesh flintshot sand
Unused automotive paint
Vacuum filter solids
Output Streams
APCD discharge0
Boiler tube soot blowdown
Cooling pond sludge
Cyclone ash
Incinerator bottom ash
Waste water treatment facility
belt filter cake residue
Waste water treatment facility
discharge water
Rotary kiln ash
Stack condensate
8
1 and b
7
2
4
I
3
3 and 9
6
5
7
1, 3, 5 to 10
1
1
4
2
2
1 to 4, 7 to 10
3
8
1 and 4
5 to 8
3 and 7
7
1 to 3, 8, 9
4
X
•x
X
X
X
X
X
X
X
X
X
X
X
X -
X
X
*
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
•x
X X
XXX
XXX
XXX
X X X
1
X X X X
aSite 1 only.
''Site 8 uses cooling ponds and recycles APCD discharge.
-------
TABLE 3. ANALYSIS PROTOCOL — TYPICAL INCINERATOR SYSTEM
Stream
Category
Waste
feeds
Sample
Description
o Solids and
sludges
» Organic
liquids
o Aqueous
liquids
Pollutant
Category
o Volatile POHC's
e Semi volatile
POHC's
o Priority
pollutant metals
• Volatile POHC's
o Semi volatile
POHC's
o Priority
pollutant metals
e Volatile POHC's
o Semi volatile
POHC's
Analysis Method3
o Purge and trap GC/MS by Methods 5030 and 8240
o Sonication extraction, concentration, direct
injection GC/MS by Methods 3550 and 8270
o Acid digestion by Method 3010, 3020, or 3050,
AA by 7000 series methods
o Direct injection (dilution as necessary) GC/MS
by Method 8240
o Direct injection (dilution as necessary) SC/MS
by Method 8270
« Add digestion by Method 3030, AA by 7000
series methods
o Purge and trap by Methods 5030 and 8240
» Continuous liquid-liquid extraction,
concentration, direct injection GC/MS by
Methods 3520 and 8270
Discharge o Solids (ash
streams samples)
o Aqueous
liquid
(scrubber
discharge)
« Priority
pollutant metals
o Volatile POHC's
• Volatile POHC's
9 Semi volatile
POHC's
o Priority
pollutant metals
o (Same" as aqueous
liquid waste
feeds')
AA by 7000 series methods
o Purge and trap GC/MS by Methods 5030 and 8240
a EP III leaching with analysis of the
resultant extract by purge and trap GC/MS by
Methods 5030 and 8240
* Sonication extraction, concentration, direct
injection GC/MS by Methods 3550 and 8270
o EP III leaching with analysis of the
resultant extract by continuous liquid-liquid
extraction, concentration, direct injection
GC/MS by Methods 3520 and 8270
o Acid digestion by Method 3010 or 3020,
AA by 7000 series methods
o EP I toxicity leach test with analysis of the
resultant extract by AA via 7000 series
methods
o EP III leaching with analysis of the
resultant extract by AA via 7000 series
methods
o (Same as aqueous liquid waste feeds)
aA11 method numbers refer to SM-846, second edition
425
-------
TABLE 4. CONCENTRATION OF VOLATILE AND SEHIVOLAT1LE ORGANICS IN INCINERATOR ASH RESIDUALS AND THEIR EP III LEACHATES
iitc HtM&er 1
Hfe« ijyscriptiofl Jttln «h
2
iciia «n
3
Mln «SH
3
toller tth
4
Cyclone ash
5
button dSh
DwU
bs
Incinerator
bottom ash
8 8
Klin Kh Inclntritor
botlo* 45h
C^un^-M
{«9/KyJ/(v*g/L]
volatile Organic;"
Nontnal Detection Limit 1 / — •
SeraivolatUe Organics1*
Nominal Detection Limit 0.1 / 2
r luorantnenK /
* *
Phunal ™- / —
Benzyi outyi pril aiaie /
ui n Duty i p t aiate /
Uietttyl phthalate — / —
/
2 Kethyt naphtha Ism. /
Coweiits: Het or Dry Asli Ifet
i («,/*a)/(«/L)
100 / —
'.
0.1 / 2
0.35 / -
0.31 / 24
',
*
.. i „
Hut
KtMttrdtiun.
jimiaa! detect iott
{•g/fcg}/(pg/L]
0,5 / «
?*»/.,
i «i / ..
0.1 / 2.
01? / —
3 / U6
4,4 / .,
-- / 6
*
*
wet
Itmtt.
1 t««/k9)/{pg/L)
0.1 / --
/ —
'.
i __
0.1 / 2
0.4 / 3U
*,
*
t
wet
W»9)/(MA)
0.5 / --
3 7 / —
5 4 / —
S 4 / .»
/ „
^
^
0.01 / 2
/ ,_
-- / --
(
'
/
/ .„
',
Pry
(^1J/(WA)
100 / —
%
y
^
^
0.1 / 2
^
-- / --
t
*,
',
II ry
(•9A9
10U
0.1
0 26
10
0 23
1.1
6.8
1.7
500
5
39
2.5
31
1 3
2.4
6.2
Dry
)/(M/L)
/ -
/ ,-
/ *„
^
/ ._
/ 2
/ 10
/ 20
/ 8
/ 60
/ 90
/ 30
/ 14
/ --
/ 580
-------
TABLE 5. CONCENTRATION OF PRIORITY POLLUTANT METALS IN INCINERATOR RESIDUALS
Site number
Stream description
1
Kiln ash
2
Kiln ash
3
Kiln ash
3
Boiler ash
4
Cyclone ash
5
Urge incinerator
bottom ash
Concentration3
Antimony
Arseni c
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Comments
Wet or Dry Ash
(rag/kg)/!mg/L)/{ing/L)
2 /<0.05 / 0.04
4 / 0.23 /<0.01
<2 /<0.01 /<0.01
120 / 0.10 / 0.22
6900 / 8.6 / 16
220 / 2.3 / 3.5
<0.05 /<0.001/<0.001
190 / 0.49 / 0.45
<1 /<0.05 / 0.02
11 /<0.01 /<0*01
<1 /<0.01 /<0.02
160 / 0.14 / 0.42
Wet
(mg/kg)/(aig/L)/(mg/L)
6
2
<2
<1
110
840
100
1.5
7300
6
8
<1
640
Met
/<0.01 /<0.01
/<0.01 /<0.01
/<0.01 /<0.01
/ 0.09 / 0.10
/ 3.7 / 7.9
/<0.01 /<0.01
/<0.001/<0.001
/ 6.9 / 6
/ 0.2 / 0.05
/ 0.05 /<0.01
/<0.01 /<0.02
/ 1.8 / 2
(mg/kg)/(mg/L)/(nig/L)
18 / 0.06 /<0.01
3 /<0.01 /<0.01
<7 /<0.01 /<0.0l
<1 /<0.01 /<0.01
660 / 0.03 / 0.06
400 / 0.02 / 0.09
610 / 0.04 /<0.01
<0,1 /<0.001/<0.001
240 / 0.79 /13
13 / 0.17 / 1.4
4 / 0.02 / O.OS
7 /<0.01 /<0.02
21000 / 27 /300
Met
(mg/kg)/{mg/L)/(mg/L)
190
14
6
61
1800
780
5000
0.2
4700
13
190
9
32000
Wet
/<0.01 /<0.01
/<0,01 /<0.01
/<0.01 / 0.08
/ 8.6 / 6.7
/ 0.03 / 0.36
/ 31 /21
/ 4.4 / 4.5
/<0.001/<0.001
/ 20 /13
/ <0.1 / 1,4
/ 0.09 / 0.05
/ 0.7" /<0.02
/ 1400 /1200
(mg/kg)/(mg/L)/(mg/L)
<2
<1
7
<4
<1
<0.1
25
<1
120
<1
200
Dry
/<0.01 /<0
/
-------
chemical, o-chlorobenzalmalononitrile, the
volatile orgardes found also appear in the
waste feed. The cyclone ash from Site 4
shows several compounds that would appear
to be PICs. Because the cyclone ash was
collected In the cyclone, periodically
emptied, and allowed to free fall through
air during the cyclone draining procedure,
it was likely that the volatiles observed
were adsorbed while the ash was in the
cyclone and/or during the free fall
through air upon draining.
Several of the seraivolatiles detected
are phthalates and frequently associated
with laboratory contamination, especially
when observed at low values. Site 6 may
be an exception since one input, lacquer
chips, contains approximately 7 percent by
weight Bis(2-ethylhexyl)phthalate and was
detected in the ash at a concentration of
500 mg/kg.
Most detected compounds are less than
10 ppm. Most sites quench ash with water.
Especially if a rotary kiln discharges ash
and unburnt material too quickly, it is
possible for some of the organics to not
be subjected to high enough temperatures
for complete combustion (thus, the
appearance of the organics in the
analyses). Also, the quench water may
experience a buildup of these organic
compounds and contaminate the ash (c.f.,
wet and dry ash from Site 8).
PRIORITY POLLUTANT METALS
The metal concentrations presented in
Table 5 indicate that only 1 metals
measurement of the EP leachate out of 84
exceeded the maximum concentration of
contaminants for characteristics of EP
Toxicity (standards are set forth in
Table 4 of 40 CFR 261.24), hence only the
boiler ash at Site 3, due to cadmium being
8.6 mg/L versus an allowable standard of
1 mg/L, would be considered a hazardous
waste for metals if not already listed in
40 CFR Subpart D. The EP III leachate,
1f subjected to the same standards, would
have 3 measurements out of 84 exceeding an
allowable concentration. Site 3 boiler
ash would exceed the standards for cadmium
at 6.7 mg/L and selenium at 1.4 mg/L
versus an allowable standard of 1 mg/L.
Site 6 ash would exceed the standard for
lead at 12 mg/L versus an allowable
5 mg/L. In general, the results from the
two different extraction procedures were
within a factor of three.
Leachate concentrations (in mg/L) are
expected to be about 20 times less than ash
reported values (in mg/kg) for 100 percent
soluble metals. Although several metal in
ash concentrations are less than detectable
limits and cannot be further evaluated,
solubility generally ranged from 1 to
10 percent. Metal concentrations greater
than 1000 mg/kg of ash included chromium
(Site 3), copper (Sites 1, 7, and 8) lead
(Sites 3, 5, 6, and 8), nickel (Sites 2 and
3), and zinc (Sites 3 and 8).
These high concentrations in the ash
did not always yield a good mass balance.
Outputs were greater than inputs by a
factor of ten for chromium (Site 3), copper
(Site 1), and lead (Site 6) and by a factor
of 100 for copper (Site 7). Since process
data were not gathered for Site 5 and all
streams were not sampled for Sites 7 and 8,
mass balance statements cannot be
accurately made for those sites. To
improve the representativeness of the ash
samples and better close a mass balance
would require sampling and analysis of
other streams.
Most of the leachate measurements for
antimony, arsenic, beryllium, cadmium,
lead, selenium, silver, and thallium
yielded measurements less than detectable
limits of nominally 0.01 to 0.05 mg/L of
leachate. All mercury leachate
measurements were less than 0.001 mg/L of
leachate.
ACKNOWLEDGEMENTS
This project was jointly sponsored by
the U.S. Environmental Protection Agency's
Hazardous Waste Engineering Research
Laboratory (HWERL). The valuable
assistance and guidance of Paul Warner and
Robert Olexsey in HWERL are acknowledged.
The cooperation of all test sites and
the efforts of man/ individuals in the
Environmental Engineering Department and
Analytical Chemistry Laboratory of the
Energy & Environmental Division of Acurex
Corporation are also acknowledged.
428
-------
PRACTICAL LIMITATION OF WASTE CHARACTERISTICS
FOR EFFECTIVE INCINERATION
Edward J. Martin, Leon W. Weinberger
PEER Consultants, Inc.
Rockvllle, Maryland 20852
Joseph T. Swartzbaugh
PEER Consultants, Inc.
Dayton, Ohio 45432
Alexander P. Mathews
Kansas State University
Manhattan, Kansas 66502
C. C, Lee
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
This report gives Information which can be used to assess the 1nc1nerab111ty of
wastes which are subject to being banned from'landfill disposal under the 1984
Amendments to the Resource Conservation and Recovery Act. Existing Incinerator
technology 1s evaluated and the characteristics of wastes which Identify their 1nc1n-
erabiHty are identified. Publicly accessible waste Information data bases were
reviewed and found to be deficient 1n requisite waste characterization. Commercial
incinerator operators were Interviewed to assess the costs of Incineration and the
characteristics of wastes which Impact those costs.
Based upon published data from trial burns, a computerized thermodynamic model
of incinerator systems with different air pollution control devices was developed
which predicts the emission pathways taken by various metallic species. An extensive
analysis of the potential risks to human health and the environment resultant from
organic and metallic emissions were calculated.
Finally 1t was determined that incineration should be considered as the
preferred method of treating/disposing of organics-laden hazardous waste and a
general approach to assessing a waste's incinerability was developed. A similar
approach was developed for determining the need for and the effects of pretreatment
of non-lncirterable'wastes to make them acceptable for Incineration. Both of these
are presented as formal logical processes which together could form the basis for an
approach to identifying waste incinerability.
INTRODUCTION
The 1984 Amendments to the Resource mental Protection Agency take steps to
Conservation and Recovery Act (RCRA) restrict many hazardous wastes from land
require that the United States Environ- disposal and/or establish criteria for
429
-------
treatment of wastes. The amendments
require that all RCRA listed characteris-
tic wastes be passed through an analysis
to assess the feasibility of banning
them from further land disposal.
Certain wastes will be banned on a given
schedule, but In other cases the EPA is
required to either ban or show why these
are not detrimental to human health and
the environment when being disposed to
thfi land. High-temperature Incineration
is an established and recognized
technique for destroying many, predomin-
antly organic, waste streams. As such,
incineration is the technique which will
be considered first as an alternative to
land disposal for many organic streams,
which may be banned but which are not
typically incinerated in current
Industrial/commerical practice.
The basic objective of the subject
work assignment was to determine
criteria for Incinerability of all waste
groups which are required to be banned
as a first priority, and other wastes
which are likely to be banned. The
final product 1s a set of criteria which
can be used by the generator, Superfund
site manager, or facility owner/operator
for use 1n evaluating incineration,
ASSESSMENT OF INCINERABILITY
Although waste characterization
data are very limited, there exists
sufficient information about Incinerator
capabilities and limits to Indicate
that, even though no one type of Indner-
tor can readily accept every waste, for
any waste there is at least one type of
incinerator capable of burning that
waste. Thus, a major conclusion of this
study is that for hazardous wastes
containing some organic fraction,
incineration should be considered the
preferred method of treatment/disposal
unless one or more of the following
conditions hold:
A, The waste cannot be physically
introduced into an incinerator,
even after pretreatment.
0. Constituents are present 1n the
waste which would destroy the
incinerator or result In its
rapid deterioration.
C, No site and/or disposal method
1s practicably available for
the environmentally sound
disposal of ashes and other
residues.
D. For wastes having a heat
content too low to sustain
combustion, no supplementary
fuel Is available.
E. New and overriding environmen-
tal regulations are put forth
which cannot be dealt with,
even using new control
technology (e.g., NOX, SOX)
F. Agency or Congressionally man-
dated "acceptable risk" levels
'are too low to be met by
Incineration systems.
In addition to this major conclusion,
the following paragraphs summarize the
other conclusions of the study.
The authors surveyed representatives
of the Incineration service
Industry, Incinerator manufacturing
industry and the waste-as-fuel
brokering Industry. The information
gathered did confirm a general
Industry consensus that capacity at
present is sufficient and expansion
capability exists.
Quality characteristic data (full-
composition analyses) for hazardous
wastes are practically non-existent.
This lack of quality data signifi-
cantly hampers the performance of
such tasks as the analysis of metal
behavior 1n the process, the
assessment of the capability of
APCDs to remove metals, and the
estimation of risks resultant from
emissions.
The determination of 1nc1nerab1lity
for given wastes should be done on a
case-by-case basis by matching the
waste to the available Incinerators
based on the specific limitations of
each Incinerator. The Input
capabilities of an incinerator
depend on the specific design of
that Incinerator's feed mechanism
and, to a limited extent, on the
incinerator ash-handling system.
430
-------
Extensive data exist which describe
the destruction and removal
efficiencies achieved for various
wastes in several types of
combustion systems. These data are
sufficient to make realistic
estimates of the risk to human
health.
Based on available PIC and POHC
data, the risk to human health
from Incineration does not appear to
be significant. The quantity of
metals in the Input waste streams
impacts the risk levels to exposed
individuals and populations. Under
the assumptions and conditions used
in this study these levels exceed
the one-1n-a-mill1on lifetime cancer
risk for carcinogens and the
acceptable dally Intake (ADI)
levels, for high metals content
waste streams and for large
throughputs. .The potential risks
due to metals are higher 1n general
than those resulting from exposure
to PICs,
It is possible to model the metal
partitioning aspects of hazardous
wastes combustion between the bottom
and fly ash fractions, and to
estimate the removability with
various APCDs.
An improved basis for estimating metals
partitioning would enhance the
understanding of metals behavior In
combustors, removability of the various
fractions in bottom ash and APCD
residues, and the risks to exposed
individuals. This improved
understanding would 1n turn, allow
permit writers to ensure better control
of metals in permitted Installations
through the permit conditions.
The waste characteristic data
necessary to develop an effective
detailed pretreatment strategy were
not available. A general 'strategy
was developed and Is presented
herein as a recommendation.
The cost of incineration 1s directly
impacted by the factors Btu content
(Inversely), water content and solids
content. The maximum unit cost for
incineration of hazardous waste
which was confirmed during the survey
was $30Q/ton. The maximum unconfirm-
ed cost was $600/ton. It 1s likely
that generators are paying even
higher rates than these. These
rates are considerably higher than
previously thought for Incineration,
and leaves a considerable window for
application of pretreatment 1n a
trade-off analysis. The cost of
disposal of residues could become
the most significant aspect of the
cost of Incineration and especially
of any necessary pretreatment.
SUGGESTED STRATEGY FOR .DETERMINING
THE BURNABILITY OF SPECIFIC •
WASTE STREAMS
The conceptual strategy for
determining the wastes which should be
burned 1s shown schematically 1n Figure 1,
The major decision points for this
assessment are;
I. The first postulated requirement 1s
that all dioxin and furan-containing
wastes must be Incinerated.
2. All Wastes containing PCBs will also
be required to be Incinerated.
3. Liquid and solid wastes containing
more than 1000 ppm of halogenated
organ1cs can be Incinerated. For
halogenated wastes 1t will be
necessary to evaluate alternatives
to Incineration. However, 1n
assessing such alternatives, data
and Information could be required to
be submitted for permit applications
which demonstrate to the Agency that
equal destruction 1s achievable by
the application of chemical,
physical or biological treatment.
4. If the waste being considered 1s
hazardous by characteristic 1t will
be necessary to evaluate the
relative hazard associated with "non-
burning" disposal or treatment
options and compare those
levels of hazard with the probable
hazards to human health and
environment resulting from
Incineration.
5. The four above criteria would
deal with the large fraction of
431
-------
Incinerate
Does waste contain
Dioxins or Furans?
Does waste
Contain >50 ppm
PCBs?
Incinerate
Identify all state-
of-the-art disposal
and treatment options
Does waste contain
>1000 ppra Halogens?
Assess relative
hazard to HHE for
each option compared
to incineration
Is the
waste a "characteristic"
waste?
Is heat
content (as received)
>8500 Btu/lb?
Estimate costs
of each option
having acceptable
"hazard" level
Incinerate or
use as fuel
/use lowest cost
y^ "acceptable
^•v^
Is heat
content (as received)
between 2500 and
8500 Btu/lb?
Incinerate with
auxiliary fuel
Figure 1. Conceptual strategy for determining waste burnability,
432
-------
available organic waste groups. For
the remaining wastes, I.e., the
"survivors" of the evaluation, the
considerations given below would
apply. The remaining two decision
steps are based on the premise that
all wastes which are organic 1n
nature (including those which are
hazardous for reasons other than
toxicity) should
be incinerated.
The next step in the evaluation approach
is based on heat content of wastes. The
heat content is the primary determinant
of burnability and this takes the water
content into account indirectly. Ash
content is more a determinant of the
type of incinerator to be used rather
than the burnability.
Above approximately 8500 Btu/lb» the
wastes can be considered a fuel, and
under the regulations for wastes used as
fuels, would be required to be burned
with a RCRA permit under Subpart 0,
Hastes at this level of heat content and
above would sustain combustion 1n most
furnaces.
Between 2500 and 8500 Btu/lb, the waste
can be considered a "prime candidate for
incineration" although requiring the use
of auxiliary fuel in many Instances.
Again, it would be necessary to evaluate
the relative hazards of burning compared
to other disposal or treatment options.
Below 2500 Btu/lb, the wastes may
require pretreatment before incineration
to concentrate the wastes and this could
involve application of chemical,
physical or biological treatment.
A GENERAL CONCEPT FOR A
PRETREATMENT STRATEGY
The generalized strategy presented
is based upon a series of questions
which must be asked about any candidate
waste. That which constitutes acceptable
answers to these questions Is dependent
upon the limits imposed by each specific
incinerator system. Thus, the
conceptual pretreatment strategy, as
presented, is based on the assumption
that the assessment is being made for a
specific existing incinerator system.
Further, it is assumed that the analysis
of the candidate waste stream is known.
The waste parameters which were
identified as being important to
1nc1nerab1lity and for which costs
Increments are charged by commercial
incinerator owners/operators are water
content, solids or ash content and heat
content. These then provide the basic
questions and are the focus for
determining a pretreatment requirement,
I.e., must one lower solids and/or water
content of a waste and must one Increase
the heat content of a waste to enhance
its incinerability. If some regulatory
requirement for pretreatment is to be
utilized, then the process for
determination of 1nc1nerab1lity for
given Incinerator/waste combinations
would contain the pretreatment
feasibility determination as an Integral
part. The pretreatment decision process
is shown in Figure 2. If required, this
decision process would have to be added
to the overall Incinerability decision
process shown 1n Figure 1.
It should be noted that there are
essentially two phases to the
pretreatment feasibility assessment.
The first phase 1s composed of a series
of questions which must be asked to
determine 1f pretreatment 1s required
for acceptance Into an Incinerator
system and the second phase is a set of
risk assessment.steps that must be
performed to determine the composite
resultant risk to human health and the
environment both from the combustion of
the waste 1n that Incinerator system and
from the ultimate disposal of all
residues of the pretreatment process and
of the incineration process.
The first consideration as to
whether a waste 1s indnerable Is the
determination of whether or not the
gross physical form of the waste 1s one
which can be fed Into the incinerator.
If it is not, then that question Is
Immediately followed by the question of
whether or not .some pretreatment could
be performed which would modify the
physical form of the waste so that it
could be introduced Into the
incinerator. If again the answer is no
then another alternative disposal method
for this waste is required.
433
-------
Oewatering
Process
Gross Physical
Form O.K.
Moisture Conten
Too High
Is
Metals Content
Too High
IS
Btu Content
Too Low
Can
Physical Form
be made O.K.
Modify Physical
Form as Needed
S ^NX*^ v
Content^ .s h
Highx-^
I* . . .... .
Solids Removal
Process
Metal
Removal/Reduction
Process
Augment with
High Btu Waste
or Auxiliary Fuel
utner
Alternatives
Figure 2. Pretreatment option logical decision flow chart.
434
-------
Estimate Probable Emissions
- Bottom Ash
- Fly Ash
- Scrubber Haters
- ParticuJate Emissions
Gaseous Emissions
Assemble List of all
Pretreatment Residues
- Liquids
- SolIda
- Sludge* •
Identify BAT lor
Treatment/Disposal
of all Residues
(Incinerator and Pretreatment)
Perform Composite Risk
Assessment for all .
Emissions and all.
Disposal Residues
Assess Total Process
Cost for Incineration
with Pretreatment
Compare Cost "Effectiveness*
with other Incineration/Pretreatment
Combinations or other Competing
Technologies
Use Host Cost Effective Option
Figure 2. (Continued)
435
-------
A second Important question to be
asked about the waste is the question of
whether or not its moisture content 1s
too high for the specified Incinerator
system. If so, then some form of
dewatering process 1s required.
The next question to be asked 1s
Hhether the sol Ids content of the waste
Is too high, 1f so then the waste must
be subjected to a solIds removal or
solids reduction process.
The fourth question which must be
asked is whether or not the metal
content of the waste (as treated to this
point) Is too high. This question 1s
one for which the limits of acceptable
answers will probably be set by the
expected metal emission rate for the
candidate Incinerator when burning metal
bearing wastes and the resultant risk
which would thus obtain. If the metal
content is too high then the waste must
be subjected to some metal reduction or
metal removal process.
The final question about the nature
of the waste 1s whether or not the
waste's heat content Is too low. If the
heat content 1s too low then the waste
to be burned In the candidate
incinerator must have Its heat content
augmented somehow 1n order to ensure
proper combustion at the appropriate
temperatures within the Incinerator
system.
Now 1t Is necessary to enter Into
the second phase of the evaluation
process. The following steps in the
logic flow chart are those which are
required to assess the risks resulting
from all aspects of pretreatment and
combustion of the candidate'waste.
First it 1s necessary to estimate the
probable emissions from the Incinerator
system. This Includes Incinerator
bottom ash, fly ash, any scrubber
waters, and the partlculate emissions
and gaseous emissions. Once one knows
Che extent and probable character of the
various emission pathways, then one 1s
in a position to estimate the risks
resultant to human health and the
environment from those emissions as a
result of burning the subject waste 1n
the candidate incinerator. The next
step Is to assemble a 11st of all
pretreatment residues resultant from the
various steps in the pretreatment
process. These residues are likely to
Include solids, liquids or sludges. For
• each of these residues, the best
available technology for treatment and
disposal must be Identified. This
actually 1s required for all residues
both from the Incinerator and from the
pretreatment processes. Once all of the
Incinerator emissions and the resultant
emissions from all treatment processes
applied to the pretreatment residues
have been Identified, then a composite
risk assessment can be performed for all
of the emissions and all of the disposed
residues.
Finally, one should assess the total
process cost for Incineration Including
the cost of pretreatment and including
the cost of treating and disposing of
the Incinerator and pretreatment '
residues. Then compare the cost
effectiveness of the entire approach
with other candidate Incinerator/
pretreatment combinations or any other
competing technologies. The final step
in the decision process 1s to select and
use the most cost-effective option
available for the candidate waste.
436
-------
INCINERATOR AND CEMENT KILN CAPACITY
FOR HAZARDOUS WASTE TREATMENT
Gregory A. Vogel
Alan S. Goldfarb
Robert E. Zler
Andrew Jewell
The METRE Corporation
McLean, ¥irginla 22102
Mr. Ivars Lids
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
Estimates of incinerator and cement kiln capacities for hazardous waste
treatment are required to evaluate the Impacts of banning land disposal of hazardous
wastes. RCRA Part B permit applications were reviewed to obtain information about
incinerator design capacity, utilization and the Incinerated hazardous wastes. MITRE
identified 208 incinerators within the RCRA regulatory program that are presently
destroying approximately two million metric tons of hazardous waste annually. The
unused potential capacity of these units is estimated to be one million metric tons
of waste per year. The Congressional Budget Office' estimates that 265.3 million
metric tons of hazardous waste are generated annually.
MITRE estimates that the annual hazardous waste treatment capacity available in
cement kilns ranges between two and six million metric tons. Less than five percent
of the potential hazardous waste treatment capacity in cement kilns has been permit-
ted under RCRA. Factors affecting this low utilization include the large geographic
distances separating some major waste generation sites from cement kilns, marginal
economic benefits, and the uncertainty of some kiln operators about regulatory
requirements.
INTRODUCTION
The U.S. Environmental Protection
Agency (EPA) has been authorized to ban
the land disposal of some hazardous
wastes under the 1984 Hazardous and
Solid Waste Amendments to the Resource
Conservation and Recovery Act (RCRA).
If insufficient capacity exists to dis-
pose of banned wastes using alternative
treatment technologies, EPA is autho-
rized to delay the effective date of
such a ban. Incineration and thermal
destruction of specific wastes in cement
kilns and incinerators are preferred to
land disposal. EPA requested that METRE
prepare an estimate of incinerator and
kiln capacity to determine whether any
delays are warranted.
The purpose of this study is to
estimate the potential hazardous waste
437
-------
destruction capacities of incinerators
and cement kilns beyond current utiliza-
tion. Between 1980 and 1982 The MITRE
Corporation conducted several studies
of domestic hazardous waste Incinerator
manufacturers, owners and operators.
The information from these studies and
new data obtained from RCRA Part B
permit applications in 1985 were used
to estimate incinerator capacity.
All facilities operating hazardous
waste incinerators after 19 November
1980 were required to file RCRA Part A
permit applications by EPA. The haz-
ardous waste incineration facilities
Included In this study have filed RCRA
Part A permit applications and the Part
A information has been verified.
EPA has requested RCRA Part B
permit applications for Incinerators
and the applications are presently being
received by the EPA regional offices.
The Part B information requirements are
much more detailed than the Part A
applications. The Part B applications
and trial burn results are reviewed at
the EPA regional offices or by offices
within states authorized by EPA. Per-
mits to incinerate hazardous wastes are
issued or denied based on evaluations
of the Part B applications. MITRE
visited four EPA regional offices to
obtain Information from incinerator
Part B applications.
Seventeen (17) Part B applications
were reviewed at the Region 2 offices,
17 applications were reviewed at Region
3 offices, 30 applications were reviewed
at Region 4, and 38 applications were
reviewed at Region 5. The status of
incinerators at other regional offices
was determined through data verification
forms for the Incineration Permitting
Study conducted by A.T. Kearney, Inc.
of Alexandria, Virginia. These forms
were completed during November and
December 1985. In addition, offices In
California, Louisiana and Texas were
contacted to verify permit status. The
results of these reviews and a survey
of the hazardous waste incinerator man-
ufacturing industry are summarized in
this paper. An estimate of the hazard-
ous waste destruction capacity available
in cement kilns is also presented.
HAZARDOUS WASTE INCINERATOR CAPACITY
MITRE identified 341 incinerators
for which Part A applications had been
filed. Of these units, 34 have RCRA
operating permits, 174 have filed Part B
permit applications that are being eval-
uated, 99 have withdrawn from the RCRA
system and the status of 34 units could
not be determined. Permit applications
are withdrawn if incinerators cease
operation, no longer burn RCRA hazardous
wastes, or burn hazardous wastes that
have been delisted. Most incinerators
have ceased operation through voluntary
action, although a few have been closed
through regulatory enforcement. Most
of the incinerators for which the per-
mit status could not be determined are
located in Texas, where a large number
of permit applications are still being
classified. More than half of the
incinerators in the RCRA. regulatory
program are located in EPA Regions 5
and 6.
The Incinerator design capacities
were obtained for 87 percent of the 208
units that are permitted or have filed
an application. The design, or name-
plate, capacity of these units is 6.28
billion Btu/hour. Extrapolating this
statistic to include all 208 inciner-
ators in the RCRA regulatory program,
the national capacity could be 7.2
billion Btu/hr, which is equivalent to
burning approximately three (3) million
metric tons of hazardous waste per year.
The incinerator capacities and
other data are itemized by combustion
chamber design in Table 1. Rotary kilns
have the largest average capacity and
are most likely to have air pollution
control equipment. The relatively high
utilization of rotary kilns is expected
because of their high equipment cost.
Utilization of liquid injection incin-
erators is relatively low and less than
half are equipped with air pollution
control equipment. Many of these units
are operated intermittently as needed.
The average design capacity for fume
incinerators in Table 1 represents only
the liquid destruction capability}
Installed units have additional capacity
to burn fumes. The high utilization
results from integration of fume incin-
erators with continuously operating
438
-------
production processes. Hearth Inciner-
ators have the smallest average capacity
and the lowest incidence of air pollu-
tion control equipment installation.
The available capacity estimates
in Table 1 are derived from the average
design capacity and the utilization
statistics. The total available capac-
ity estimate of 2.53 billion Btu/hr is
roughly equivalent to one million metric
tons of waste a year or 0.4 percent of
the hazardous waste generated annually
in the United States. Half of the
incinerators are equipped with air
pollution control devices for burning
halogenated wastes. Most incinerator
air pollution control systems Include
scrubbers. Approximately 350,000 metric
tons of available capacity in rotary
kilns and hearths could be used to
incinerate solid hazardous wastes.
Information for 26 commercial
incinerators is included in the data
summaries for the 208 units in the RCRA
program. The design capacity of the
commercial incinerators that are per-
mitted or have filed applications totals
781,000 metric tons of waste annually.
However, 34 percent of this capacity
has not yet been constructed. The
utilization of commercial incinerators
is generally regarded as confidential
business information but is probably
not significantly different from the
utilization of private units. Nearly
all commercial incinerators have air
pollution control equipment.
Information about the charac-
teristics of incinerated wastes were
obtained for approximately 85 percent
of the 208 units known to be the RCRA
system. These facilities indicated that
1.72 million metric tons of hazardous
wastes are destroyed annually. An
annual volume of two million metric tons
for all 208 incinerators in the RCRA
program may be extrapolated from these
statistics. This estimate correlates
with the design capacity estimate of
three million metric tons and the
average utilization of 67 percent
presented in Table 1.
Based on the available informa-
tion, the waste incinerated in the
greatest amount is corrosive waste iden-
tified by EPA waste code D002, account-
ing for 29 percent of the weight of
wastes incinerated under the RCRA pro-
gram. Approximately 8 percent of the
wastes are ignltable (D001), 8 percent
are reactive (D003), 5 percent are spent
halogenated solvents (P001) and the re-
mainder of the wastes are t, U and other
F codes as identified in 40 CFR 261.
The average heating value of all
reported wastes is 8,580 Btu per pound.
Forty—six percent by weight of the
wastes are halogenated with an average
halogen content of 33.2 percent. The
average solids content of the reported
wastes is 7.9 percent and the average
water content is 50.5 percent.
HAZARDOUS WASTE INCINERATOR MANUFAC-
TURING INDUSTRY
During May and June 1985, 55
incinerator manufacturers were contacted
to provide information about units they
sold between 1981, the date of the
previous MITRE survey, and mid-1985.
Attempts to contact an additional 15
manufacturers were unsuccessful.
Of the 57 companies identified as
marketing hazardous waste incinerators
in the previous study, 23 have either
gone out of business, left the hazardous
waste incinerator business, or have put
much less emphasis on this activity.
Only one new company is pursuing this
market. Apparently many of the compa-
nies that were anticipating large growth
in 'the incinerator market in 1981 have
abandoned the business as a result of
selling only a few or no Incinerators
since that time. Of the 23 companies
marketing liquid injection incinerators
in 1981, only 12 are marketing them now;
of the 17 companies offering rotary kiln
incinerators in 1981, only 11 are doing
so now; and of the nine companies offer-
ing fluidized bed incinerators in 1981,
only five remain. Of the 16 -hearth
incinerator manufacturers in 1981, 12
remain. Half of the companies offering
innovative incineration technology in
1981 have left the market place. The
active incinerator manufacturers are
listed in Table 2.
439
-------
The 35 manufacturers cooperating
in this study reported that 111 incin-
erators had been sold since 1981. The
population and design capacity statis-
tics provided for the major types are
summarized in Table 3. Nearly all of
these incinerators are equipped with
air pollution control devices.
CEMENT KILN CAPACITY
The available hazardous waste
destruction capacity in cement kilns is
estimated from data including cement
production capacity, kiln fuel consump-
tion and the percentages of the thermal
input that can be provided by hazardous
wastes. MITRE estimates that the
present annual capacity for cement
production in the United States is 92.1
million tons. Government statistics
indicate that 71.3 million tons were
produced in 1983.
Cement is produced by wet and dry
processes, depending on whether the raw
materials are reduced in size using
water. The current trend favors the dry
process because less energy is required
than in the wet process where consid-
erable amounts of water must be evapo-
rated and heated. A breakdown of cement
production capacity by process type and
the associated energy consumption is
shown in Table 4.
Fuel requirements for cement kilns
range from 3 million Btu per ton of
product for dry kilns to 6 million Btu
per ton of product for wet kilns. Using
these statistics, the annual energy
requirement for cement kilns is esti-
mated to be approximately 400 trillion
Btu. Wastes may supply between 10 and
60 percent of the kiln heat input and a
typical value is approximately 30 per-
cent.
The heating values of wastes
burned in cement kilns range from 8,000
to 18,000 Btu per pound based on current
practice. The wastes with low heating
values are probably burned at low firing
rates to prevent kiln upsets. Wastes
•with high heating values similar to
fuels can replace large percentages of
fuel input. Waste destruction capaci-
ties in cement kilns estimated below
indicate a probable upper and lower
bound and a typical value:
Fuel Replacement Haste Heating Annual Cement Kiln
Rate Value Waste Capacity
(Percent)
10
30
60
(Btu/lb)
8,000
12,000
18,000
(Million of
Metric tons)
2.27
4.54
6.05
Most of the wastes reported to
have been burned in cement kilns are
either spent solvents, paint wastes or
still bottoms from solvent recovery
operations. These liquid wastes contain
metals such as titanium, lead, chromium,
manganese, zinc and barium. A limited
amount of metal oxides can be incorpo-
rated in cement without affecting the
quality of the product and particulate
emissions are controlled by existing
fabric filters, electrostatic preclpi-
tators or other devices.
Cement kiln operators typically
place limits on selected waste charac-
teristics to ensure a uniform high
quality product. A summary of the range
of acceptable waste characteristics is
presented in Table 5 for the 12 docu-
mented cases of waste incineration in
cement kilns available to MITRE. Other
important characteristics of acceptable
wastes include a sufficiently low
viscosity to permit atomizatlon, low
volatility, and being single-phase and
non-corrosive.
The quantity of wastes destroyed
in three permitted cement kilns was
obtained from the Economic Analysis
Branch, Office of Solid Waste, EPA. In
1983, the three kilns burned 21,741
metric tons of hazardous waste. The
Economic Analysis Branch has estimated
that 8 to 12 cement kilns have received
hazardous waste storage permits neces-
sary to burn hazardous wastes. Extra-
polating the known waste destruction
quantities for the three kilns provides
estimates of 58,000 metric tons
destroyed in 8 kilns and 87,000 metric
tons destroyed in 12 kilns.
These estimates of the quantities
of wastes currently destroyed in cement
kilns are one to four percent of the
440
-------
estimated potential cement kiln capac-
ity. The available capacity estimates
for wastes that could be destroyed in
cement kilns range from two (2) to six
(6) million metric tons per year, or
from 0.8 to 2.3 percent of the hazardous
wastes generated annually in the United
States. The major barrier to using this
capacity is the assurance that an eco-
nomic benefit exists from waste destruc-
tion in cement kilns.
tank construction, permitting, burner
modification, additional monitoring
equipment, operating and maintenance
costs, waste analyses and the cost of
the hazardous waste which ranges from
10 to 70 cents per pound. Economic
benefits include the reduction of fuel
costs and the receipt of disposal fees.
Based on conversations with cement
kiln operators, the profitability of
waste destruction in cement kilns is
marginal. Expenses include storage
TABLE 1
BSTIKATION OF AVAILABLE INCINERATOR CAPACITY
BY INCINERATOR DESIGN
INCINERATOR
DESIGN
Rotary Kiln
Liquid Injection
Fume
Hearth
Other
Total or
Average Values
NUMBER Of
UNITS
42
95
25
32
14
208
AVERAGE DESIGN
CAPACITY
(Million Bta/hr)
58.7
36.1
29.5
22.5
23.8
37.6*
UTILIZATION
(Percent:)
77
55
94
62
—
67"
AVAILABLE
CAPACITY
(Million Btu/hr)
570
1540
40
270
110
2530
-
PERCEHT WITH
AIR POLLUTION
CONTROL EQUIPMENT
90
42
40
38
__
50
*181 Incinerators reporting
90 incinerators reporting
441
-------
TABLE 2
1985 MANUFACTURERS BY INCINERATOR TYPES
Hearth Incinerators
Basic Environmental Engineering - Glen Ellyn, IL
Bayco - San Leandro, CA
Bum-Zol - Dover, NJ
Econo-Therm Energy Systems - Tulsa, OK
Ecolaire ECP - Charlotte, NC
Epcon Industrial Systems, Inc. - The Woodlands, fX
Midland-Rosa - Toledo, OH
Therm-Tech - Tualatin, OR
Hashbura and Granger - Patterson, NJ
Liquid Injection Incinerators
Brule' - Blue Island, IL
C&H Combustion - Troy, MI
CE Raymond - Chicago, IL
CJS Energy Resources, Inc. - Albertoon, NY
Coen - Burlingase, CA
Hirt Combustion - Montebello, CA
HPD, Inc. - Napierville, IL
Kelley Co., Inc. - Milwaukee, VI
McGill - Tulsa, OK
Peabody International - Stanford, Cf
Freneo - Madison Heights, HI
Shirco - Dallas, IX
Sur-Lite - Santa Fe Springs, CA
Irane Thermal - Conohohocken, FA
John Ziak - Tulaa, OK
Rotary Kiln Incinerators
CE Raymond - Chicago, IL
CSH Conbustion - Troy, MI
Environmental Elements — Baltimore, MD
Fuller Company - Bethlehem, FA
Induotronlca - South Windsor, CT
International Incinerators - Columbus, GA
Thermal1, Inc. — Peapack, NJ
Trofe Incineration - Mt. Laurel, NJ
Vulcan Iron Works - Hilkes Barre, PA
U.S. Smelting Furnace - Belleville, IL
Fluidlzed Bed Incinerators
CE Raymond - Chicago, IL
Copetech - Oak Brook, IL
Dorr Oliver - Stanford, CT
Fuller Company - Bethlehem, FA
GA Technologies - San Diego, CA
Sur-Lite - Santa Fe Springs, CA
Other Types of Incinerators
Midland-SosB-Rotary Hearth - Toledo, OH
Rockwell-Molten Salt - Canoga Park, CA
Shireo-Infrared - Dallas, IX
442
-------
TABLE 3
THERMAL EATINGS OF NEW HAZARDOUS WASTE
INCINERATORS REPORTED BY MANUFACTURERS
INCINERATOR
TYPE
Liquid Injection
Hearth
Rotary Kiln
RANGE OF
RATINGS*
(106 Btu/hr)
4 - 200
4-48
0.5 - 100
AVERAGE
RATING*
(106 Btu/hr)
56
20
44
OF
UNITS SOLD**
57
36
14
23 manufacturers reporting.
*35 manufacturers reporting.
TABLE 4
CEMENT KILN CAPACITIES BY PROCESS TYPE
PROCESS
Wet kiln
Dry kiln
Both Wet and Dry
ANNUAL CEMENT CAPACITY
(Thousands of tons)
26,783
39,384
kilns
at same location 17,172
Process Unknown
TOTALS
8.803
92,142
ESTIMATED .
ENERGY USE RATE
(Million Btu per
ton of cement)
6
3
4,5
5
ESTIMATED ANNUAL
ENERGY CONSUMPTION
(Trillion Btu)
160.70
118.15
77.27
44.02
400.14
443
-------
TABLE 5
RANGE OF ACCEPTABLE WASTE FEED
CHARACTERISTICS FOR CEMENT KILNS
Waste Parameter
Acceptable Range
Heating Value
Sulfur
Ash
Water
Chlorine
pH
Lead
Chromium
Zinc
Barium
Titanium
Mercury
Arsenic
8,000 Btu/lb to
1% to
5% to
1% to
3% to
4 to
less than
1,500 to
1,000 to
less than
less than
less than
less than
18,000 Bti
3%
12%
10%
10%
11
4,000 ppm
3,000 ppm
3,000 ppm
3,000 ppm
6,000 ppm
10 ppm
10 ppm
444
-------
NONSTEADY-STATE TESTING OF INDUSTRIAL BOILERS BURNING HAZARDOUS WASTES
Carlo Castaldim", Ramsay Chang, Harold Lips, and Charles Pickett
Acurex Corporation
Energy & Environmental Division
485 Clyde Avenue
P.O. Box 7044
Mountain View, California 94039
ABSTRACT
Results from several field tests of industrial boilers burning liquid hazardous waste
have indicated that these devices achieve destruction and removal efficiency (ORE) of
principal organic hazardous constituents (POHC) in excess of 99.99 percent when operating
under typical steady-state combustion conditions. Recent EPA-sponsored field tests have
focused on the effect of nonsteady-state and transient boiler operation on ORE and product
of incomplete combustion (PIC) formation. This paper presents test results on
three industrial boilers cofiring liquid wastes spiked with hazardous organic compounds.
The boilers tested comprise the major watertube design types: one gas- and oil-fired
packaged boiler, one field erected gas-fired boiler, and one coal-fired stoker. Emission
sampling included detailed evaluations of volatile and semi volatile organic pollutants in
the flue gas for POHC, ORE, and PIC determinations under several steady conditions and
induced operational upsets. In the case of the coal-fired stoker, emission evaluations
also included detailed organic and metal analyses of flue gas and flyash samples from
collection devices.
INTRODUCTION
Cofiring of combustible hazardous
wastes with conventional fuels in
industrial boilers is widely practiced in
industry as an economical disposal method
and for energy recovery. The environmental
implications of cofiring are currently
under evaluation by the Environmental
Protection Agency (EPA). Regulation of
boiler cofiring for combustible wastes was
exempted in the 1981 Resource Conservation
and Recovery Act (RCRA) which limited
incinerator DREs to 99.99 percent or
higher. This exemption was based, in part,
on the lack of information on cofiring
hazardous wastes and potential mitigation
strategies. In the interim, EPA's Office
of Solid Waste and Emergency Response has
conducted a regulatory impact assessment to
evaluate hazards and potential control
measures in preparation for recommending a
boiler regulation. To support this
regulatory assessment, EPA's Hazardous
Waste Engineering Research Laboratory and
Office of Solid Waste and Emergency
Response conducted a series of field tests,
Sites A through K, to evaluate the
capability and limitations of industrial
boilers for hazardous waste thermal
destruction by cofiring with conventional
fuels (1). While these tests showed high
waste destruction (composite weighted ORE
of 99.998 percent), the tests were
conducted at nominally steady operating
conditions and did not address transient
and off-design operating conditions which
may influence ORE.
To complete this evaluation,
information was then needed on boiler
operational limits yielding satisfactory
waste destruction, and on the effects of
routine, nonsteady operation on destruction
performance. In 1984 one further test,
Site L, was performed to evaluate these
operating conditions and obtain sufficient
ORE data to deduce an acceptable operating
445
-------
window. The relatively high ORE results of
these tests (composite ORE for all runs
equal 99.998 percent) reported at the
Eleventh Symposium (2) have dispelled some
of the initial concerns about waste
destruction during nonsteady boiler
operation. However, confirmatory tests
were necessary to validate Site L results
using alternative boiler designs.
OBJECTIVE
The overall objective of these tests
was to quantify DREs of POHCs for nonsteady
and off-design operating conditions during
cofiring In industrial boilers. The study
was also to provide background information
for determining regulatory requirements for
organic, metal, and chloride emissions
during cofiring of hazardous wastes.
APPROACH
Three watertube boilers were selected
for these tests. Table 1 summarizes design
characteristics and test operating
conditions for these boilers.
Site L is a forced draft Combustion
Engineering (CE) package watertube boiler
with a maximum rated capacity of
13.9 kilograms per second (kg/s)
(110,000 pounds per hour (lb/hr)) of
superheated steam. Natural gas and/or
No. 6 oil is fired through a single dual
air register burner. The burner is
retrofitted with two steam atomized liquid
waste guns designed to fire a distillation
byproduct containing methylmethacrylate
(MMA). The waste was spiked with CC14 and
monochlorobenzene in concentrations ranging
from 0.7 to 4.5 percent by weight. MMA
concentration varied from 6 to 8.4 percent
by weight. Spiked waste feedrate during
tests accounted for 7.3 to 56 percent of
the total heat input to the boiler. This
waste feedrate encompasses the range of
typical plant operating practice of 15 to
40 percent of the total heat input. Boiler
loads tested were as low as 23 percent of
TABLE 1. BOILER DESIGN AND OPERATING CHARACTERISTICS
Sits Boiler* Primary Wastt fuel
Identification type and fuel POHC1, wtals
capacity
(i(>3 ib/nr)
1 FO-HT Ho. 6 TSB
1 burner oil CCld,HCB,MHA
14 (110)
Natural TSB
gas CCljj.HCB.HHA
K FD-HT Natural Waste oil
4 burners gas CCla.HCB,
44 (350) 1,2,4-TCB
H BD-HT Stoker Sludge
Stoktr coal TC£;1,2,4-TCB
19 (ISO) rtJ.Cr
Stoker No. 2 oil
coal TCE;1,Z,4-TCB
Pb.Cr
Operation
Steam load Percent 02 POIIC
kg/s concentration
(103 lb/hr) (percent)
3.4 to 7.8 2.5 to 7.9 2.8 to 7.5
(27 to 62)
3,2 to 9.6 1.2 to 7.5 0.7 to 8.4
(25 to 76)
16 to 25 1.2 to 10 1.0 to 15
(125 to 195)
10 to 17 6.7 to 10 5 to 10
(80 to 135)
10 to 17 5.4 to 12 5 to 6
(80 to 135)
Haste/ Additional
primary variables
fuel
percent
heat Input
7.3 to 34 Load and Og
transients ;
waste flow
transients;
atomlzatlon ;
sootblow
8.8 to 56 Load and 02
transients ;
waste flow
transients ;
atoml zation ;
sootblow
20 to 30 ' Haste
Hgtitoff;
waste
Injection
location;
waste flow
transients;
atomlzatlon ;
sootblow
0 None
20 to 35 ' None
Sampling
protocol
.Continuous
monitors.
i«1n1-VOST,
periodic
full-VOST
Continuous
monitors.
i»1n1-VOST,
periodic
ful 1 -VOST
Continuous
monitors.
m1n1-¥OST,
periodic
ful 1 -VOST
Continuous
tnoni tors ,
«1n1-VOST,
periodic
full-VOST,
EPA HM5,
Andersen ,
ash
sampling
*FD-HT • forced draft watortube; BD-KT = balanced draft watertube.
*I58 « third stage bottom fro« production of methyl
methacrylate (MMA); ECU = carbon
tetrjchlorldti MCB » nonochlorotenzene; TCE » trlchloroetHylene; 1,2,4-TCB =
446
-------
capacity which is lower than normal
operating practice for this site. In
addition to tests performed at high- and
low-steam loads, excess air levels, and
waste/fuel ratios, several other nonsteady
and induced upset conditions were
investigated including transients, waste
atomlzation upsets, and sootblow cycles.
Site M is a field erected CE watertube
boiler with a rated capacity of 44 kg/s
(350,000 Ib/hr) of superheated steam. The
boiler is equipped with four gas and/or
oil-fired CE R-type burners arranged in a
square pattern. Combustion air is
preheated to about 500 degrees
Fahrenheit (°F). Typically, the boiler
operates at about 57 percent capacity using
a mixture of natural gas, waste process
gas, and a liquid waste. This waste, fired
with the lower 2 burners, consists of heavy
ends from a butanol/propanol production
unit combined with surface oil from a waste
retention pond. During the tests, the
waste oil was spiked with CC14, monochloro-
benzene, and 1,2,4-trichlorobenzene in
varying concentrations ranging from 1 to
15 percent by weight. The boiler was
tested at nominal- and low-steam load
conditions with various amounts of excess
air. Nonsteady and upset test conditions
were induced by operating the boiler with
extreme high and low excess air and during
wasteflow transients, waste Hghtoff, waste
atomization cutoff, and sootblow cycles. A
less favorable waste injection location was
also tested by switching waste firing to
the upper two burners. This caused very
unstable combustion conditions with
significant smoke generation.
Site N is a Riley balanced-draft
spreader stoker with a rated capacity of
19 kg/s (150,000 Ib/hr) of saturated steam.
Figure 1 illustrates a side view of the
boiler and flyash control devices. The
boiler disposes of liquid sludge through
two pressure atomized guns located opposite
each other in the furnace side wall at
approximately 3 feet (ft) above the grate.
The sludge, consisting primarily of water
(greater than (>) 90 percent), was spiked
FDIO fan
Legend -- sample location
A Baghoyse Inlet flue gas
B Baghouse outlet flue gas
C Mechanical separator, ash
0 Baghouse ash
E Bottom ash
F Coal
G Sludge
OFA Overfire afr
FDID Forced draft induced draft
Figure 1. Site H —
site layout (not to scale),
447
-------
with various quantities of tMchloro-
ethylene and 1,2,4-triehlorobenzene for a
total sludge concentration of 5 to
10 percent by volume. Lead and chromium
salts 1n water solution were also injected
to determine the partitioning of these
metals among the bottom ash, dust collector
ash, or baghouse ash. Boiler test loads
and excess air levels spanned the range of
the plant operating practice. Excess air
1n the furnace was varied by changing
undergrate and overfire airflows. Similar
tests were also performed with distillate
oil spiked with trichlorobenzene and
trichloroethylene to determine the effect
of waste heating value on POHC destruction
efficiency.
At Sites M and N, POHC injection pumps
were used to spike the waste streams. This
allowed better control of POHC
concentration in the waste and POHC
feedrate to the boiler. The sampling and
analysis protocol consisted of the
following:
• Continuous flue gas monitoring for
Og» C0£» CO, NOX, SQ2» and
unburned hydrocarbons
t Particulate and semi volatile
organic sampling using the EPA
Modified Method 5 (MM5) train;
post test semi volatile analyses by
gas chromatography/mass
spectrometry (SC/MS); priority
pollutant metals for Site N by
furnace atomic absorption (AA)
* Volatile organic sampling by
mini-volatile organic sampling
trains (VOST) and VOST protocols;
onslte GC/Hall for chlorinated
organics by direct desorbtion of
mini-VOST samples; offsite
laboratory GC/MS for priority
pollutant organics by direct
desorption of VOST samples
• HC1 sampling by Modified EPA
Method 6 (MM6) train; analyses by
titration
* Particulate size distribution and
metals by Andersen impactor stages
at inlet and outlet of Site N
baghouse; analyses by furnace AA
• Grab sample of waste and primary
fuels; SC/MS for POHC analyses.
and ultimate analyses by ASTM
methods
« Grab sample of ash streams at
Site N; selected analyses for
metals by AA, semi volatile
organics by GC/MS, and leachates
by extraction procedures (EP)
toxicity protocols
The bulk of volatile organic sampling and
analysis was done with a simplified
mini-VOST protocol (3). A simple tenax
trap was used with the VOST train for a
10 liter (L) sample of stack gas.
Immediately following the sampling, the
traps were taken to an onsite GC/Hall and
thermally desorbed. Trap QC preparation
was also performed onsite.
Table 2 summarizes unburned carbon
emissions recorded during cofiring test
conditions at the three test sites. Higher
CO and total unburned hydrocarbons (TUHC)
emissions at Sites L and M are indicative
of nonsteady and induced upset conditions.
During these tests the boiler operational
settings were selected to deliberately
induce smoking or higher CO and TUHC
emissions. For Site N, foreseen problems
in clinkering of coal ash on the grate and
baghouse blinding limited the degree of
.off-design boiler operation investigated.
RESULTS
Table 3 summarizes ORE results for
Site L. These tests produced a total of
112 ORE data points for the steady and
nonsteady quantltation of carbon
tetrachloride, monochlorobenzene and
methylmethacrylate thermal destruction.
Then ORE results, discussed in greater
detail in the Eleventh Symposium, show high
levels of thermal destruction despite
substantial upset boiler operating
conditions. Excluding few low ORE values,
which were considered outliers, the overall
average ORE for the entire test program was
99.998 percent. No significant ORE
correlation with boiler operating
conditions or surrogate emissions (CO and
TUHC) was evident in the data for Site L
indicating that the destruction or
pyrolysis of the POHCs near the burner
flame region was very efficient resulting
in low POHC emissions in the range of
background levels. PIC from this site
ranged between 1 and 270 micrograms per
cubic meter (wg/nH) corresponding to an
448
-------
TABLE 2. UMBURNED CARBON EMISSIONS DURING COFIRIMG
Site Primary Test load
Range in
average emissions*
identification fuel
L Gas
on
M Gas
N Coal
(percent of
capaci ty
45 to 70
23 to 32
42 to 56
25 to 30
46 to 50
36 to 40
85 to 90
53 to 60
CO
88 to 9,500
215 to 10,000
99 to 3,850
146 to 1,860
17 to 2,850
8 to 1,260
27 to 43
37 to 116
TUHC
0 to 25
0 to 10
0 to 2.6
0 to 0.2
0 to 120
8 to 250
0 to 13
0 to 27
Dry ppm corrected to 3 percent 02-
TABLE 3. ORE SUMMARIES ~ SITE L
POHC
MMA
CC14 + MCB*
CC14
MCB
CC14
CC14
MCB
MCB
No.
tests
16
96
48
48
21
27
21
27
Series
All runs
All runs
All runs
All runs
Oil firing
Gas firing
Oil firing
Gas firing
Average ORE
99.9995
99 .9949
(99.9980)t
99.9919
(99.9981}t
99.9980
99.9975
99.9874
(99.9985)t
99.9991
99.9971
Monochlorobenzene.
^Excluding one low {99.7 percent) value for
CC14.
449
-------
average PIC/POHC mass ratio of 15.
Dlchloromethane and tetrachloroethylene
were the principal organics detected.
Table 4 summarizes DREs measured
during Site M tests under all test
conditions using mini-VOST, VOST, and MM5
sampling techniques. The 91 ORE data
points are grouped according to test
conditions. Overall, except for one data
point, the DREs for the chlorinated POHCs
exceeded 99.99 percent under all test
conditions even during significant
nonsteady and smoky operation. Benzene,
the only nonchlorinated organic compound
detected in the waste oil, showed poor ORE,
as low as 94 percent. However, benzene
concentrations in the waste oil were very
low (generally 1000 parts per million
(ppm)) and benzene is readily produced as a
PIC from other organics in the waste oil
and from natural gas combustion. Thus,
benzene is probably not an accurate
indicator of actual DREs during these
tests. The high DRE values measured for
the chlorinated POHCs did not seem to be
affected by the various test conditions
aimed at producing poor combustion
environments with high smoke and CO
emissions. Values of the DREs measured
under these conditions were similar to
baseline values measured under normal load
and excess air conditions. The above
observations are further confirmed by plots
of DRE values for each chlorinated POHC as
a function of key boiler operating
parameters such as boiler load, 02 levels,
CO, TUHC, and smoke and NOX emissions. An
example of the lack of correlation is
displayed in Figure 2 where DRE of CC14 is
plotted as a function of average CO
emissions. The DRE seemed to increase with
the rate of POHC injection, a finding
similar to Site L. A plot of -log
(l-DRE)/rate of POHC injection versus TUHC
for monochlorobenzene (MCB) (Figure 3)
shows that DRE, when normalized by the POHC
Injection rate, decreased as the TUHC
emissions increased, a trend which is
consistent with poorer destruction
efficiency with poorer combustion. A
similar trend can be observed for CC14-
These results indicate that analysis of the
data should include considerations of
possible interactions among several
parameters which may affect DRE. Simply
plotting DRE versus a specific parameter,
such as CO levels without minimizing the
possible effect of other factors, may not
produce an observable trend. Total
chlorinated PIC emissions during cofiring
at Site M ranged between 2 and 360 ug/m .
The PIC/POHC ratios were consistent with
values obtained at Site L with the majority
of the values ranging between 0,5 and 5
(average 2.4 for 45 samples). The data
indicated a weak correlation of PIC
emissions with TUHC possibly indicating
some dependence on combustion efficiency.
Table 5 summarizes the DRE for
trichloroethylene measured using mini-VOST
and VOST during six combustion conditions
investigated at Site N, Results indicate
generally good agreement between the two
sampling techniques. DRE for.this spiked
POHC exceeded 99.999 percent in all tests.
Some drop in destruction efficiency seems
apparent during high excess air conditions
which likely resulted in lower combustion
temperatures. Spiked lead and chromium
were segregated into ash streams collected
by the collector and baghouse.
ACKNOWLEDGEMENT ,
This research effort was sponsored by
the Environmental Protection Agency under
Contract 68-03-3241. Ivars Lids and
Bob Mournighan of the Hazardous Waste
Engineering Research Laboratory, and
Marc Turgeon of the Office of Solid Waste
and Emergency Response were the Project
Officers. Their support and assistance
during all phases of this effort was
greatly appreciated. The authors also wish
to acknowledge the contribution and
logistical support from the management and
technical personnel at each host site.
REFERENCES
1. Castaldini, C., et a!., "Engineering
Assessment Report, Hazardous Waste
Cofiring in Industrial Boilers," Acurex
Report TR-84-159/EE to USEPA under
Contract No. 68-02-3188, June 1984.
2. DeRosier, R. J., et al., "Nonsteady
Industrial Boiler Waste Cofiring
Tests," in Incineration and Treatment
of Hazardous Waste, Proceedings of the
Eleventh Annual Research Symposium,
EPA/600/9-85/028, September 1985.
3. Spannagel, U., et al., "Mini-VOST Field
Analytical Protocol," Acurex Draft
Report FR-85-159/EED to the USEPA under
Contract 68-03-3241, July 1985.
450
-------
TABLE 4. ORE GROUPED BY TEST CONDITIONS — SITE M
Test.
number
Overall description
ORE
CC14
MCB
TCB
Benzene
HV-27 Baseline with MCB, CC14, TCB 99.99961 99.99968
MMS-7 Baseline with MCB, CC14, TCB 100 ...
MV-32 Normal load and air 99.99837
MV-33 Normal load and air 99.99767
MV-9 Maximum fan capacity . . , 99.99962 ... ...
NV-10 Maximum fan capacity ... 100 ... ...
HV-13 High air ... 99.99981
FV-6A High air 100 99.99689 . . . 99.76775
HV-28 High air 99.99988 99.99886 ,
MMS-8 High air ...... 100 ...
MV-S Low air, smoky . . . 99.99491
MV-11 Low air, hazy . . . 99.99924
HV-1Z -Low air, hazy . . . 99.99978
FV-3 Low air, hazy . . . 99.99861 . . . 94.71648
MV-21 High oil firing, smoky, low air 99.99949 99.99956 ... ....
MV-22 High oil firing, smoky, low air 99.99940 99.99855
MV-52 Low air, hazy stack 99.99937 99.99947
MV-53 .Low air, smoky, high MCB & CC14 99,99932 99.99990 ... ...
FV-13 Low air, high smoke, high MCB and CC14 99,99968 99.99938 . . . 97.49198
FV-12 Low air, hazy 99.99987 99.99860 . . . 99.68780
MM5-12 Low air, high TCB, smoky ... ... 99.99773 . . .
HH5-13 Reduced TCB, low air, hazy ... ... 99.99979 . . .
FV-10 High TCB, low air, smoky 98.64499
MV-6 Low air, atontlzation off . . , 99.99749 ... ...
FV-2 Low air, atonrtzatlon off ... 100 ... ...
HV-7 Low load, no waste gas ... 99.99166 ... ...
MV-8 Smoky, low load & air, no waste gas . . , 100 ... ...
MV-18 Low load 99.99933 99.99705 ... ...
MV-19 Low load, low air 99,99951 100
FV-4 Low load, low air 99.99681 99.99789 . . . 98.29147
MV-26 Low load, lightoff, smoky - 99.99922 99.99967
FV-5 Low load, lightoff, smoky ' 100 99.99973 . . . 99.99779
MM5-6 Low load, lightoff, smoky 99.99991 . . .
HM5-10 Unsteady oil flow, low load 99.99934 . . .
MV-55 High MCB & CCU 99.99955 99.99996
MV-54 High MCB & CC14 99.99963 99.99994
MV-51 High air, high MCB 4 CCU 99.99875 99.99989
FV-11 High air, high MCB & CCU 99.99976 99.99884 . . . 99.86947
HV-34 High CC14 99.99931
FV-6B High CC14 99.99911 ... ... 99.59178
HV-35 High CC14 99.99928
FV-7 High CC14 99.99865 98.69712
MV-38 High CC14 99.99981 ...
MM5-11 High TCB 99.99899 . . .
HV-36 Low CC14 99.99454
HV-37 Low CC14 99.99793 .
MV-4 MCB spike start . . . 99.99792
FV-1 MCB spike start . . . 99.99991 . . . 99.10608
HV-17 011 lightoff, low load 99.99890 99.99992
MV-14 Sootblowlng . . . 99.99982
MV-15 Sootblowlng . . . 99.99994
MM5-14 Sootblowlng 99.99895 . . .
MV-29 Sootblowlng 100 100
MV-30 Sootblowlng 99.99980 99.98762
HM5-9 Sootblowlng 100 ...
MV-20 High oil firing 99.99976 99.99976 ... ...
MV-23 Smoky 2 top burners only 99.99975 99.99941 ... ...
FV-9 Unsteady oil flow, low load ... ... ... 100
Average ORE 99.9991 99.9987 99.9994 98.8873
451
-------
99.99999
OJ
o
i.
03
99.9999 -
99.999 1
99.99
.5- 99.9
CU
O
99
90
0
a
D rfl
a
o
0.4 0.8 1.2 1.6 2 2.4
(Thousands)
AVG CO AT 3% 02 (ppm)
a CC14
Figure 2. ORE versus average CO, CC14 -- Site M.
2.8
2
O.
03
O
O
O
uf
Q
I
o
Q
2O
SO
AVG TUHC AT 3% O2 (ppm)
O MCB
Figure 3. DRE/injectlon rate versus average TUHC, MCB — Site M.
452
-------
TABLE 5. ORE RESULTS FOR TCE — SITE N
Test description
High load, baseline,
sludge cofiring
Low load, high air,
sludge cofiring
Low load, high air,
sludge cofiring
Low load, low air,
sludge cofiring
Low load, baseline,
sludge cofiring
Low load, baseline,
oil cofiring
TCE
injection
gpm
0.20
0.20
0.44
0.23
0.21
0.24
TCE emissions, ORE (percent)
pg s
Hini-VOST VOST Hini-VOST VOST
30 to 31 38 99.9998 99.9998
31 to 96 72 99.9996 99.9996
100 NA 99.9997 NA
<0.6 NA 100 NA
5.7 to "7. 7 <12 99.99997 >99. 99994
2.8 NA 99.99998 NA
NA — not available, measurement not performed or analytical
interference.
453
-------
BIODE6RADATION OF CHLOROPHENOXY HERBICIDES
D. Roy* and S. Mitra**
*Env1ronmental Engineering Program, Department of Civil Engineering,
Louisiana State University, Baton Rouge, LA 70803, USA
**Biology Division, P.O. Box Y12, Oak Ridge National
Laboratory, Oak Ridge, TN 37831, USA
ABSTRACT
Pesticides of all kinds as a broad class of man-made chemicals constitute approxi-
mately 2 percent of the total chemicals used by our society and herbicides make up about
50 percent of the quantities of pesticides used in the United States. Chlorinated
phenoxy alkanoic acids, particularly 2,4-D (2,4 Dichlorophenoxy acetic acid) and 2,4,5-T
(2,4,5 Trichlorophenoxy acetic acid) have maintained an important position in the herbi-
cide market because of their widespread use. In our ongoing research on biodegradation
of chlorinated herbicides and pesticides, we have isolated two new strains of Pseudo-
rnonas which can utilize chlorophenoxy alkanoic acids as their sole source of carbon and
energy. These Pseudomonas isolates were obtained through the process of selective
enrichment from a mixed microbial population of a municipal activated sludge treatment
plant. This paper reports on the first phase of the study investigating the applica-
bility of the two new isolates in detoxification and biodegradation of chlorophenoxy
herbicides.
INTRODUCTION
Pesticides of all kinds as a broad
class of man-made chemicals constitute
approximately 2 percent of the total
chemical used by our society and herbi-
cides make up about 50 percent of the
quantities of pesticides used in the
United States (EPA, 1974). Chlorinated
phenoxy alkanoic acids, particularly
2,4-D (2,4 Dichlorophenoxy acetic acid)
and 2,4,5-T (2,4,5 Trichlorophenoxy
acetic acid) have maintained an important
position in the herbicide market because
of their widespread use. They have been
extensively used to control weeds in
cropland and aquatic systems, around
railroad trucks and oil tanks; to main-
tain desirable turfs on lawn, golf
courses, parks and other land areas.
Also, there have been reports of acci-
dental leakages and spills of these
lerbicides in large quantities to the
environment posing an acute health
hazard.
The naturally occurring microbes in
the environment were not exposed to such
man-made herbicides and did not have the
necessary enzymes to degrade these com-
pounds. Due to continuous exposure of
natural microorganisms to these herbi-
cides, a few group of microorganisms have
come up with the enzymatic system to be
resistant to such toxic compounds and are
capable of degrading them, although at a
slow rate.
Biodegradation of 2,4-D and 2,4,5-T
by naturally occurring microorganisms or
laboratory constructed strains has been
the subject of intensive research in the
recent past and has been reviewed by
Shosal et a]_. (1985). Several
researchers in the past decade have
isolated a number of strains capable of
biodegrading 2,4-D and 2,4,5-T (Pemberton
and Fisher, 1977; Chakrabarty et al.,
1973; Suflita et al_., 1982). In our
ongoing research on biodegradation of
chlorinated herbicides and pesticides, we
have isolated two new strains of Pseudo-
454
-------
monas which can utilize chlorophenoxy
alkanoic acids as their sole source of
carbon and energy. These Pseudomonas
isolates were obtained through the pro-
cess of selective enrichment from a mixed
microbial population of a municipal acti-
vated sludge treatment plant. The pre-
sent study was undertaken to evaluate the
kinetics of biodegradation of 2,4-D by
two new Pseudomonas isolates and a well
characterized microorganism, A1callgenes
eutrophus, In mixed culture and mixed
substrate environment. Also, the pure
culture kinetics of these microorganisms
under similar conditions have been pre-
sented.
MATERIALS AND METHODS
Microorganisms
Through the process of selective
enrichment, two new Pseudomonas have been
isolated from the mixed microbial culture
obtained from a local municipal Sewage
Treatment Plant. Routine biochemical
tests were performed to determine the
taxonomic classification of these iso-
lates. Based on these properties, an
attempt was made to match their profiles
with those .included in the rapid NFT data
base of OMS Laboratories, Inc. (Plain-
view, New York). These new isolates were
identifiable at the genus level to be
Pseudomonas. This was further verified
by independent testing conducted by API
Analytab Products (Plainview, New York).
These two isolates will be referred
to as Pseudomonas Isolate DR 101 and
Isolate DR 201. These isolates are gram
negative, non-fermentative, motile rods.
Isolate DR 101 is white with approximate
size of 3.7 x 1.3 p with polar flagella.
The isolate DR 201 is approximately 2.7 x
1.1 p in size with a pale yellow color
and peritrichious flagella.
A. eutrophus containing 2,4-D
degrading plasmid was originally isolated
by Professor J. M. Pemberton of the Uni-
versity of Queensland, Australia, and
provided to us by Professor A. M. Chakra-
barty of the University of Illinois,
Chicago, USA.
Chemicals
Stock solutions of 2,4 Dichloro-
phenoxyacetic acid (2,4-D) (Pfaltz and
Bauer, Inc., Connecticut) was prepared by
dissolving appropriate quantities in
0.1 N NaOH to produce final concentration
of 25 mg/ml.
Basal Salt Media (BSM) (IQx)
Basal salt medium (BSM) was prepared
by dissolving 58.0 g K9HPO,,, 45.0 g
KH2POa, 20.0 g (NH.LsO,, T.6 g MgCl?,
200.(Fmg CaCU, 2070 mg^NaMoQ. and L
10.0 mg MnCljTln 1 liter HgO.
Determination of Chloride Release
Chloride ion concentration in
culture fluids was determined using a
chloride ion-specific electrode Model
No. 94-17 and a reference electrode Model
No. 90-02 (Orion Research).
Culture Media
Culture media for kinetic studies
was prepared by appropriate dilution of
lOx BSM media and the stock solution of
the chemicals. Microbiological media for
biochemical reactions and growth were
prepared as per Sergey's Manual (1984)
and BBL Manual of Products (1969)
(Cockeysville, MD).
Experimental Techniques
The kinetic experiments were per-
formed in batch reactors at a constant
temperature of 30°C in a shaker water
bath. No air other than that transferred
due to shaking was provided. Pure cul-
tures from log growth phase were added as
seed microorganism at a 1:20 ratio of
seed volume to total culture fluid
volume. Approximately 5 ml samples were
withdrawn aseptically at periodic inter-
vals to monitor the growth and substrate
utilization. Growth was monitored by
measuring optical density at 540 nm wave-
length using a Zeiss spectrophotometer
Model No. PMQ-2 (W. Germany) with a
1.0 cm cell path. Substrate utilization
was monitored by measuring chloride ion
release due to biodegradation of 2,4-D.
RESULTS AND DISCUSSION
Single Substrate Studies
A1callgenes eutrophus isolated by
Pemberton and Fisher (1977) has been
studied extensively in the past. It has
455
-------
been shown to degrade 2,4-D and the
degrading enzyme has been shown to be
coded on plasmid pJP4 (Don and Pemberton,
1981). A. eutrophus was used in this
study as~~the reference microorganism to
compare and contrast the results of the
new Pseudomonas isolates DR 101 and
DR 2UT.
Growth of A. eutrophus utilizing
2,4-D was studied in a batch reactor at
30°C. The results of this experiment
shown in Fig. 1 indicate a lag period of
100 hours during which neither microbial
growth nor substrate utilization was
noted. Following the lag period, micro-
bial growth and substrate utilization as
monitored by the chloride release were
noted to proceed at a rapid rate and then
leveling off after approximately
40 hours. The growth kinetics followed
the typical pattern of batch growth. It
should be noted that the seed micro-
organism was obtained from a fresh
culture of Alcaligenes growing on 2,4-D,
thereby eliminating the need for any
acclimation period. This observation
suggests that the enzyme system necessary
for the complete degradation of 2,4-D is
inducible rather than constitutive in A.
eutrophus.
Growth kinetics of Isolate DR 101
utilizing 2,4-D was studied in a batch
reactor at 30°C. The results presented
in Fig. 2 show a similar pattern as that
for the Alcaligenes. The lag period for
Isolate DR 101 was approximately
60 hours.
The kinetics for the growth of
Isolate DR 201 of 2,4-D at 30°C is pre-
sented in Fig. 3. Isolate DR 201 appears
to have approximately 5 days lag period
followed by a rate of growth which is
comparable to that of Isolate DR 101
(Fig. 2).
Further experiments were performed
using respirometer to monitor the oxygen
uptake associated with the growth of
Isolates DR 101 and 201 on 2,4-D. It was
verified from these experiments (results
not shown here) that the oxygen consump-
tion followed a similar trend as shown in
the growth curves (Figs. 1-3) with an
extended period of lag followed by a
rapid rate of degradation.
Based on the pure culture growth
kinetics of the Pseudomonas Isolate
DR 101 and DR 201, it is apparent that
like Alcaligenes, the two isolates pos-
sess a 2,4-D degradative enzyme system
which is inducible rather than constitu-
tive.
It is apparent from the growth
curves presented above that models such
as Monod's type would be inappropriate to
explain the extended lag period observed
in the degradation of 2,4-D by Pseudomo-
nas Isolates DR 101 and DR 201. Research
work is in progress to develop kinetic
models based on the concept of substrate
inhibition and/or other semi structured
model approaches. The kinetic constants
for these models will be presented in the
future.
Multiple Substrate Studies
Batch culture of pure organics grown
on a mixture of two carbon and energy
sources has been reported to follow a
diauxic growth pattern which is caused by
sequential substrate utilization (Monod,
1949; Harder and Dijkhuizen, 1982).
Also, it has been shown that the presence
of a preferred substrate permitting
higher growth rate prevents the utiliza-
tion of a "poorer" substrate in batch
culture (Harder and Dijkhuizen, 1976).
This argument seems to be logical in
light of the fact that the enzymes for
biodegradation of poorer substrates are
coded on plasmid DNA, and the plasmid DNA
could be lost if the microorganism is
grown on a simple source of carbon and
energy (Kilbane et al_, 1982; Roy, 1985).
On the contrary, it has often been hypo-
thesized that biodegradation of hazardous
and toxic herbicides and pesticides by
microorganism can be enhanced by the
presence of a non-toxic carbon source.
This kind of fortuitous biodegradation of
an organic compound which is not neces-
sarily a growth substrate can be accom-
plished by the process of cometabolism.
Cometabolism is defined as the trans-
formation of a non-growth substrate in
the obligate presence of a growth sub-
strate or another transformable compound
(Dagley, 1971; Dal ton and Stirling,
1982). Application of this concept in
the field of biodegradation of recalci-
trant compounds has been reviewed by
Dal ton and Stirling (1982). To study
this effect, growth of pure culture of
456
-------
0.8
0.7k
0.6
0.5
O Q4
O
0.3
0.2
0.1
FLASK
ODs«o
1.6
1.4
1.2
1.0
tu
0.63
Ul
or
0.6^
O
0.4
0.2
^ iiiiiiiiiiiiiiiiiiiiii
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 ISO 160 170 ISO 190 200 210 22O 230
TIME (hrs.)
Figure 1. Growth kinetics of A. eutrophus using 2,4-D at 30°C.
0.7
0 20 4O 60 80 100 120 I4O 160 180 200 220 240 260 280 300 320 340
TIME (hrs)
Figure 2. Growth kinetics of Pseudomonas Isolate DR 101
using 2,4-D at 30°C.
457
-------
2o 4o so eo 100 120 MO teo mo 200 so wo zsoao 900 220 MO
TIME Chrsl
Figure 3. Growth kinetics of Pseudomonas Isolate DR 201
using 2,4-D at 30°C.
0 20 40 60 80 IOO 120 140160 ISO 200 220 240 260 280 300 32O 340 36O 280 4OO 420 44O 460
TIME (hrs.)
Figure 4. Growth kinetics of Isolate DR 101 using multiple
substrate of glucose plus 2»4-D at 30°C.
458
-------
Isolate DR 101 was investigated using
multiple substrate of 2000 mg/1 of
glucose and 1500 mg/1 of 2,4-D. The seed
culture for this experiment was Isolate
DR 101 grown on 2,4-D. The results of
this experiment presented in Fig. 4
indicate a similar pattern of growth as
noted for Isolate DR 101 on glucose only.
However, growth of the microorganism was
not associated with the release of
chloride ion. This observation suggests
that the growth of Isolate DR 101 in
mixed substrate of glucose and 2,4-D was
at the expense of glucose only without
any degradation of 2,4-D.
Mixed Culture Studies onSingle Substrate
Mixed culture experiments using
2,4-D as the sole source of carbon and
energy was studied using bacterial
strains Pseudomonas isolate DR 101,
DR 201 and A. eutrophus. Results of
these studies are shown in Figs. 5-6.
A. eutrophus present with either Isolate
DR 101or OR 201 was noted to initiate
the growth without any lag period a,nd the
rate of total growth of microorganisms
monitored by QDj-an did not seem to be
very different from the rate of growth of
pure cultures (Figs. 5 and 6). The
extended lag period for Isolates DR 101
and 201 could be eliminated when they are
present in mixed culture (Fig. 7),
although the rate appears to be slower
than the growth of individual micro-
organisms. From a comparison of the pure
culture studies and the mixed culture
studies using Pseudomonas Isolate DR 101
and DR 201 and A. eutrophus, it is appar-
ent that the dehalogenase enzyme systems
present in these microorganisms are not
identical and they complement each other.
This observation is apparent from the
enhanced growth noted using mixed cul-
tures. It is also possible that the
utilization of the rate limiting inter-
mediates in the metabolic pathways of
individual microorganisms are different
from each other. In the event of that
possibility, one intermediate which was
slowing down the overall process of uti-
lization by one microorganism is utilized
by the other microorganism at a faster
rate resulting in an enhanced removal by
mixed cultures. However, the exact
reasons for the improved performance of
biodegradation by mixed cultures have not
been experimentally demonstrated for the
microorganisms reported in this study.
Growth of individual microorganisms
present in mixed culture was further
studied by monitoring flow cytometric
profiles of mixed cultures at different
time. The results of this study (not
shown here) using A. eutrophus and
Isolate DR 101 on 2,4-D indicated that
starting with almost equal distribution
°f A- eutrophus and Isolate DR 101,
A. eutrophus seem to prevail in the
initial stage followed by a gradual shift
towards the growth of Isolate DR 101
before reaching a final mixed population
level.
CONCLUSIONS
Based on the findings of this
research, the following conclusions may
be drawn:
1. Biodegradation of 2,4-D by the two
new Pseudomonas Isolates DR 101 and
DR 201 is comparable to that by
Alcaligenes eutrophus.
2. The dehalogenase enzyme system
necessary for complete biodegrada-
tion of 2,4-D by these microorga-
nisms are likely to be inducible
rather than constitutive.
3. For all three microorganisms
reported in this study, there is an
extended lag period before which no
biodegradation of 2,4-D is noted.
4. Mixed substrate utilization by a
pure culture indicated that there
was no biodegradation of 2,4-D, a
"poorer" substrate, when glucose is
present.
5. Biodegradation of 2,4-D by mixed
cultures of any one of the new
Pseudomonas isolates and A^_ eutro-
phus is noted to be superior to that
observed by the pure cultures.
ACKNOWLEDGEMENT
The investigation presented in this
study was funded by the U.S. Environmen-
tal Protection Agency (EPA) under ORD/EPA
cooperative agreement CR 809714 to the
Hazardous Waste Research Center, Louisi-
ana State University. However, the study
does not necessarily reflect the views of
459
-------
0.8
O.7 -
0 IO 20 30 40 SO 60 70 80 90 100 110 120 130 140 ISO
TIME (hrs.)
Figure 5. Mixed culture growth kinetics of Pseudomonas Isolate
DR 101 plus A. eutrophus on 2,4-D at 30°C.
0.7
0 10 20 30 40 SO 60 TO 80 90 100 110 120 130 140 150 160 170 160
TIME (hrs.)
Figure 6. Mixed culture growth kinetics of Pseudomonas Isolate
DR 201 plus A. eutrophus on 2,4-D at 30°C.
460
-------
0.7
20 40 60 80 100 ISO 140 160 180 200 22O 24O 260 280 300
^TIME (hrs.)
Figure 7. Mixed culture growth kinetics of Pseudomonas Isolate DR 101
plus Pseudomonas Isolate DR 201 on 2,4-D at 30°C.
the EPA and no official endorsement is
inferred.
REFERENCES
1. Chakrabarty, A.M., Chou, 6. and
Gunsalus, I.C., 1973. Genetic
regulation of octane dissimilation
plasmid in Pseudomonas. Proc. Natl.
Acad. Sci., 70. 1137-1140.
2. Dagley, S., 1971. Cometabolism of
aromatic compounds by microorgan-
isms. Adv. Micro. Physiol., 6,
1-46.
3. Dalton, H. and Stirling, D.I., 1982.
Co-metabolisms. Phil. Trans. R^
Soc. Lond., B 297, 481-496.
4. Don, R.H. and Pemberton, J.M., 1981.
Properties of six pesticide degrada-
tion plasmids isolated from Alcali-
genes paradoxus and Alcaligenes
eutrophus^. JT"Bacteriol. ,1457 3',"
681-686."
5. EPA, 1974. -HERBICIDE Report by
Hazardous Materials Advisory Commit-
tee. USEPA Report No. EPA-SAB-74-
001, May.
Shosal, D., You, I.S., Chatterjee,
O.K. and Chakrabarty, A.M., 1985.
Microbial degradation of halogenated
compounds. Science, 228, 4696,
135-142.
Harder, W. and Dijkhuizen, L., 1976.
Mixed substrate utilization in
microorganisms. In: Continuous
Culture, Vol. 6, A.C.R. Dean, D.C.
Ellwood, C.6.T. Evans and J. Melling
(Eds.), Chi Chester and Oxford:
Ellis Norwood.
Harder, W. and Dijkhuizen, L., 1982.
Strategies of mixed substrate utili-
zation in microorganisms. Phil.
Trans. R. Soc. Lond., B 297, 459-
Kilbane, J.J., Chatterjee, O.K.,
"Karns, J.S., KeTlog, S.T. and
Chakrabarty, A.M., 1982. Pure
culture of Pseudomonas cepacia.
Appl. Environ. Microbiol., 44, 1,
72-78.
461
-------
EARTHEN LINERS: A FIELD STUDY OF TRANSIT TIME
Albrecht, K, A., B. L. Herzog, and R, A. Griffin
Illinois State Geological Survey •
Champaign, IL 61820
ABSTRACT
Performance of a field-scale com-
pacted earthen liner will be evaluated in
a three-year USEPA-funded multidiscipli-
nary project which began in 1985. Objec-
tives are to (1) determine transit times
for water flow and nonreactive solute
transport through a partially-saturated
field-scale liner and (2) test the
accuracy and practicality of available
methods for predicting these transit
times. Problems being addressed include
the accuracy of lab tests in predicting
field behavior; observation scale effects
at the field-scale, sample volume-scale
and pore-scale (the possible presence of
a biocontinuum system); and hysteresis.
Saturated hydraulic conductivity,
texture, clay mineralogy, and compaction
tests will be performed on a number of
native soils to select the most suitable
material for the liner. Laboratory
column studies will simulate field be-
havior of moisture movement and transport
and aid in the prototype and final liner
designs. Construction of a prototype (of
sufficient size to accomodate full size
compaction equipment) in 1986 will deter-
mine whether the moisture content/density/
hydraulic conductivity relationships
determined in laboratory studies can be
achieved in the field and verify the re-
quired amount of compaction. Construction
of the field-scale liner (about 25 x 30 x
1 meters) will be in 1987. The liner will
be ponded with water and a series of dif-
ferent nonreactive tracers will be applied
at staggered time intervals. Movement of
the water through the liner will be moni-
tored by measuring soil-water matric po-
tentials, water contents, infiltration and
drainage. Transport through the liner
will be monitored by measuring tracer con-
centrations in the drainage and from soil-
water suction samplers. Sampling schemes
and data analyses will take advantage of
spatial dependence of soil properties over
small land areas by applying geostatlstical
concepts. Lab and field results will be
used in unsaturated nonsteady flow and
transport models to develop performance
criteria such as times to reach specific
fluxes or concentrations at the liner
bottom.
462
-------
DEVELOPMENT OF PROCESS MONITORS FOR HAZARDOUS WASTE INCINERATION
Sharon L. Nolen, Raymond G. Merrill
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Larry Name!
Acurex Corporation
Research Triangle Park, NC 27713
ABSTRACT
The Technical Support Office (TSO)
of the Air and Energy Engineering Research
Laboratory (AEERL) has been involved in
the evaluation and development of instru-
mentation for monitoring the operation of
hazardous waste incinerators. The result
of the primary thrust of this work is the
Hazardous Air Pollutants Mobile Laboratory.
The laboratory is housed In a 27 foot long
van and includes continuous monitors for Qg,
CO, C02» NOX, SOg, and HC1; a gas chromato-
graph (GC) equipped with a Flame lonization
Detector (FID); and a mass spectrometer (MS)
for identification of the organics and con-
tinuous monitoring of up to eight selected
ions. All data are acquired by two on-
board computers.
The laboratory was designed for re-
search of hazardous air pollutants and
incineration. It is easily moved to
location and may be set up for operation
in 2 days. The laboratory has been
used for monitoring incineration of
hazardous wastes, woodstove emissions,
and other combustion sources. Typical
results from two'combustion sources,
EPA's rotary kiln incinerator and wood-
stoves, will be presented.
When the mobile laboratory was used
to monitor the rotary kiln, the continuous
monitoring data reflected rapid fluctu-
ations in emissions due to the batch
loading sequence used by the kiln. This
fluctuation complicates the organic
sampling procedures. A modification of
the EPA Volatile Organic Sampling Train
(VOST) was used during kiln waste charges
to collect multiple loadings. The
concentrated sample was then heat-desorbed
onto the GC/MS for analysis.
Continuous monitoring data and organic
samples were taken with the mobile labor-
atory from a woodstove stack. The
continuous monitors were used to confirm
steady state combustion typical of
woodstoves. Organic samples were taken
from the stack and from a woodstove
impacted atmosphere for comparison. A
family of related compounds seems to be
characteristic of residential wood
combustion.
In addition to the mobile laboratory,
TSO is evaluating other innovative
approaches to continuous monitoring. , The
Fourier Transform Infrared Spectrometer
(FTIR) is currently being evaluated as a
continuous monitor for organic compound
classes from hazardous waste incinerators.
The project is in its early stage; sensi-
tivities are being evaluated and ways to
eliminate interference from water and C02
are being developed. Techniques for
interpreting relative concentration from
functional groups continuously over the
duration of a test burn are also being
explored. Currently available hardware
and FTIR software show promise in these
areas. Progress to date and preliminary
information related to the feasibility
of the FTIR as a continuous monitor will
be presented.
463
-------
EPA EVALUATION OF IN-TANK METHODS FOR DETECTING
LEAKS FROM UNDERGROUND STORAGE TANKS
James W. Starr
Enviresponse, Inc.
Livingston, NJ 07039
The U.S. Environmental Protection
Agency is charged by the 1984 amendments to
the Solid Waste Disposal Act with issuing
regulations to eliminate leaks from
underground storage tanks, insofar as
possible. Preparatory to determining the
policy on which these regulations are
based, supporting research is being
undertaken in several areas. One of the
areas concerns the methods of determining
leaks in underground storage tanks for
chemicals or petroleum fuels.
A prior study identified 36 leak
detection techniques that may be subdivided
into quantitative methods, non-quantitative
methods, inventory methods, and monitoring
methods. The primary concern of this
evaluation is the first two groups, which
contain methods that are used inside the
tank itself to determine whether a tank is
leaking (and for quantitative methods, also
how fast). A full-scale, controlled-
condition test apparatus is being designed
and fabricated. This apparatus will be
used to generate high quality data that
will be used to screen the methods to
estimate under what sets of conditions each
may be expected to work well. Later, the
practitioners of selected methods will have
an opportunity to verify the estimate for
representative conditions. The resulting
EPA research report is expected to contain
a description of the test apparatus, the
screening process, the verification
process, and the results. Preliminary
results are scheduled to be available by
the time of this conference. These results
are expected to include a lower threshold
detectable leak rate (and the probabilities
of leak detection and false alarms at that
rate) for each method evaluated.
464
-------
EPA HAZARDOUS WASTE CONTROL TECHNOLOGY DATA BASE
Cathy S. Fore
DOE Hazardous Waste Management Program
Oak Ridge, TN 37831
ABSTRACT
The Environmental Protection
Agency (EPA) Hazardous Waste Control
Technology (HWCT) Data Base functions
as a comprehensive resource for
detailed information 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 data and information
features of the data base incorporate
(1) technical engineering data on
permit applications for existing, new,
and research development and
demonstration (RD&D) facilities, and
trial burn and design data; (2)
specialized interactive input,
retrieval, and report generation
capabilities; and (3) methods for
conducting similarity analyses. Data
input and/or edit screens have been
designed for completion of Parts A
through D for all permit conditions
applicable to thermal treatment.
Interactive menu-driven retrieval
sessions have been designed to generate
summary reports on a personal
computer. Currently, the data can be
retrieved by permit application,
thermal treatment process, specific EPA
region, heating value, waste component,
and EPA waste code. In addition, the
data base can be used to conduct
similarity analyses of the data for
(1) reevaluating the technology data to
compare with actual industry
performance and operating conditions;
(2) providing a reference guide in
order to meet new and existing
regulatory standards; and (3) providing
a means for calculating the theoretical
performance of trial burns for research
capabilities. Overall, the HWCT data
base functions as a means of tracking
the status of permit applications and
RD&D, assists 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 development of this
data base is a joint effort by EPA and
the U.S. Department of Energy's
Hazardous Waste Remedial Actions
Program.
465
-------
THE HWERL DETOXIFICATION RESEARCH LABORATORY
Alfred Kernel, Charles J. Rogers, Harold L. Sparks
U. S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
The in-house HWERL Detoxification
Laboratory was set up to identify, in
vestigate and evaluate innovative chemical
and physical methods for decontamination of
soils, sediments and other sources which
are tainted with halogenated organics such
as PCBs, chlorodioxins or other high hazard
organic materials. The goals of this lab-
oratory are to validate selected alterna-
tive methodologies for decontamination
schemes which are both environmentally
sound and cost effective. The laboratory
1s equipped to handle a variety of bench
scale experimental processes as well as the
analytical instrumentation required for
routine and sophisticated quali- and quan-
titative organic analyses. The laboratory
work requires that rapid analytical results
be obtained so that processes can be moni-
tored and adjusted for maximum efficiency.
Most process results can be obtained with.
1n a few minutes to 24 hours after a basic
experiment 1s Initiated.
State-of-the-art analytical instrumen-
tation consists of both a gas chromatograph-
data system (GC-DS) and gas-chromatograph-
mass spectrometer-data system (6C-MS-DS).
The GC 1s a Spectraphysics-7100 (SP-7100)
with both hydrogen flame ionization and
electron capture detectors. The GC is cap-
able of using packed as well as capillary
columns and a variety of inlets from flash
vaporization to cold on-column injection
for capillaries. The detectors outputs are
interfaced to an IBM CS-9000 laboratory
data system for data storage, manipulation
and presentation.
The GC-MS-DS system is a Hewlett
Packard (HP) MSD 5970B based system. The
GC system permits use of capillary columns
and either split/split!ess or direct cold
on-column injection both of which can be
automated by use of the HP-7673A auto
sampler. The MS portion of the system is
based on HP hyperbolic quadrupole tech-
nology and has a useable scan range of from
10 to 820 Daltons. Either full scan or
selected ion monitoring capabilities can be
utilized as sensitivity requirements dic-
tate. The DS of the HP GC-MS fully con-
trols all of the GC and MS functions, it
consists of color work station, 55 MByte
Hard disk with Linus tape Interface, dual
5 1/4" floppy drives printer and associated
interfaces.
The SP-7100 system is routinely used
for both liquid and gaseous analysis for
quanti- as well as qualitative results.
The HP GC-MS-DS can be used for identifica-
tion of compounds as well as quantitation
of known, and in some cases unknown or
reaction products.
Some examples of current in-house
research include PCB/Aroclor analysis in
APEG dehalogenation systems, destruction
of haloaliphatics, such as EDB, EDC, di-
chloromethane, carbontetrachloride, etc by
means of catalytically induced dehydrohalo-
genation and similar mechanisms, and meth-
ods for decontaminating PCB laden sediments.
These studies require use of physical as
well as chemical procedures in their imple-
mentation. The analytical work involves
measurements relating to disapearance of
starting materials as well as product
identification in gaseous, liquid and solid
states.
466
-------
LABORATORY EVALUATION OF MODEL SLURRY WALLS
Richard M. McCandless and Andrew Bodocsi
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
The purposes of this multi-phase
project are to identify specific areas of
current slurry wall technology where basic
research and better substantiated methods
are needed to make recommendations, where
appropriate, and to initiate research on
several aspects of slurry wall design and
performance. Research in support of the
first two objectives was initiated in 1984.
The latter objective involves model slurry
wall and bentonite filter cake studies
using -an experimental test tank and column.
Results of initial testing using this
equipment are described.
The test tank is designed to permit
evaluation of soil-bentonite walls under
simulated field hydraulic and surcharge
loading conditions. The primary objectives
are to evaluate the effects of in situ
consolidation on the hydraulic conductivity
of soil-bentonite barriers and to determine
the potential for self-remediation of
"micro windows" formed by cracking, clay
flocculation, or entrapment of slurry
within the soil-bentonite backfill. Seven
long-term tests were performed using a
clean fine sand (+200 sieve size) as a
model In situ soil. The first two tests
employed a 2% soil-bentonite wall permeated
at two different hydraulic gradients.
Results were considered questionable due to
leakage across the top of the wall, but the
tests served to identify necessary
equipment and procedural modifications.
Excavation of the wall after the tests
revealed the presence of small pockets
of entrapped sand from the trench wall
within the soil-bentonite wall, despite
a high level of care exercised in
constructing the wall. A second series of
tests employed a 1% soil-bentonite wall
permeated at three different hydraulic
gradients under three surcharge (effective
-overburden stress) conditions. Results
describe the decrease in hydraulic
conductivity of the soil-bentonite wall due
to in situ consolidation as a function of
the combined effects of increasing
effective vertical overburden stress and
increasing hydraulic gradient involving
additional seepage-induced horizontal
consolidation. Also hydraulic conductivity
values were lower than but comparable with
those obtained on the same soils using
conventional rigid-wall permeameters.
These findings, combined with an analysis
of elastic deformation and consolidation
effects observed In the tank, together
demonstrate the sufficiency of current tank
design and test procedures, and suggest
that laboratory-scale hydraulic
conductivity testing of soil-bentonitss for
QA/QC may be appropriate.
The filter cake column is designed to
permit evaluation of slurry penetration,
filter cake formation and effectiveness,
and various techniques to enhance or
predict the impact of these parameters in
various soils. A total of nine tests were
completed in a washed fine sand (+200 sieve
size) and a washed medium sand (+40). Data
describing the initial depth of slurry •
penetration, internal filter cake zonation,
and variations in filter cake hydraulic
conductivity with depth were generated for
the two test sands.
In all cases, the filter cake
initially formed was eventually breached
via mechanical erosion apparently related
to-hardware ,configuration and elastic
deformation (compression or rebound) of
the sand medium in response to planned
changes in hydraulic gradient during the
test. Nevertheless, It was possible to
estimate "breakthrough" times for ideal
467
-------
soil-bentonite barriers based upon the hydraulic conditions the barrier in the +40
composite Initial hydraulic conductivity (medium) sand was two to three times more
of the filter cake developed in each sand. effective than a similar barrier in a
deposit of +200 (fine) sand.
Results suggest that under equivalent
468
-------
ENVIROMMENTAL EMERGENCY RESPONSE UNIT
TECHNICAL INFORMATION EXCHANGE (EERU-TIX)
Maxine Karesh and A. C. Gangadharan
Enviresponse, Inc.
Livingston, NJ 07039
Hugh Masters
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
Edison, NJ 08837
The Environmental Emergency Response
Unit Technical Information Exchange
(EERU-TIX) in Edison, MJ» is establishing
and will maintain a central depository for
specialized information on hazardous waste
site cleanup and emergency response
technology. Federal, state, and local
government groups, private industry,
academia, and the general public will have
access to the data as well as to visual
materials. Computerized reference,
referral, and literature searches, a file
of Superfund contacts, and Information on
appropriate books, journals, and technical
reports are available to research and
emergency response personnel. EERU-TIX,
administered by Enviresponse, Inc., the
current EERU contractor, works closely with
the EPA's Center for Environmental Research
Information to provide the latest in
hazardous material research.
In addition, EERU-TIX will operate and
maintain a computerized bulletin board to
provide field and research personnel with
response data from past EPA clean-up
activities. Two systems are currently
undergoing final testing. The Case
Histories File will contain completed
hazardous material responses that the user
may find similar to a new or proposed
response effort. The user will be able to
examine the nature of the site, the
weather, the materials released, the
actions taken, etc., and use this
information to help solve the new problem.
The Countermeasures Selection System is a
computer-parallel to the Manual of
Countermeasures for Hazardous Substances
Release. The System is designed to speed
the evaluation of the Manual's extensive
tables and step-by-step procedure for
identifying appropriate responses to a wide
variety of hazardous materials and will be
used before consulting the manual. An
on-site mobile treatment equipment
performance evaluation data base is planned
for future development.
469
-------
HYDROSTATIC TESTING OF FMLS FOR DETERMINING
CRITICAL FAILURE FACTORS
David W.R. Shultz, Richard E. DesChenes
Department of Geosciences
Southwest Research Institute
San Antonio, Texas 78284
and
Mary Ann Curran
Land Pollution Control Division
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The performance of flexible membrane
liners (FMLs) installed in landfills
or surface impoundments is influenced
by the quality and stability of the
supporting subgrade. Inadequate subgrade
design and preparation often result
in FML failure. Subgrade composition
and ground water conditions may also
be contributing factors. Rock fragments
in a soil subgrade may cause punctures
under hydrostatic loading from liquid
over the liner. Seasonal high water
table conditions may also contribute
to hydrostatic pressure on the bottom
of the FML.
Knowledge of the various reactions
between the FML and subgrade, and the
factors which influence these reactions
is useful for developing subgrade prepa-
ration and FML selection, design and
installation guidelines.
An FML testing program has been
established to determine the factors
which influence liner/subgrade inter-
actions. Specific project objectives
are tos
(1) Determine the puncturability
effects of fixed point loads in a sub-
grade under slowly increasing downward
water pressure when applied to three
FML materials;
(2) Determine the effects of
changing the loading rate on FML perfor-
mance ;
(3) Determine the change in effects
when the tests are performed at elevated
temperature;
(4) Evaluate the protective effects
of geotextiles under the FML specimens;
and
(5) Prepare an economic assessment
of these factors relative to selection
of FML materials.
The Southwest Research Institute
FML test facility will be used to perform
the required FML tests. The test fa-
cility includes 36 pressure vessels
capable of operating under various
hydrostatic pressures and temperatures.
Three FML materials have been selected
for testing. They are high density
polyethylene, polyvinyl chloride, and
chlorosulfonated polyethylene. Two
thicknesses for each material will
be used to evaluate the effect of thick-
ness. Tests will be performed concur-
rently at ambient (23°C) and elevated
(50°C) temperatures.
Each FML specimen will be placed
over a sand subgrade containing three
470
-------
loading points. Each specimen will
be tested over three different loading
point heights protruding above the
sand subgrade. This configuration
will allow determination of the influence
of loading point height on FML penetra-
ting susceptability. Test sequences
call for varying the pressure over
the FML specimens for specified periods
of time.
The beneficial effects of geotex-
tiles will be evaluated upon completion
of the above tests. Selected geotextiles
will be placed over the subgrade and
loading points, and the FML specimen
will then be placed over the geotextile.
New specimens of the FML material which
previously failed the testing will
be used. The same test sequence will
be followed. These tests will provide
the data to evaluate critical failure
factors and determine how they are
interrelated.
471
-------
DATA REQUIREMENTS FOR REMEDIAL ACTION TECHNOLOGY SCREENING
Paul J. Exner, Donald M. Dwight, Lisa L. Parrel!, Thomas
-------
DEVELOPMENT OF AN ENVIRONMENTAL TESTING CHAMBER
FOR SOIL REAGENT TESTING
John A. Glaser and Michael I. Black
U. S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio 45268
The study of chemical and biochemical
means to detoxify soil contaminated with
hazardous waste constituents is quite per-
plexing. Studies, to date, have taken
control methodology developed at the lab
bench directly to a field test setting.
Each field setting can be quite different
and has a multitude of possible variables
which can impinge on the optimum perform-
ance of a control technology. Clearly, an
intermediate stage of study or development
is necessary for more cost effective and
better designed field experimental studies.
Our research is directed to the development
of this intermediate testing stage and is
embodied in a device(s) that we choose to
call an "Environmental Testing Chamber."
The first prototype is based on the work of
Nash. We look forward to a continued
development of the chamber to incorporate
advanced ideas such as microclimate con-
trols so that this testing device can
closely mimic field testing conditions.
473
-------
PARAMETRIC STUDIES DELINEATING THE OCCURRENCE OF
TRANSIENT PUFFS IN A ROTARY KILN SIMULATOR
M. P. L1nak, J. D, Kilgroe, J. A. McSorley, J. 0. L, Wendt* and J, E. Dunn2
Air and Energy Engineering Research Laboratory
U. S, Environmental Protection Agency
Research Triangle Park, NC 27711
In the operation of practical rotary kiln Incinerators, the hazardous waste
charge Is often introduced In drums or containers in a batch mode. The ensuing
transient, caused by rapid devolatilizatlon of the waste, can lead to incomplete
destruction of the waste within the kiln, and to a momentary overcharging or
"puff" in the afterburner system. This presentation describes results of a
research program focused on the delineation of how waste properties and kiln
operating parameters govern the occurrence and magnitude of these puffs. The
objective 1s to delineate general engineering guidelines that allow waste proper-
ties and kiln operating parameters to be properly matched so that the likelihood
of transient overcharging problems is minimized. For example, for certain volatile
wastes, it may be beneficial to operate the kiln at lower, rather than at higher,
temperatures, in order to retard devolatilization of the waste and to thus enhance
mixing and oxidation within the kiln.
The research 1s conducted on a 250,000 Btu/hr (73 kW) rotary kiln simulator,
which is described in detail. This kiln simulates the complicated features of
full-scale units, but is still sufficiently well-defined to allow the important
fundamental phenomena governing the process to be identified, and specific operat-
ing parameters to be varied, one at a time.
1 University of Arizona, Tucson, AZ 85721
2 University of Arkansas, Fayettville, AR 72701
474
-------
CANINE OLFACTIONs -A NOVEL TOOL FOR INVESTIGATING
DECONTAMINATED EQUIPMENT & LEAKING UNDERGROUND STORAGE TANKS
Herbert S. Skovronek, Barbara Kebbekus, and Stewart Messur,
Hew Jersey Institute of Technology, Newark, NJ 07102
Lorenz D. Arner, Randolph, NJ 07869
Hugh Masters, US Environmental Protection Agency, Edison, NJ 08837
ABSTRACT
The feasibility of using trained dogs
to perform preliminary surveys for two types
of environmental hazards is being studied in
this project. Soil and groundwater pollu-
tion by leaking underground storage tanks
(LUST) is a growing problem, and the verifi-
cation of decontamination of equipment used
in spill cleanup is a task that may well be
simplified.
In the"decontamination* phase of the
project, the dogs are trained to respond to
certain chemicals, with xylene and 1,1,1-
trichloroethane being .chosen as the first
target compounds. Dog-proof devices con-
taining diffusion tubes with know emission
rates.at ambient temperature have been
developed for training. This allows the dog
to be trained on a higher emitter, with
successively lower-rate emitters being used
as the animal becomes more proficient. In
field tests, the dog is expected to find
plants of the target chemicals on equipment.
Air samples are taken in the vicinity of
the planted samples using Tenax adsorption
tubes. These samples are analyzed by high
resolution gas chromatography to determine
the concentration to which the dog is
responding. In actual use, the dog would
be used to screen the piece of equipment,
and the decontamination process would be
deemed complete if the dog detected no
contamination. In those cases where contam-
ination was found, further cleaning may be
required,-
To have the dog detect vapors arising
from leaking underground storage tanks and
distinguish such vapors from surface spills
is a more difficult task. A series of GC
analyses are performed to profile the compo-
sition of gasoline headspace and contrast
these with vapors that have percolated
through soil. This allows selection of a
compound or mixture o'f substances for the
dog to seek, and helps to differentiate the
vapors from the leaking tank from surface
spills. The dog may also be used to track
a leak back to its source, when conditions
are favorable.
During the project, the dogs are
regularly examined'by a veterinarian and,
while working with known chemicals, wear an
OSHA-type passive vapor sampling badge. The
levels to which the dogs and their handlers
are exposed should in no case exceed accept-
ed OSHA workplace standards, and are usually
very much lower.
The ability of dogs to detect various
odors is well known and has been used in
such varied applications as the detection
of contraband fruit, explosives, drugs,
leaks from pipelines, and termites. A
previous project (Arner, Skovronek, Masters)
has shown that the dogs can detect a chem-
ical such as toluene at low levels. This
project will delineate the usefulness as
well as the limitation of dogs for these
applications.
475
-------
DATA BASE FOR ESTIMATING HAZARDOUS WASTE VOC EMISSIONS
John M. Harden, Winton Kelly, Clinton BurkHn
Radian Corporation
Research Triangle Park, North Carolina 27709
ABSTRACT
The 1984 amendments to RCRA call
for EPA to promulgate regulations for
air emissions from hazardous waste
management. The Office of Air Quality
Planning and Standards (OAQPS) has
responsibility for development of these
regulations for all hazardous waste
management facilities except incinera-
tion and other combustion source
emissions.
This project - development of a data
base describing hazardous waste opera-
tions - supports OAQPS in its regulation
effortf the data base provides informa-
tion on the types of wastes, their
composition, their properties, their
quantities, and their treatment,
storage, and disposal operations. The
data base is a compilation of informa-
tion from several sources.
There were 5 specific objectives in
this program:
• Develop a comprehensive national
data base on the chemical properties of
all RCRA waste streams.
• Rank the wastes according to
their VOC potential (defined as the
product of waste volume x vapor
pressure).
• Estimate annual emissions for
each stream from each unit operation in
its management chain,
• Rank each stream by its
estimated annual VOC emission rate,
* Rank the various treatment,
storage, and disposal operations
according to their contribution to
annual VOC emissions.
A major advantage of this data
base is its breakdown of D and F series
wastes into more easily characterized"
subcategories. All D, F, K, P, and U
wastes listed in 1981 are accounted for
in this data base.
The Hazardous Waste VOC Data Base
is stored on diskette for the IBH-PC!M.
It is displayed on a LOTUS 1,2,3™
spreadsheet.
476
-------
APPLICATIONS 0-F CATALYTIC OXIDATION TO THE
• TREATMENT OF HAZARDOUS WASTE
Howard L. Greene
The University of Akron, Akron, Ohio 44325
Robert Mournighan
U.S.E.P.A., Cincinnati, Ohio 45268
ABSTRACT
Heterogeneous catalytic
oxidation of wastes such as
halogenated organic liquids
provides a potential method for
their effective disposal. Homo-
geneous thermal incineration
approaches to oxidation of these
stable materials at traditional
conditions means very high tem-
peratures, large residence times,
and almost no control-over pro-
duct selectivity. Catalytic
methods can avoid these short-
comings because of the potential
for totally different surface
reaction mechanisms.
Choice of catalyst supports,
catalytic agents (noble metal
versus transition metal oxide
types), and oxidizing conditions
are reviewed for the destruction
of chlorinated biphenyls and other
simpler halogenated molecules.
Results, including temperatures,
destruction efficiencies, resi-
dence times, and diffusion limi-
tations are discussed. Catalyst
selectivities for oxidation of
high-halogen content versus low-
halogen content organics are also
.critically compared. v .
477
-------
THERMAL DECOMPOSITION CHARACTERISTICS OF CHLORINATED METHANE MIXTURES
Philip H. Taylor, Douglas L. Hall, and Barry Del linger
University of Dayton
Research Institute
Dayton, Ohio 45469
ABSTRACT
The thermal decomposition behavior of
a mixture of chloromethanes has been anal-
yzed with relation to stability of the
initial component and the production of
thermal reaction products. Two mixtures
{Cl/H atomic ratios of 5.0 and 1.0} were
subjected to pyrolytic and oxidative (high
temperature degradation) at a mean gas-
phase residence time of 2.0 seconds. The
POHC stabilities were measured over a tem-
perature range of 300°C to 1050°C using a
quartz reactor coupled to a high resolution
gas chromatograph. Identifications of PICs
were deduced from an identical reactor
system coupled to a SC/MS analytical system.
In all four cases studied, 99 percent
destruction of a given POHC did not corre-
late with the lowest bond dissociation
energy. This finding is in agreement with
a previous thermal decomposition study of
the pure compounds in air and illustrates
the importance of complex reaction kinetics
in determining POHC stability. Results
indicate that decreasing oxygen concentra-
tion increased the stability of chloro-
methane; however, this effect was much less
pronounced for methylene chloride and
chloroform. As was the case for the other
components, the destruction efficiency of
carbon tetrachloride was independent of
Cl/H ratio under oxidative conditions.
However, under pyrolytic conditions, carbon
tetrachloride exhibited anomalous behavior
being thermally labile under low chlorine
conditions»
Identification of thermal reaction pro-
ducts under oxidative conditions are dom-
inated by chioroethy!ene compounds,
irrespective of Cl/H ratio. PIC identities
in the absence of air are strongly depen-
dent on the amount of chlorine present.
For low chlorine loadings, chloroacetyilene,
trichloroethylene, and tetrachloroethylene
are the dominant high temperature products.
For higher chlorine loadings, additional
high temperature products include 1,3-
hexachlorobutadiene and hexachlorobenzene.
On the basis of PIC formation.and POHC
stability as a function of reactor tempera-
ture and atmosphere, reaction pathways for
each component of the mixture have been
elucidated.
478
-------
WEATHERING EFFECTS ON EXPOSED SOIL LINERS
David C. Anderson, Janic F. Artiola, Walter W. Crawley
James A. Rehage, Ronald L. Shiver and Marvin Unger
K. W. Brown and Associates, Inc.
College Station, Texas 77840
ABSTRACT
Soil liners may be exposed to
weathering prior to being, covered with
other liners, leachate' collection systems,
and waste. Exposure to weathering can
degrade performance and prolonged exposure
can result in complete failure .(loss of
hydraulic integrity) of a soil liner.-
Mechanisms by which weathering may result
in failure include the following:
(a) Cracks formed due to drying
conditions;
(b) Cracks formed due to freeze-
thaw cycles; and
(c) Swelling and subsequent slough-
ing or dispersion due to flood-
ing. - .
Desiccation cracks in a.soil liner
are a potential problem wherever soil
liners are constructed. The problem is
most acute in areas where the evaporation
rate is- high .and the mineralogy of the .
soil clays is predominantly smectitic
(expanding lattice clays). Vertical
shrinkage cracks measuring over 3 centi-
meters wide and 2.5 meters deep have been
observed in a soil liner exposed to drying
conditions.
Freeze-thaw cycles may result in
degraded performance of a soil liner. For
instance, HI situ freezing of water can
result in soil fabric rearrangements.
There may also be a loss of liner strength
following the eventual thaw of the soil.
Under certain conditions, both vertical
and horizontal cracks may form in a soil
liner as the result of frost heaving.
Swelling due to flooding will tend to
decrease both strength and density of a
soil liner. Swelling has been observed to
increase the .hydraulic conductivity of
soil liners. Soil liners used in high
moisture situations (e..g., surface
impoundments) need to be chemically
modified or the soil appropriately selectr
ed to handle the excess water.
-•-)-! i ' '
Discussions with engineers experienced in
the construction of soil liners revealed
only one effective method for the repair
of soil liners damag'ed from exposure to
weather. The method included ripping out
the affected area, breaking down the size
of clods, bringing the soil back to the
desired water content and recompacting to
the density given in the design specifica-
tions.
The best approach is to prevent
weathering from taking place, which is
easily accomplished by covering the soil
liner immediately after construction.
If the final liner covering is not avail-
able at the time of liner completion, an
inexpensive (2 mil ) plastic sheet can
serve as a temporary cover. Protection
from freezing requires a liner covering
with some insulating properties.
479
-------
RESEARCH ACTIVITIES AT THE HAZARDOUS WASTE RESEARCH CENTER
-AN EPA CENTER OF EXCELLENCE-
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: environ-
mental media/waste interaction, inciner-
ation and alternate methods of treatment/
destruction. Although administered through
LSU's College of Engineering, individual
research are conducted by multidisciplinary
groups representing a number of academic
departments.
Two recent or ongoing research pro-
jects will be highlighted in the poster
presentation. These projects address
stabilization/solidification of hazardous
organic wastes and Industry Associates
Program for short term applied research
on hazardous waste treatment.
The results of three year's research
by M.E. Tittlebaum, F. Cartledge and H.C.
Eaton on stabilization/solidification will
be presented. Their research project,
"Identification of Bonding and Interfering
Mechanisms Associated with Stabilized/
Solidified Hazardous Organic Wastes," has
examined the microscopic distribution of
chemical elements and compounds in
selected organic hazardous waste/binder
chemical elements and compounds in
systems. The results of examinations
have shown that chemical interaction
projects occurs but that fixation as
defined by U.S. Environmental Agency
is not achieved.
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 findings will be applied
to current industrial problems in
hazardous waste treatment and disposal.
Short term applied research will be
conducted on generic problems faced by
companies who become members of the
program. The operation and progress
of this report will be presented.
480
-------
HAZARDOUS WASTE RESEARCH, DEVEI0PMENT, AND DEMONSTRATION PERMITS
Nancy Pomerleau
U.S. Environmental Protection Agency
tfeshington, DC
The Hazardous and Solid Waste .Amend-
ments of November 8, 1984, gives EPA a. new
permit authority to issue research, develop-
ment, and demonstration permits (RD&D).
The primary purpose of RD&D permits is to
aid in the development of safe alternatives
to land disposal of hazardous wastes by
expediting the permitting process to demon-
strate experimental and innovative tech-
nologies and processes.
In issuiJKpRD&D permits, EPA may
modify or waive the standard permit appli-
cation and permit issuance requirements
applicable to other hazardous waste permits,
except that financial responsibility
requirements and public participation pro-
cedures must be met. RD&D permits also
must include terms and conditions that
assure protection of human health and the
environment.
EPA will tailor the information re-.
quired in individual permit applications
to the nature and scope of the proposed
activity. The types of information appli-
cants should be prepared to submit are a
description of the proposed process or
technology, a list of the types and quan-
tities of hazardous wastes that will be
treated, a summary of the procedures that
will be used for responding to emergency
situations, a description of the general
qualifications of the staff that will
operate the activity, and a closure plan
and evidence of financial assurance. The
level of detail required from each appli-
cant will be determined by a combination
of site-specific factors such as facility
size, type and quantity of waste, duration
of the experimental activity, potential
for damage to health or the environment,
and staff experience and qualifications.
Applicants should submit proposals to
the EPA Regional Office that serves the
area,where the activity will occur. Because
each proposal is unique, EPA encourages
applicants to send a letter to the Regional
Office describing their proposal before
attempting to prepare a full application.
The Regional' Office will generally review
the proposal within 30 days of receipt to
tentatively determine if it is eligible
for an RD&D permit and to identify the
additional information the applicant needs
to submit for the Agency to process the
application. During this period, the "
Regional Office will also coordinate with
the State and determine if any State author-
ities may apply. The applicant should be
prepared to meet with EPA if necessary to
discuss the proposal.
EPA assigns a high priority to pro-
cessing RD&D permit applications and will
work closely with applicants to expedite
reviews. Based on limited experience to
date, EPA expects it will take a minimum
of six to nine months to issue each RD&D
permit. This includes the required 45
days for public comment.
To assist applicants, EPA has prepared
the Guidance Manual for Research, Develop-
ment, and Demonstration Permits. The
manual provides criteria'for determining
which proposals may qualify for an RD&D
permit, procedures for pre-application
reviews, descriptions of the types of
information that should be provided in an
RD&D permit application, and the types of
permit terms and conditions that will be
used in issuing RD&D permits. Copies of
the manual are available from the EPA
Regional Offices.
481
-------
CENTER HILL RESEARCH ACTIVITIES
Department of Civil and Environmental Engineering
University of Cincinnati "
Cincinnati, Ohio 45221
ABSTRACT
The Center Hill Facility is a U.S. EPA
Hazardous Waste Engineering Research
Laboratory research facility operated under
contract by the University of Cincinnati.
This facility provides technical support
for various Superfund and Resource
Conservation and Recovery Act (RCRA)
projects. Two RCRA projects that were
completed during FY'85 are the basis for
this poster.
DESIGN STUDY ON THE FEASIBILITY
OF UTILIZING AN ABOVE-GRADE LANDFILL
FOR THE DISPOSAL OF HAZARDOUS WASTES
Hark T. Bowers, Elizabeth Jones
Above-grade landfills built over sloped
double liners on a sawtoothed base offer
several advantages for the containment of
leachate from hazardous waste. These
advantages include continuous gravitational
leachate drainage, minimal reliance on
flexible membrane liner seams, easy
maintenance of leachate collection systems,
and monitoring of leachate volumes. The
state-of-the-art in above-grade landfill
technology as well as needed data and
designs have not been accumulated and
published, if indeed such have been
determined. This information is necessary
to assess the potential usefulness of
above-grade waste disposal.
This design study comprised three
phases of work as follows:
1. A 1iteraturereview assessing the
availability of sufficient information
to design, construct and operate an
above-grade landfill
2. A design phase in which all pertinent
components from the base through the
cover are designed in accordance with
1984 RCRA amendments, and
3. A perailt review phase in which the
design is submitted to regulatory
agencies for review and comment
The goal of this work assignment is to
determine what deficiencies exist in
current above-grade landfill technology, if
licensing agencies are prepared to assess
the technical soundness of designs for
above-grade landfills, and to determine
what experimental research should be
carried out to further the above-grade
disposal concept. The results of the study
not only provide advancement in the area of
above-grade hazardous waste disposal, but
provide important findings for the
performance of Subtitle D municipal solid
waste research as well.
SAS CHARACTERIZATION AND
MICROBIOLOGICAL ANALYSIS OF REFUSE
IN LANDFILL SIMULATORS
Riley Kinman, Janet Rickabaugh,
David Nutini, Martha Lambert,
This study describes the termination
of a five-year, pilot-scale experimental
landfill project. The five-year project
evaluated methane production and gas
enhancement techniques in sanitary
landfills through the addition of various
nutrients and moisture and the use of
temperature elevation.
The termination study which is
discussed here consisted of characteri-
zation of the trace volatile constituents
of the experimental landfill gas and
microbiological analysis of the refuse.
The trace volatile organic compounds were
found in higher concentrations than
482
-------
previously reported in the literature.
Xylenes, ethyl benzene, methylene chloride,
toluene and benzene were found in all of
the gas samples analyzed. The levels and
types of trace organlcs found in the gas
indicate landfill gas could be potentially
corrosive and may contain toxic levels of
some compounds.
All refuse samples had relatively high
aerobic and anaerobic plate counts,
Clostrldium perfrlngens and fungi levels.
These same samples indicated relatively low
levels of total col 1 forms, fecal coliforms,
fecal streptococci, and gram negative rods.
Relative numbers and types of micro-
organisms seemed to reflect the enhancement
techniques applied to the cell. For
example, microorganism levels were
generally lower in cells that received
leachate recycle as an enhancement tech-
nique. The highest level of microorganisms
was found in a cell which had a sewage
sludge enhancement.
433
-------
THE U.S. EPA COMBUSTION RESEARCH FACILITY
Larry R. Waterland, Robert W. Ross, II., Ralph H, Vocque,
Joannes W. Lee, and Thomas H. Backhouse
Aeurex Corporation
Combustion Research Facility
Jefferson, Arkansas 72079
ABSTRACT
During the past year, Incineration
testing at the EPA Combustion Research
Facility (CRF) has focused on evaluating
the incinerability of actual hazardous
waste. These tests have generally been
in support of EPA Regional Office remed-
iation efforts at priortiy Superfund,
sites. Specific wastes test-burned in
the rotary kiln incinerator included the
toluene stillbottoms waste from tri-
chlorophenol production at the "Vertac
Chemical Company, which contained about
40ppm 2,3,7,8-TCDD. In addition, a test
burn of PCB liquids (50 percent Askarel
in fuel oil) 'was also conducted in the
CRF liquid injection incinerator in
support of at-sea incineration sampling
protocol validation.
Detailed results from the Vertac
test will be' presented in this poster
presentation. Preliminary results of
other tests will be discussed as appro-
priate.
484
-------
AUTOMATED EXPERT ASSISTANCE FOR RCRA PERMIT REVIEWS/EVALUATION
OF REMEDIAL ALTERNATIVES FOR SUPERFUND SITES
Daniel G. Greathouse
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
It is estimated that during the next
3-7 years the U.S. Environmental Pro-
tection Agency and the State Permitting
agencies will have to make decisions con-
cerning permit applications for 770 land
disposal sites; 125 existing inciner-
ators; and 1800 other Hazardous Waste
Treatment, Storage, and Disposal Facili-
ties. Additionally, for a large number
of abandoned sites (Superfund) alterna-
tive remedial action alternatives will
have to be evaluated. Due to the unique
characteristics of the sites and the
inherent complexities and uncertainties
associated with most of these assessments,
considerable time and resources are
needed for each site evaluation. It has
been estimated that on the average the
cost of processing each permit for a
land disposal facility will cost fifty
thousand dollars. However, even at this
relatively large price there is no
guarantee that the evaluation will
reflect the latest available research
results or that there will be consistency
among reviewers (i.e. similar permits
will be comparably evaluated by different
reviewers). Because of the expected
heavy review workload and potential
difficulties, the Hazardous Waste
Engineering Research Laboratory is using
the principles of expert systems to
develop automated systems to assist in
these reviews.
Hazardous Waste Permit reviews and
Superfund evaluations have been
separated Into self .contained tasks or
components (flexible membrane liners,
covers, leachate collection systems,
closure plans, assessments of alternative
remedial actions). These components and
priorities for development of automated
systems have been formulated with input
from EPA Headquarters staff and permit
reviewers from the EPA regional offices
and states. Currently, prototyped systems
have been developed for evaluation of
chemical compatibility of PVC liners,
assessment of waste characteristics, and
evaluation of remedial action alternatives.
These efforts are being expanded and new
systems initiated during 1986. This paper
will present the overall development
plan for the automated systems, example
results of the prototype systems, and
status of on-going work.
485
-------
PREDICTION OF SOLUBILITY OF HAZARDOUS
MATERIALS IN WATER
Frank R. Groves, Raj Doshi
Chemical Engineering Department
Louisiana State University
Baton Rouge, Louisiana 70803
ABSTRACT
The transport and fate of spilled
hazardous materials in the environment is
a major concern for environmental
protection. For example, when organic
materials leak from an underground
storage tank they may contact ground
water and dissolve. Modelling of this
transport requires basic thermodynamic
dota including solubility. This paper
describes research on predicting aqueous
solubility of sparingly soluble organics
including effects of temperature and
salinity. The present study is limited
to single components in water but the
ultimate objective is to deal with multi-
component solutes.
The solubility of a sparingly
soluble material is related thermo-
dynamically ia a simple way to the
activity coefficient of the solute in the
solution. This research explores two
methods of predicting the activity
coefficient: the UNIFAC group contri-
bution method and the solubility
parameter method.
Conventional UNIFAC group inter-
action parameters are based on data for a
wide range of solutes and solvents, most
of them exhibiting complete miscibilitiy.
In this research revised group inter-
action parameters were fitted .to data
limited to aqueous solubilities of
sparingly soluble organics. The revised
parameters give better results for
solubility predictions. Temperature
dependence was determined for the revised
parameters. The temperature dependent
parameters successfully predicted the
effect of temperature on solubility of
paraffins and alkylbenzenes.
The solubility parameter method was
used successfully to correlate activity
coefficient and solubilities for alkyl-
benzenes and paraffins. The data were
fitted using a single polar solubility
parameter for water plus interaction
parameters for various classes of
compounds. For alkylbenzenes in sea
water it was found that a slight decrease
in the interaction parameter permitted
solubility in sea water to be predicted
from fresh water solubility.
The methods investigated here should
be useful for solubility prediction in
more complex systems. In particular,
both the UNIFAC method and the solubility
parameter approach are readily general-
ized to multicomponent solutes.
486
-------
REMOVAL AND RECOVERY OF HAZARDOUS ORGANIC VAPORS BY
ADSORPTION ONTO ACTIVATED CARBON AND XAD4 RESIN
Kenneth E, Noll
Industrial Waste Elimination Research Center
Illinois Institute of Technology
Chicago, IL 60616
ABSTRACT
A comparative study has been conducted that at low concentrations activated
on adsorption/desorption of six hazardous carbon adsorbed more organic vapor
organic vapors on synthetic resin (XAD4) than synthetic resin. At higher,
and activated carbon, using a differential industrial concentrations, the
reactor involving the expansion of a quartz resins adsorbed more vapor. The
spring. While both sorbents can effective^ resin also showed much higher desorp-
ly remove the organic vapors, it was observed tion.
487
-------
QUALITY ASSURANCE/QUALITY CONTROL SUPPORT DURING SHAKEDOWN
AND DEMONSTRATION OF EMERGENCY RESPONSE EQUIPMENT
Uwe Frank and Michael Gruenfeld
US Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, New Jersey 08837
John D. Koppen
Princeton Aqua Science
Edison, New Jersey 08837
The Comprehensive Environmental
Response, Compensation, and Liability Act
of 1980 (Superfund) required the develop-
ment of countermeasures (mechanical de-
vices, and other physical, chemical, and
biological agents) to clean-up inactive
hazardous waste disposal sites and hazard-
ous substances released from these sites.
As a consequence, the Releases Control
Branch (RCB) of the Hazardous Waste Engi-
neering Research Laboratory (HWERL) is
engaged in the shakedown and demonstration
of several major pieces of equipment, such
as: a Mobile Incineration System designed
for the thermal destruction of hazardous
substances in an environmentally safe and
controlled manner; a Mobile Carbon Regen-
erator System designed to purify spent
carbon in the field for reuse during
responses to hazardous substance emergen-
cies; a Mobile Soils Washing System de-
signed to extract spilled hazardous ma-
terials from soil; and a Mobile In-Situ
Treatment Unit designed to inject chemical
agents into the ground to detoxify conta-
minated soil.
The shakedown and field testing of
this equipment depends on an extensive
number of analytical measurements to
provide information on their operation,
efficiency, reliability and safety. This
poster describes the quality assurance/
quality control (QA/QC) procedures that
are used to ensure that the analytical
data that is generated on a daily basis is
of known and adequate quality to support
technical decisions that need to be made
for proper equipment operation. This.
QA/QC support activity complements the
overall quality assurance program of HWERL
which is directed by the Laboratory's
Quality Assurance Officer.
Routine QA/QC support activities are
being performed by Enviresponse, Inc.,
under contract to RCB, and include the
following: reviewing QA/QC procedures to
assure compliance with approved quality
assurance project plans; reviewing sample
collection and analytical methods; moni-
toring analytical performance through the
use of split and replicate samples; main-
taining and performance checking analyti-
cal instruments to assure reliable results;
developing and evaluating rapid "referee"
methodologies to confirm analysis results;
and preparing periodic progress reports
showing an assessment of the technical
soundness, statistical accuracy, and pro-
per documentation of analytical informa-
tion and data associated with equipment
operation, shakedown, and field demonstra-
tion.
488
-------
DETERMINATION OF EFFECTIVE POROSITY OF SOIL MATERIALS
Robert Horton and M. L, Thompson
Iowa State University
Ames, Iowa 50011
ABSTRACT
Hazardous waste disposal landfills
require liners constructed of compacted
soil material to help prevent the migra-
tion of hazardous wastes from the landfill.
The performance of a compacted soil liner
is partly a function of the porosity.
Porosity is important because the trans-
port of materials through the liner will
occur via the nonsolid or the pore space.
The main purpose of this project is to
study the pore spaces of compacted soil
materials, and to estimate the effective
porosity which is the portion of the pore
space where the most rapid mass transport
occurs.
The pore space of three soil
materials, till, loess, and paleosol, is
studied using mercury intrusion porosi-
metry, water desorption, and image
analysis. These analyses provide cumula-
tive porosity curves which lead to
estimates of pore size distribution.
Theory is developed to estimate the
effective porosity of a compacted soil •
material based upon its pore size distri-
bution. The effective porosities of the
compacted till, loess, and paleosol
materials are estimated to be 0.04, 0.08,
and 0.09, respectively. These effective
porosities are between 10 arid 201 of the
total porosities.
Comparisons between measured and
predicted Cl" travel times through com-
pacted soil samples are made in order to
verify the estimated effective porosities.
The estimated effective porosities are
reasonable because predicted Cl~ first
breakthrough times are similar to the
measured first breakthrough times in com-
pacted till, loess, and palesol materials.
For the three soil materials used in this
laboratory study, predicted first break-
through times were 5 to 10 times earlier
when using effective porosity in the
estimates as compared to estimates based
solely on soil permeability. This result
implies that effective porosity may be an
important parameter to use when charac-
terizing field sites.
489
-------
HAZARDOUS WASTE RESEARCH AT THE USEPA TEST AND EVALUATION FACILITY
Douglas W. Grosse
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
The U.S. Environmental Protection
Agency operates a Test and Evaluation
(T&E) Facility which houses a variety
of bench- and pilot-scale research exper-
iments supporting wastewater and hazardous
waste research programs in the Office of
Research and Development. With an in-
creasing emphasis on hazardous waste
treatment research, several projects have
been initiated at the T&E Facility. The
Treatment Technology Staff of the Thermal
Destruction Branch, Alternative Technology
Division (ATD), is involved with existing,
state-of-the-art treatment technologies
which can be applied as an alternative to
land disposal of hazardous residues. The
Chemical and Biological Detoxification
Branch, ATD, is responsible for the
development and/or testing of innovative
technologies with novel approaches applied
to hazardous waste detoxification. The
Containment Branch of the Land Pollution
Control Division is involved with research
on solid and hazardous waste land disposal,
including clean-up practices for uncon-
trolled hazardous waste sites. This
poster session describes the participating
research groups, facility layout and
services, current and proposed research
projects, and relevant data on results
that can be derived from the present
research program.
490
-------
EVALUATION OF CUTOFF WALL CONTAINMENT EFFICIENCY USING AQUIFER
STRESS ANALYSES - GILSON ROAD NPL SITE
NASHUA, NEW HAMPSHIRE
Matthew J. Barvenik, David M. Brown and Thomas J. Kern
Goldberg-Zoino & Associates, Inc.
320 Needham Street
Newton Upper Falls, Massachusetts 02164
Michael A. Sills
New Hampshire Water Supply and Pollution Control Commission
Concord, New Hampshire
ABSTRACT
Soil/bentonite backfilled cutoff walls are being used as passive physical barriers to the
movement of groundwater containing hazardous wastes. Although many cutoff walls have
been successfully installed over the past forty years for construction dewatering, their
effectiveness for containing hazardous chemicals has not been adequately evaluated in the
field. Because these barriers typically have been constructed under performance
specifications, it is important to be able to differentiate between seepage through the
wall and that through the underlying aquitard for construction contract enforcement. The
effectiveness of a soil/bentonite cutoff wall keyed to fractured bedrock has been
evaluated for the Gilson Road NPL site in Nashua, New Hampshire. Closed-form analysis of
a bedrock pumping test and a finite difference numerical sensitivity analysis of cutoff
wall conductivity were performed. The results show that the bulk hydraulic conductivity
of the cutoff wall is several orders of magnitude lower than that of the fractured
bedrock. Based on these analyses, the efficiency of the cutoff wall was determined to be
greater than 93 percent.
IHTROBOCTION
Soil/bentonite cutoff walls have been
used successfully as an adjunct to pumping
for construction dewatering and general
groundwater control for over forty years.
More recently, soil/bentonite (S/B) cutoff
technology has been applied to the more
critical application of containing
hazardous wastes at US EPA National
Priority List (NPL) sites, as well as
private industrial sites.
As is often the case with initial
technology transfer, these cutoff wall
containments were designed and specified
by professionals associated with the
emerging hazardous waste field, rather
than that of cutoff wall design/
construction. These professionals,
although familiar with the hydrogeologic
constraints of the application, generally
lack expertise and experience in
cutoff-wall- specific technology. As
a result, specifications were strictly
performance based, thus allowing
contractors to execute the work in a
manner consistent with past dewatering
practices. Unfortunately, fluid flux
rates which were insignificant for
dewatering purposes are now unacceptable
with respect to hazardous waste
containment. As a result, these
containments have not performed up to
expectation in some cases. The reaction
of the regulatory community has been
negative in response to these failures.
As part of the remediation of the
Gilson Road Site (NPL No. 23) in Nashua,
New Hampshire, a S/B cutoff wall was
employed to physically isolate a
contaminant pluma from adjacent surface
waters, locally occurring high
491
-------
permeability eskers, and the rest of the
aquifer in general. The cutoff wall was
4,000 feet in length and up to 110 feet
deep. The objectives of this barrier were
to allow cost efficient hydrodynamic
isolation and groundwater
reoirculation/treatment within the
containment for aquifer restoration. In
recognition of the importance of the
cutoff wall to the successful operation of
the hydrodynamic isolation/reciroulation
system, a rigorous design phase was
implemented including an advanced S/B
backfill design methodology (Schulze,
1981) and development of new equipment and
procedures for quality control field
testing (Barvenik, 1985). This work was
incorporated into the specifications as
minimum design standards keyed to specific
aspects of the actual construction process
(Ayres, 1983). The Gilson Road project
therefore represents perhaps the first
case where the level of cutoff' wall
design, specification, and construction
quality control (Q/C) were commensurate
with application to hazardous waste
containment.
The research described herein
documents the hydraulic effectiveness
achieved with a cutoff wall constructed in
1982 under state-of-the-art design and Q/C
guidance. This work represents the final
portion of a three phase research and
development contract funded as a
cooperative agreement between the Hew
Hampshire Water Supply and Pollution
Control Commission (NHWSPCC) and the US
EPA. The first two phases consisted of an
evaluation of the precision and accuracy
of the new Q/C testing developed for and
used during construction of the Gilson
Road wall, and assessment of the
suitability of electrical piezocone
instruments for post-construction anomaly
deteotion (GZA, 1986).
Site Description and Geology
The Gilson Road uncontrolled hazardous
waste site was the subject of the first
cooperative agreement signed under EPA's
Superfund program. Clandestine dumping of
toxic volatile chemicals into an abandoned
gravel pit resulted in a contaminant plume
which severely degraded residential air
quality as it discharged into a local
stream and also threatened downstream
municipal drinking water supplies. A more
detailed site history was presented
earlier (Ayres, 1983).
The site is underlain by primarily
granular ice-contact and glacial lake
deposits. These consist of fine to coarse
sands and gravels, with layers and lenses
of silts and silty fine sands. Below the
stratified drift is a granular basal till.-
This, in turn, is underlain by a
moderately fractured schist bedrock, with
a weathered, highly fractured upper zone.
For additional geologic description see
Koteff, 1973.
The sand and sand/gravel deposits
exhibit an average hydraulic conductivity
of 10~2 cm/sec based on pumping test
results. However, these deposits vary
over a range from 10"^ cm/sec for the
silts and silty fine sands to an estimated
10~^ cm/sec for open work gravel seams
found in the esker deposits. The glacial
till deposits are typically less permeable
(10~5 cm/sec) than the average values for
the overlying stratified drift. However,
the till is generally thin and
discontinuous in nature, thus constituting
a relatively imperfect aquitard. The
hydraulic conductivity of the upper 50
feet of the bedrock was estimated to range
between 10~5 and 10~3 cm/sec, based on
investigation phase packer testing. The
bedrock pumping test executed as part of
this post-construction R&D project,
however, indicated that the upper 10 to 15
feet is shattered, with a conductivity of
over 10~1 cm/sec. This condition exists
at least locally, in the vicinity of the
downgradient wall.
Remedial Measures
Final remediation for the site
consisted of complementary active and
passive physical barriers to flow, and
treatment of the captured contaminant
plume. This interactive system was
adopted because neither a purely passive
barrier nor a purely active barrier alone
could eliminate off-site containment
migration. A cutoff wall down to the top
of bedrock and a synthetic cap form the
passive elements of the system. Extending
the passive barrier through the fractured
portion of the bedrock was evaluated
during initial design phases. This could
be accomplished by excavating through the
bedrock or grouting the rook fractures.
Albeit feasible, neither method was found
to be a cost-efficient solution for
492
-------
limiting contaminant flux through the
bottom of the containment. The active
component of the total containment vessel
is a hydrodynamics isolation system which
compensates for the inadequacies of the
bedrock as an aquiclude. This system
redistributes the groundwater within the
containment and eliminates head
differences across the cutoff wall to
achieve a no flow condition (zero
gradient). The groundwater
extraction/recharge associated with the
hydrodynamic isolation also allows for
treatment of the recirculated
water, resulting in aquifer restoration.
The restoration process requires a number
of flushing cycles for complete
contaminant removal.
A purely active system of plume
control via hydrodynamic
extraction/recharge was found to be
unacceptable during the feasibility study
due to the heterogeneity of the aquifer.
Groundwater extraction at, a rate to,
eliminate plume migration would
necessarily entail pumpage of significant
volumes of clean water from regions of
high conductivity outside the plume and
nearby surface waters. This dilution of
the plume at the point of recharge and the
concomitant spreading quickly results in a
contaminant plume width greater than the
extraction system's capture zone.
Even in a homogeneous aquifer,
contaminants can still escape active
containment if the recharge point does not
lie directly upstream of the extraction
well (i.e., along a flow line). Finally,
even with a perfect recirculation system,
contaminant dispersion across the flow
lines at the edge of the containment
results in a. further loss of contaminant
mass from the system.
Although the cutoff wall provides
enhanced hydraulic control, no practical
hydrodynamic isolation system,
incorporating a finite number of
extraction and recharge points, can
completely eliminate all gradients across
the physical barrier. Therefore, to
preclude off-site contaminant migration,
the system is operated with a conservative
bias by forcing an inward gradient
everywhere along the cutoff wall. The
required net recharge deficit, relative to
pumpage, necessitates an external purge
stream. The purge stream volume is, in
part, a function of the conductivity of
the containment perimeter (cutoff wall).
To attain an extremely high water quality
in the purge stream, as mandated for
off-site disposal at this CERCLA site,
this portion of the pumpage must undergo
additional treatment steps as compared to
the total recirculation flow. Therefore,
the hydraulic efficiency of the cutoff
wall directly affects purge stream
treatment costs, which are a high
percentage of the two million dollar/year
overall operation and maintenance cost of
the on-site treatment plant.
Objectives
The objective of the post-construction
verification study described hereafter was
to hydraulically evaluate cutoff wall
performance over a large scale. Hydraulic
evaluation required two separate steps:
1. a bedrock pumping test to determine
the transmissivity of the fractured
.bedrock which forms the containment
bottom, and
2. a numerical sensitivity analysis of
wall conductivity on vertical head
distributions to assess cutoff
efficiency.
The ability to better determine the
site-specific conductivity of the bedrock
and cutoff wall allows more accurate
calibration of the three-dimensional
numerical model used to design and operate
the hydrodynamic isolation/treatment
system. This results in increased
operational cost efficiencies via
reduction in the required volume of the
highly treated purge stream.
BEDROCK PUMPING TEST
The pumping well was located at the
downgradient end of the site, adjacent to
the inside face of the cutoff wall. The
open section of the well was completed in
the bedrock and isolated from the
overburden above. The bedrock pumping
test was conducted in September, 1985, and
ran for ten days at a rate of 88 gallons
per minute (gpm). Drawdowns were
monitored at over seventy single- and
multi-level installations in the bedrock
and overburden, both inside and outside of
the cutoff wall. The locations of the
pumping well and monitoring installations
493
-------
are shown in Figure 1.
Results
Contours of equal drawdown in the
bedrock at the end of the test are shown
in Figure 2. While the top of rock varies
considerably in elevation throughout the
site, it was considered to be one
continuous hydrologie unit for analysis
due to the presence of the less conductive
glacial till above. This simplification
was justified based on the magnitude and
distribution of the drawdowns measured. A
maximum study-state drawdown of less than
3 fset at a pumping rate of almost 90 gpm
succinctly demonstrated the large
transmissivity of the bedrock aquifer.
The drawdown contours were also notable in
that they are uncharacteristically
cirularfor fractured media. This
indicates that the rock is isotropic in
the horizontal, which is indicative of a
high degree of fracturing, and thus a high
transmissivity.
Contours of equal drawdown in the
overburden, or stratified drift, are shown
in Figure 3. Drawdowns were slightly
larger outside the cutoff wall than inside.
This was due to mounding from the recharge
trench, which was felt more strongly
CUTOFF WALL
+-
n
+
h-
RECHARGE
^-"TRENCHES
•-V
0' 50' 100 200'
FIGURE 1: PUMP TEST NETWORK
494.
-------
-CUTOFF WALL
RECHARGE
TRENCHES
•*••••••.»••••*
0' 50' 100'" ~200'
• MEASURED •••••••ANALYTICAL
FIGURE 2: BEDROCK DRAWDOWN CONTOURS - MEASURED VS. ANALYTICAL
account for the low conductivity barrier
within the overburden. Therefore, the
pumping well was located as close as
possible to the cutoff wall such "that the
wall would lie along the anticipated
radial flow paths.
The maximum drawdown in the overburden
was approximately 0.6 feet. While this
drawdown is small compared to the
thickness of the phreatic aquifer,
enabling linearization of the flow
equation, the drawdown is significant
withrespect to fluid flux. Inasmuch as
the phreatic aquifer cannot be treated as
the constant head source bed in a leaky
confined aquifer analysis, a multi-aquifer
analysis, which considers drawdown in all
layers, was used.
The time-dependent drawdown in two
aquifers separated by an aquitard, due to
pumping in one aquifer, was derived by
Hantush (1967), For large time, this
solution diverges due to the lack of a
source of water other than depletion of
storage. At the site, however, the water
pumped from the lower aquifer was
recharged to the upper aquifer. To
account for this, two of Hantush's
transient solutions were superimposed.
Details of the calculation are given in
495
-------
JL_
0' 5d iOO' 200'
MEASURED ****** ANALYTICAL
FIGURE 3: OVERBURDEN DRAWDOWN CONTOURS - MEASURED VS. ANALYTICAL
GZA, 1986. It suffices here to mention
that the divergent terms in the two
transient solutions differ in sign, and
that, in the limit of large time, the sum
of the two is finite. Therefore, the
solution is one of steady state, which is
reasonable because a mathematical source
of water now exists.
The analytical results for each aquifer
were contoured and compared visually to
the contours of measured drawdown. The
parameters which gave the best match are
70 cm2/sec (6,500 ft2/d) for the bedrock
transmlssivity, 38 cn2/sec (3,500 ft2/d)
for the overburden transmisslvity,
and H x 10~6 sec'1 (0.33 d1"1) for the
leakanoe of the glacial till. (The
leakance of an aquitard may be defined as
the flow per unit head difference across
the unit's thickness, and is thus equal to
the stratum's vertical conductivity
divided by the thickness). The drawdowns
calculated using these parameters, as
compared to the actual drawdowns, are
shown in Figures 2 and 3 for the bedrock
and overburden, respectively.
Discussion
The transmissivities and leakance thus
calculated-are effective homogeneous
496
-------
values. As such, they can only represent
a kind of mean behavior of the system.
Heteogeneities (such as the cutoff wall)
and boundary conditions (such as Lyle Reed
brook) are, in a sense, lumped or averaged
into the properties of each hydraulic unit.
The analytically derived heads for the
overburden do not capture the variability
seen in the heads actually measured. The
calculated transmissivity for the
overburden shows this effect. The derived
transmissivity is low compared to
estimates based on previous overburden
pumping tests in this area of the site.
Therefore, its calculated transmissivity
is useful only for lumping the behavior of
the stratified drift as it affects the
glacial till and fractured bedrock. The
leakance found for the glacial till
aquitard, however, agrees well with visual
classifications made during the
exploration program. In the immediate
vicinity of the pumping well, the till is
approximately 15 feet thick, thus implying
a vertical conductivity of 10~3 cm/s.
This value is higher than is probable for
the intact glacial till and therefore
elucidates the discontinuous nature of the
stratum.
Fractured rock formations do not
typically exhibit porous medium, or
continuum, behavior (Brown and Gelhar,
1986). However, at this site, the low
magnitude and high radial symmetry of the
drawdowns in the bedrock aquifer support
the representation of the fractured rock
as an equivalent porous medium. Drilling
and water yield data obtained during
installation of the pumping well and
monitoring points indicate that the
bedrock transmissivity is concentrated in
its upper 10 to 15 feet. As such, the
conductivity of the upper bedrock is over
10"^ cm/sec. During previous attempts to
calibrate the finite difference model of
the site, it was necessary to use bedrock
conductivities much larger than had been
indicated by packer testing. The pumping
test has demonstrated that the model was
essentially correct. In fact, the test
has justified incorporation of even larger
bedrock conductivities, thus resulting in
a more accurate calibration.
CUTOFF WALL
EVALUATION
BULK CONDUCTIVITY
During Phase I of the R&D project, the
intact hydraulic conductivity of the wall
was evaluated. Initially, Q/C samples of
the backfill were obtained during
construction and tested in the laboratory.
This work was followed with post-
construction undisturbed sampling/testing
of the in-place wall two years after
installation. Both data sets indicated
that the intact conductivity of the cutoff
wall met the design specification of 1 x
10-7 cm/sec (GZA, 1986). Knowledge of the
intact backfill conductivity, however,
does not allow direct evaluation of
overall wall performance. This requires
assessment of the bulk conductivity, which
includes potential imperfections in the
wall such as windows, cracks, and
inadequacies from keying to bedrock.
Phase II research was directed at
development of a modified electrical
piezocone probe and its use for anomaly
detection in the completed wall. These
data indicated that the wall is generally
intact (GZA, 1986). However, the bulk
conductivity of the containment can only
be verified via hydraulic performance
evaluation of large sections of the wall
subjected to hydraulic stress conditions.
A three-dimensional numerical model is
required to correlate the head data
generated during such a stress test and
thus arrive at an evaluation of cutoff
wall performance. This is particularly
true when the head measurements reflect
the geologic heterogenity inherent in most
actual containment applications.
Model Description and Calibration
The three-dimensional finite
difference flow model of the site .was
originally developed as part of the
design, operation and monitoring effort
for the permanent hydrodynamio isolation
system. The code used was written by
Trescott (1975), as modified by Torak
(1982). The present application utilizes
a W8 x 50 grid to cover the 425-acre
modeled area, with five layers in the
vertical. Each of the 12,000 nodes was
assigned parameter values determined from
over 150 borings and 2,000 grain size
samples, numerous borehole permeability
and bedrock packer testa, and overburden
pumping tests executed during the RI/FS
phases. Overburden transmissivity values
for the downgradient portion of the
containment were further refined based on
head data developed during operation of
the emergency hydrodynamic isolation
497
-------
system. This temporary active barrier was
implemented to limit plume migration prior
to cutoff wall construction and
incorporated five pumping wells and a 300-
foot-long recharge trench located 400 feet
upgradient. Bedrock conductivities were
also refined based on the bedrock pumping
test described herein. The model was
subsequently calibrated generally to
within 1/2-foot for four different system
stress states:
1. a period of medium recharge before
the cutoff wall was installed
(January,1982),
2. a period of high recharge after
installation of the cap and wall but
before commencement of hydrodynamie
isolation (June, 1983),
3. a subsequent period of low recharge
without hydrodynamie isolation
(October, 1983), and
4. a period of low recharge with the
hydrodynamie isolation system in
operation (September, 1985).
The last period represents the initial
state of the site just prior to the
bedrock pumping test.
The runs were performed in steady
state. The transient nature of
precipitation variation was accounted for
with steady state approximations. For
example, in the run for June, 1983, an
areal recharge of %2 in/yr was used. This
would be high if it were applied for an
entire year. It may, however, be
considered as the difference between the
actual recharge rate at the time and any
water entering storage associated with
increases in water level. During model
oalibration, a cutoff wall conductivity of
1 x 10~7 cm/sec was used. This value was
selected based on the previously discussed
measurements of intact conductivity.
Detailed descriptions of boundary
conditions and hydrologic stresses for the
model are given in GZA, 1986.
Sensitivity Analysis
A sensitivity analysis was performed
to determine if the intact conductivity of
1 x 10~f cm/sec was also a valid
representation of the bulk wall
conductivity, thus indicating a high
degree of wall integrity. This analysis
consisted of artifically increasing the
cutoff conductivity above the intact value
and observing the change in the vertical
head distributions generated by the model.
The higher conductivities represented the
effects of imperfections in the otherwise
intact cutoff wall. The simulated
vertical head distributions were then
compared to the actual field heads and
conclusions drawn with respect to the
plausibility of the artifically higher
conductivity value. This procedure was
executed incrementally until the mis-match
between the simulated and actual head
distributions precluded the possibility of
a wall bulk conductivity of the magnitude
input for that run. An upper bound for
the cutoff wall bulk conductivity was thus
established as an indication of barrier
integrity.
For the purpose of assessing cutoff
wall bulk conductivity, the wall must be
hydraulically stressed. Ideally, the head
difference across the wall would be at
least ten to twenty times the accuracy of
the model (0.25 feet in the vicinity of
the wall). A data set obtained on a
routine basis prior to this research
(June, 1983) was therefore selected. This
data set provided the greatest measured
head drop across the wall (over 3 feet).
A more recent data set, obtained
specifically for the the sensitivity
analysis, would have been preferable
because it would have included the
additional monitoring points installed for
the bedrock pumping test. However, the
hydrodynamie isolation system was already
operational prior to the installation of
the new monitoring points. Therefore, the
head differences across the wall at that
time were too small to be used for the
sensitivity analysis.
Figure 4 shows the m'easured
piezometric heads versus depth and the
modeled heads at the homologous nodes for
two multi-level installations, M-10R and
M-14. Flow is occurring downward at M-14
(inside the containment) and upward at
M-10R (outside the containment). The
model was run with the wall conductivity
at 10, 100, 1000, and 10000 times the
previously verified intact conductivity of
1 x 10"' cm/sec. The results of the
analysis are shown in Figure 4. For the
first increase (10~^ cm/sec), there is
almost no effect on the heads. With the
498
-------
76
P1EZOMETRIC HEAD (ft.)
77 78
79
78
P1EZOMETRIC HEAD (ft.)
79 80 81
•-30
SIMULATED BULK
CONDUCTIVITY
I07, I06 CM/SEC
IOS CM/SEC
104 CM/SEC
VERTICAL HEAD
DISTRIBUTION
SIMULATED BULK
CONDUCTIVITY
|63 CM/SEC
MEASURED, 6/86
VERTICAL HEAD
DISTRIBUTION
**.**•***•****
FIGURE 4: MODELED VERTICAL HEAD DISTRIBUTIONS FOR DIFFERENT WALL CONDUCTIVITIES
conductivity at 10~-* cm/sec, there is a
noticeable effect, but the change is
still within the error band of the model.
However,- at 10"^ cm/sec, the change
exceeds the error band. Therefore, it is
bulk, large-scale hydraulic conductivity
of the cutoff wall is 10~5 cm/sec. Please
note that this value is not an estimate of
the actual bulk conductivity. Rather, it
is a value below which the actual cutoff
wall bulk conductivity must, fall.
Discussion
The leakance of the glacial till
aquitard is moderately high, and thus the
bedrock and cutoff wall behave analogous
to electrical resistors in parallel. The
bedrock transmissivity is so high relative
to the transmissivity of the cutoff wall,
-that it dominates, the flow, which then
controls the vertical and horizontal
piezometrie head gradients. This effect
can be seen in the fluid fluxes out of the
containment by node type as shown in Table
1 for-, .the range of-wall conductivities
used in the sensitivity analysis.
As the conductivity of the wall is
slowly increased from 10~? cm/sec, the
bedrock flow still dominates the heads,
499
-------
TABLE 1
MODELED FLUXES OUT OF CONTAINMENT
Flux by Node Type (gpd)
Hall K (em/sec)
__7
icr6
lo-s
10-4
10-3
Rock
73,668
73,287
69,056
50,764
35,351
Wall
93
830
8,361
45,036
90,814
Hote; Average overburden conductivity is
10~z cm/aec.
and thus the wall flux increases linearly
with its conductivity. As the wall flux
eventually becomes significant with
respect to the bedrock flux, a
concomitant decrease in the overall
gradient occurs as reflected in the
vertical head distribution. If the
bedrock had a lower transmissivity, then
this critical point would occur at a
lower wall conductivity. An upper bound
value could then be established which
would allow a meaningful evaluation of
cutoff wall bulk conductivity relative to
the design value of 1 x 10"^ cm/sec.
However, the high bedrock transmissivity
at the site precludes a determination of
wall conductivity below the upper bound
value of 10~5 cm/sec. This limitation
applies to any strictly hydraulic
analysis.
CUTOFF WALL EFFICIENCY
To verify the contractor's ability to
adhere to a performance based
specification, evaluation of the actual
bulk hydraulic conductivity of the cutoff
wall is required. As discussed
previously, this can only be accomplished
by hydraulically stressing the in-place
containment wall. This procedure is
subject to the limitations imposed by the
in-aitu hydrology. For the Gilson Road
site, the high bedrock conductivity,
relative to the cutoff wall design
specification, precludes the usefulness of
such an evaluation for construction
contract enforcement purposes. This
limitation, inherent in the procedure for
evaluation of cutoff wall work executed
under performance specification, serves to
underscore the need for Q/C testing during
construction (Barvenik, 1984).
With respect to the more important
issue of containment effectiveness,
however, it suffices to evaluate the
cutoff wall based on its ability to
preclude contaminant flow through the
overburden. Hence, cutoff wall efficiency
can be defined as the flux from the site
through the overburden which was
eliminated by the wall divided by the flux
which would have occurred without the wall.
The flux value without the wall was
estimated by running the June, 1983
scenario and substituting native
permeabilities into the wall nodes. The
computed flux was approximately 122,000
gpd. Inasmuch as an upper bound value for
bulk cutoff wall conductivity is 10~5
cm/sec, then the wall flux is at most
8,000 gpd (see Table 1). Therefore, a
lower bound on the cutoff wall's
efficiency is 93 percent.
CONCLUSIONS
Analysis of a pumping test conducted
in the fractured bedrock at the Gilson
Road NPL site resulted in estimates of the
transmissivity of the bedrock and
overburden and the leakance of the
intervening glacial till aquitard. The
transmissivity of the bedrock is estimated
to be 70 cm2/sec (6,500 ft2/d), and the
leakance of the till is estimated to be
4 x 10~6 sec"1 (0.33 d1"1). These values
agree well with descriptive field data.
The effective homogeneous transmissivity
of the overburden was estimated to be 38
em2/sec (3,500 ft2/d), a value reflecting
an average of the cutoff wall and native
materials. It is therefore useful only
for lumping the behavior of the stratified
drift as it affects the other
geohydrologic units.
A sensitivity analysis of cutoff wall
bulk hydraulic conductivity on vertical
distributions of piezometric head was
performed using a three-dimensional finite
difference flow model. Given the large
bedrock transmissivity and the practical
limits on the calibration of the model, it
was possible only to determine an upper
bound for the conductivity of the wall.
This upper bound is approximately 10~5
cm/sec. The actual bulk conductivity of
500
-------
the cutoff wall is probably lower, but
could not be more specifically determined
via hydraulic stress analysis of the
containment. Therefore, it is felt that
for construction contract oversight and
enforcement, the post-cons truet ion
numerical modeling approach for estimating
cutoff wall bulk conductivity may not be
appropriate for all sites. Modeling may be
more appropriate for estimating the
overall efficiency of such containment
systems, however, and is essential for
establishing withdrawal, recirculation,
and recharge locations within the
containment to optimize treatment of the
captured groundwater.
Although the actual value of the
cutoff wall bulk conductivity could not be
determined via hydraulic stress analysis,
the efficiency of the wall with respect to
containment could be evaluated.
Containment efficiency can be defined as
the ratio of flux through the overburden
which is eliminated by the passive barrier
to the flux which would have occurred
without the barrier. Based on the results
of the sensitivity analysis, the
efficiency of the cutoff wall at the
Gilson Road site was found to be greater
than 93 percent. This is based solely on
c he upper bound value for bulk
conductivity established by the
sensitivity analysis and as such, is a
worst case condition. The actual cutoff
wall efficiency is likely to be higher as
based on the Q/C data amassed during
construction as well as the other
post-construction verification studies.
ACKNOWLEDGEMENTS
The research and development work
presented herein was associated with the
Gilson Road NPL site and as such was
sponsored under a cooperative agreement
between the State of New Hampshire Water
Supply and Pollution Control Commission
and the 0.S. EPA. The authors would_
therefore like to express appreciation to
Michael Donahue, Paul Heirtzler, and Ralph
Wickson of the State of New Hampshire and
Stephen James, Walter Grube, and Jon
Herrmann of the U.S. EPA for their
continuing cooperation and assistance. In
addition, the authors would like to
acknowledge Scott Perkins, without whose
dedicated field involvement this paper
would not have been possible.
REFERENCES
Ayres, J.E., D.C. Lager, and M.J.
Barvenik, "The First EPA Super fund Cutoff
Wall: Design and Specifications,"
presented at the Third National Symposium
on Aquifer Restoration and Groundwater
Monitoring, Columbus, Ohio, 1983.
Barvenik, M. J,, W. E. Hadge, aad D. T.
Goldberg, "Quality Control of Hydraulic
Conductivity and Bentonite Content During
Soil/Bentonite Cutoff Wall Construction,"
Eleventh Annual Research Symposium, EPA,
Cincinnati, Ohio, April, 1985.
Brown, D. M., and L. W. Gelhar, "The
Hydrology of Fractured Rocks: A
Literature Review", MIT Parsons Laboratory
Technical Report, 1986.
GZA, "Cutoff Wall Testing Program, Gilson
Road Site, Nashua, New Hampshire."
Prepared for the USEPA Office of Research
and Development, Hazardous Waste
Engineering Research Laboratory,
Cincinnati, Ohio, 1986.
Hantush, M. S., "Flow to Wells in Aquifers
Separated by a Semipervious Layer,"
Journal of Geophysical Research, 72(6),
pp. 1709 - 1720, 1967.
Koteff, Carl, and Richard P. Volchman,
"Surficial Geologic Map of the Pepperell
Quadrangle," USGS, GP-1118, 1973.
Schulze, D., M. J. Barvenik, and J. E.
Ayres, "Design of the Soil/Bentonite
Backfill Mix for the First EPA Superfund
Cutoff Wall," Proceedings of the 4th
National Symposium and Exposition on
Aquifer Restoration and Groundwater
Monitoring, Sponsored by NWWA, Columbus,
Ohio, May, 1984.
Trescott, P. C., "Documentation of
Finite-Difference Model for Simulation of
Three-Dimensional Groundwater Flow," USGS
Open File Report 75-438, 1975.
Torak, L. J., "Modifications and
Corrections to the Finite-Difference Model
for Simulation of Three-Dimensional
Groundwater Flow," USGS Water Resources
Investigation 82-4025, 1982.
501
U.S. GOV1BHMEHT PRISTISO OFtICE:1986~ 6H6-116/ •» 0 6 6 3
-------
------- |