o1
United States Hazardous Waste Engineering EPA/600/9-85/013
Environmental Protection Research Laboratory April 1985
Agency Cincinnati, OH 45268
Research and Development
&ERA Lancj Disposal of
Hazardous Waste
Proceedings of the
Eleventh Annual
Research Symposium
Do not remove. This document
should be retained in the EPA
Region 5 Library Collection.
-------
EPA/600/9-85/013
April 1985
LAND DISPOSAL OF HAZARDOUS WASTE
Proceedings of the Eleventh Annual Research Symposium
at Cincinnati, Ohio, April 29-May 1, 1985
Sponsored by the U.S. EPA, Office of Research & Development
Hazardous Waste Engineering Research Laboratory
Land Pollution Control Division
Containment Branch
and
Alternative Technologies Division
Thermal Destruction Branch
Coordinated by:
JACA Corp.
Fort Washington, Pennsylvania 19034
Contract No. 68-03-3131
Project Officer
Naomi P. Barkley
Cincinnai, Ohio 45268
HAZARDOUS WASTE ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
U.S. Environment?.! Protection Agency
Region v, ! "'"••<:'•'
230 Soul-, L •;;•.>:,c>n 'trect
Chicago, Illinois 60604
-------
DISCLAIMER
These proceedings have been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for
use.
-11-
-------
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
solid and hazardous wastes. These materials, if improperly dealt with, can
threaten both public health and the environment. Abandoned waste sites and
accidental releases of toxic and hazardous substances to the environment also
have important environmental and public health implications. The Hazardous
Waste Engineering Research Laboratory assists in providing an authoritative
and defensible engineering basis for assessing and solving these problems.
Its products support the policies, programs, and regulations of the Environ-
mental Protection Agency, the permitting and other responsibilities of State
and local governments, and the needs of both large and small businesses in
handling their wastes responsibly and economically.
These Proceedings present the results of completed and ongoing research
projects concerning the land disposal of hazardous waste.
David 6. Stephan, 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 and
treatment of hazardous waste. These projects are funded by the U.S. Environ-
mental 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.
-iv-
-------
ABSTRACT
The Eleventh Annual Research Symposium on Land Disposal, Remedial
Action, Incineration and Treatment of Hazardous Waste was held in Cincinnati,
Ohio, April 29-May 1, 1985. The purpose of the symposium was to present to
persons concerned with hazardous waste management the latest significant
findings of ongoing and recently completed research projects funded by the
Hazardous Waste Engineering Research Laboratory's Land Pollution Control
Division and Alternative Technologies Division.
This volume is a compilation of speakers' papers and poster presenters'
abstracts for Sessions A and C concerning hazardous waste land disposal.
Areas covered are: 1) Remedial Action, 2) Pollutant Assessment, and 3)
Pollution Control. Subjects include landfill design and operation, waste
leaching and analyses, pollutant migration and control, waste modification,
surface impoundments, technology assessments, remedial action techniques, and
cost/economics.
This document covers hazardous waste land disposal only. A separate
document for Session B, Hazardous Waste Incineration and Treatment, will be
published by the Hazardous Waste Engineering Research Laboratory.
-v-
-------
CONTENTS
Page
SESSION A - HAZARDOUS WASTE LAND DISPOSAL
REMEDIAL ACTION
Technical Resource Documents for Hazardous Wastes
Norbert B. Schomaker, U.S. Environmental Protection Agency 1
Evaluation of Chemical Grout Injection Techniques for Hazardous
Waste Containment
James H. May, Robert J. Larson, Philip G. Mai one, and
John A. Boa, Jr.
USAE Waterways Experiment Station 8
Permeable Materials for the Removal of Pollutants from Hazardous
Waste Leachates
James E. Park, University of Cincinnati 19
Mechanisms of Containment Migration Through a Clay Barrier - Case
Study, Wilsonville, Illinois
R.A. Griffin, B.L. Herzog, T.M. Johnson, W.J. Morse, R.E. Hughes,
S.F.J. Chou, and L.R. Follmer, Illinois State Geological Survey . . 27
Reclamation and Redevelopment of Uncontrolled Hazardous Waste Sites
G.L. Kingsbury and R.M. Ray, Research Triangle Institute 39
Remedial Action Modeling of the Western Processing and Occidental -
Lathrop Sites
David B. Watson and Stuart M. Brown, CH2M-Hill
Frederick W. Bond, Battelle Pacific Northwest Laboratory 49
Demonstration of a Geographic Information System for Hazardous Waste
Site Analysis
Margrit von Braun, University of Idaho. ... 57
Quality Control of Hydraulic Conductivity and Bentonite Content
During Soil/Bentonite Cutoff Wall Construction
M.J. Barvenik, W.E. Hadge, and D.T. Goldberg
Goldberg, Zoino and Associates, Inc 66
Control of Fugitive Dust Emissions at Hazardous Waste Cleanup Sites
Keith D. Rosbury, PEI Associates, Inc.
Stephen C. James, U.S. Environmental Protection Agency 80
Practical Methods for Decontaminating Buildings, Structures, and
Equipment at Superfund Sites
M.P. Esposito, J.L. McCardle, A.H. Crone, and
J.S. Greber, PEI Associates, Inc.
R. Clark, S. Brown, J.B. Hallowell, A. Langham, and
C.D. McCandlish, Battelle Columbus Laboratories 88
-vn-
-------
In Situ Treatment of Contaminated Groundwater and Soils, Kelly
Air Force Base, Texas
Roger S. Wetzel, Susan M. Henry, and Phillip A. Spooner
JRB Associates
Stephen C. James, U.S. Environmental Protection Agency
Edward Heyse, U.S. Air Force Engineering and Services Laboratory. . 97
Case Study Investigations of Remedial Response Programs at
Uncontrolled Hazardous Waste Sites
Claudia Furman, JRB Associates
S. Robert Cochran, Earth Technology 107
Effect of Freezing on the Level of Contaminants in Uncontrolled
Hazardous Waste Sites - Part I: Literature Review and Concepts
I.K. Iskandar
U.S. Army Cold Regions Research and Engineering Laboratory
J.M. Houthoofd, U.S. Environmental Protection Agency . 122
Operation of the Center Hill Pilot Plant and Soils Laboratory for
Superfund and RCRA Research Projects
Joseph Burkhart, U.S. Environmental Protection Agency
Gerard Roberto, University of Cincinnati 130
Reclamation and Reuse of Contaminated Land in the United States,
England, and Wales
Tayler Bingham and Carrie L. Kingsbury
Research Triangle Institute 131
Computer Analysis of Hazardous Waste Dikes
Andrew Bodocsi and Richard McCandless, University of Cincinnati . . 132
Evaluation of Stabilized Dioxin Contaminated Soils
William H. Vick and William D. Ellis
Science Applications International Corp.
Donald E. Sanning and Edward J. Opatken
U.S. Environmental Protection Agency 133
Two Automated Systems for On-Scene Coordinators: The Counter-measure
Selection System and the Case History File of Responses to Hazardous
Substances Releases
R.F. Hildreth and S. Fullerton
Science Applications International Corp.
R.A. Griffiths, U.S. Environmental Protection Agency 134
POLLUTANT ASSESSMENT
Volatile Organic Emissions Reductions from Surface Impoundments by
the Use of Wind Fences and Wind Barriers
Louis J. Thibodeaux, Louisiana State University
Rebecca Parker and Charles Springer, University of Arkansas .... 136
-viii-
-------
Investigation of Floating Immiscible Liquids to Control Volatile
Organic Chemical Emissions from Surface Impoundments
K.T. Valsaraj, Charles Springer, and Trac Nguyen
University of Arkansas
Louis J. Thibodeaux, Louisiana State University 145
Leaching Potential of 2,3,7,8-TCDD in Contaminated Soils
Danny R. Jackson, Henry M. Grotta, Steve W. Rust, J. Scott
Warner, Mickey F. Arthur, and Fred L. DeRoos
Battelle-Columbus Laboratories
Mike H. Roulier, U.S. Environmental Protection Agency 153
Development of Standardized Batch Adsorption Procedures:
Experimental Considerations
W.R. Roy, C.C. Ainsworth, S.F.J. Chou, R.A. Griffin, and
I.G. Krapac, Illinois State Geological Survey 169
Clay Liner Systems: Current Design and Construction Practices
L.J. Goldman, R.S. Truesdale, and C.M. Northeim
Research Triangle Institute 179
Effects of Inorganic Leachates on Clay Soil Permeability
J. Jeffrey Peirce and Thomas A. Peel, Duke University 182
Effective Porosity of Geologic Materials
James P. Gibb, Michael J. Barcelona, Joseph D. Ritchey and
Mary H. Lafaivre, Illinois State Water Survey 190
Solubility Parameters for Predicting Membrane-Waste Liquid
Compatability
Henry E. Haxo, Jr., Nancy A. Nelson, and Jelmer A. Miedema
Matrecon, Inc 198
Analysis of Leachate Production in Closed Hazardous Waste Landfills
Randy R. Kirkham, Scott W. Tyler, and Glendon W. Gee
Pacific Northwest Laboratory 213
Precision and Reliability of Laboratory Permeability Measurements
John L. Bryant and Andrew Bodocsi, University of Cincinnati .... 225
Simulating Landfill Cover Subsidence
Harry J. Sterling, Daniel International Corp.
Michael C. Ronayne, ATEC Associates, Inc 236
Leachate Collection Systems - Summarizing the State-of-the-Art
Jeffrey M. Bass, Arthur D. Little, Inc 245
Verification of Leachate Collection Efficiency with Physical Models
Paul R. Schroeder, USAE Waterways Experiment Station 252
-IX-
-------
Page
Chemical Resistance of Flexible Membrane Liners Immersed in Selected
Aqueous Solutions
Gordon Bellen and Rebecca Corry, National Sanitation Foundation . . 261
Assessment of Synthetic Membrane Successes and Failures at Waste
Storage and Disosal Sites
Jeffrey M. Bass, Warren J. Lyman, and Joseph P. Tratnyek
Arthur D. Little, Inc 262
Relationship of Laboratory and Field Determined Hydraulic
Conductivity in Compacted Clay Soils
A.S. Rogowski, USDA-ARS, Northeast Watershed Research Center. . . . 263
Critical Review and Summary of Gas and Leachate Production from
Municipal Solid Waste Landfills
Frederick G. Pohland and Stephen R. Harper
Georgia Institute of Technology 264
Hydraulic Conductivity of Two Prototype Clay Liners
Steven R. Day, Geo Con, Inc.
David E. Daniel, University of Texas at Austin 265
Clay-Chemical Compatibility and Permeability Testing: A Review
G.L. Kingsbury and R.S. Truesdale, Research Triangle Institute. . . 266
Clay Liner Transmit Time Prediction Methods and Effects of Leachate
Composition on Liner Infiltration
Ashok S. Damle and Scott M. Harkins, Research Triangle Institute. . 268
Effective Porosity of Compacted Clay Soils Permeated with Organic
Chemicals
J.W. Green, Arthur D. Little, Inc.
K.W. Brown and J.C. Thomas, Texas A&M University 270
Influence of Concentrations of Organic Chemicals on the Colloidal
Structure and Hydraulic Conductivity of Clay Soils
K.W. Brown and J.C. Thomas, Texas A&M University 272
Laboratory Determination of Dielectric Constant and Surface Tension
as Measures of Leachate/Liner Compatibility
C.I. Mashni, H.P. Warner, and W.E. Grube, Jr.
U.S. Environmental Protection Agency 273
Quick Indicator Tests to Characterize Bentonite Type
Andrew Bodocsi and Richard M. McCandless
University of Cincinnati 274
Estimating Transit Times of Noninteracting Pollutants Through
Compacted Soil Materials
Robert Horton, Michael L. Thompson, and John F. McBride
Iowa State University 275
-x-
-------
Page
POLLUTANT CONTROL
Waste-Liner Compatibility Studies
Warren J. Lyman, Arthur D. Schwope, Joseph P. Tratnyek and
Peter P. Costas, Arthur D. Little, Inc 283
Recovery and Testing of a Synthetic Liner from a Waste Lagoon After
Long-Term Exposure
Nancy A. Nelson and Henry E. Haxo, Jr., Matrecon, Inc.
Peter McGlew, U.S. Environmental Protection Agency 296
Strength and Durability of Flexible Membrane Liner Seams After
Short-Term Exposure to Selected Chemical Solutions
Mary Ann Curran, U.S. Environmental Protection Agency
Ronald K. Frobel, U.S. Bureau of Reclamation 307
Chemical Mass Transport Measurements to Determine Flexible Membrane
Liner Lifetime
Arthur E. Lord, Jr., Robert M. Koerner, and Eric C. Lindhult
Drexel University 313
Development of a Technique for Retrofitting Impoundments with
Geomembranes
David W. Schultz, Daren L. Laine, and John W. Cooper
Southwest Research Institute 321
The Role of Impoundment Size in Controlling Environmental Risk at
Hazardous Waste Surface Impoundments
Gordon B. Evans, Jr., Jey K. Jeyapalan, Joe Sai, Ronald Shiver,
and David C. Anderson, K.W. Brown and Associates, Inc 332
Update: Storage of Hazardous Waste in Mined Space
Ronald B. Stone, Fenix & Scisson, Inc 339
Factors Affecting Stabilization/Solidification of Hazardous Waste
Jerry N. Jones, R. Mark Bricka, Tommy E. Meyers, and
Douglas W. Thompson, USAE Waterways Experiment Station 353
Construction Quality Assurance for Hazardous Waste Disposal
Facilities
Coleen M. Northeim and R.S. Truesdale
Research Triangle Institute
Alessi D. Otte, U.S. Environmental Protection Agency 360
Field Verification of Landfill Cover System Construction to Provide
Hydrologic Isolation
Richard C. Warner, Nathaniel Peters, and James E. Wilson
University of Kentucky 369
-XT-
-------
Page
Maintenance Free Vegetative Systems for Landfill Covers
John C. Rodgers, Fairley Barnes, Teralene Foxx, and Gail Tierney
Los Alamos National Laboratory 370
An Electrical Technique for Detecting Leaks in Membrane Liners
David W.R. Schultz and Bob M. Duff, Southwest Research Institute. . 371
-xn -
-------
TECHNICAL RESOURCE DOCUMENTS FOR HAZARDOUS WASTES
Norbert B. Schomaker
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Environmental Protection Agency is preparing documents to assist permit
applicants and permit writers responsible for hazardous waste landfills, surface
impoundments, land treatment facilities and waste piles. These Technical Resource
Documents (TRDs) describe current technologies and methods for evaluating the performance
of the applicant's design in the area of liners, leachate management systems, covers, and
other aspects of the containment facilities. These documents are basically compilations
of research efforts of the Land Pollution Control Division (LPCD) to date in these areas
and are developed and published to assist in the implementation of the RCRA regulations
concerning hazardous waste disposal facilities. The TRDs support RCRA Technical Guidance
Documents prepared by EPA's Office of Solid Waste (OSW). The Guidance Documents provide
details of design and operating procedures that contribute to the satisfactory
performance of a facility.
INTRODUCTION
Land disposal of hazardous waste is
subject to the requirements of subtitle C
of the Resource Conservation and Recovery
Act (RCRA) of 1976. 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. In reviewing and evaluating
the permit application for acceptance or
rejection, the permit official must make
all decisions in a well defined and well
documented manner. The TRD's are being
developed and published to assist the
permit applicants and permit review
official 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 review and evaluation based
on good engineering practices. There may
be alternative and equivalent methods for
conducting the review and evaluation.
However, if the results of these methods
differ from those of the Environmental
Protection Agency's method, they may have
to be validated by the applicant.
Eight TRDs have been completed to
date. Six of the TRDs are related to
landfills, and one document each is
related to surface impoundments and land
treatment. A listing of these documents
is shown below, along with a brief
description, publication number, and the
contact person's name in parentheses.
TRD 1: EVALUATING COVER SYSTEMS FOR
SOLID AND HAZARDOUS WASTE (SW-867) - This
document presents a procedure for
evaluating cover systems for solid and
hazardous waste. It contains a
checklist, with supporting documentation,
-------
that a permit writer can use to evaluate
proposed cover systems.
*055-000-00228-2 (R. E. Landreth)
TRD 2: HYDROLOGIC SIMULATION ON
SOLID WASTE DISPOSAL SITES (SW-868) -
This document provides a computer package
to aid planners and designers by
simulating hydrologic characteristics of
landfill covers to predict percolation as
a function of cover design and climate.
This document has been updated to include
the two-dimensional aspects of landfill
cover systems. It is now identified as
the "Hydrologic Evaluation of Landfill
Performance (HELP) Model" SW-84-009 and
SW-84-010. These documents have been
published for public comment. The HELP
model document will replace the SW-868
document. *055-000-00225-8 (D. C. Ammon)
TRD 3: LANDFILL AND SURFACE
IMPOUNDMENT PERFORMANCE EVALUATION
(SW-869) - The evaluation of leachate
collection systems using compacted clay
or synthetic liners to determine how much
leachate will be collected and how much
will seep through the liner into
underlying soils is presented. The
adequacy of sand and gravel drain layers,
slope, and pipe spacing is also covered.
*055-000-00233-9 (M. H. Roulier)
TRD 4: LINING OF WASTE IMPOUNDMENT
AND DISPOSAL FACILITIES (SW-870) -
Information and guidance on liner systems
is provided. The document discusses
waste, liner types, compatibility, liner
selection, specifications, design of
leachate collection systems, and case
study analysis methodology. It also
includes a glossary of liner system
related terms.
*055-000-00231-2 (R. E. Landreth)
TRD 5: MANAGEMENT OF HAZARDOUS WASTE
LEACHATE (SW-871) - This document
discusses leachate composition, leachate
generation, selected management options,
available treatment technologies,
*These documents have been published and
the reports are available from GPO by
requesting the stock number. Copies can
be obtained for a price from the:
Superintendent of Documents
U. S. Government Printing Office
Washington, DC 20402
(202) 782-3238
unit-treatment options, and system
economics. It presents management
options that a permit writer or hazardous
waste landfill operator may consider in
controlling and treating leachate.
*055-000-00224-0 (S. C. James)
TRD 6: GUIDE TO THE DISPOSAL OF
CHEMICALLY STABILIZED & SOLIDIFIED WASTES
(SW-872) - Basic information on
stabilization/solidification of
industrial wastes to insure safe burial
of wastes containing harmful materials is
presented. A summary of major physical
and chemical properties of treated wastes
is presented along with a listing of
major suppliers of stabilization/
solidification technology and a summary
of each process.
*055-000-00226-6 (R. E. Landreth)
TRD 7: CLOSURE OF HAZARDOUS WASTE
SURFACE IMPOUNDMENTS (SW-873) - The
methods, tests, and procedures involved
in closing a surface impoundment are
discussed and referenced. Problems
related to abandoned impoundments 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.
*055-000-00227-4 (M. H. Roulier)
TRD 8: HAZARDOUS WASTE LAND
TREATMENT (SW-874) - Land treatment
medium, characteristics of hazardous
waste, allowable waste loadings, facility
planning for land treatment, land
treatment permit processes, closure and
items to be considered during permit
evaluation are discussed.
*055-000-00232-1 (C. C. Wiles)
Each of the completed TRDs is
described in greater detail in the
following section of this paper.
TECHNICAL RESOURCE DOCUMENT SUMMARIES
Evaluating Cover Systems for Solid
and Hazardous Waste (SW-867)
A critical part of the sequence of
designing, constructing, and maintaining
an effective cover over solid and
hazardous waste is the evaluation of
engineering plans. This TRD presents a
procedure for evaluating closure covers
on solid and hazardous wastes. All
-------
aspects of covers are addressed in detail
to allow for a complete evaluation of the
entire cover system.
Covers may be evaluated for
conformance to requirements for closure
covers by the following eleven (11)
general procedures in sequence:
1) examine soil test data,
2) examine topography,
3) examine climate data,
4) evaluate composition,
5) evaluate thickness,
6) evaluate placement,
7) evaluate configuration,
8) evaluate drainage,
9) evaluate vegetation,
10) evaluate post-closure plan, and
11) evaluate contingencies plan.
This document describes current
technology for landfill covers in three
broad areas: Data Examination,
Evaluation Steps and Post-Closure Plan.
The data examination discusses test data
review procedure, topographical data
review and climatological data review
procedures. The evaluation steps include
cover composition, thickness, placement,
configuration, drainage and vegetation.
The post-closure aspects include
maintenance and contingency plan
evaluation procedure. 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.
Hydrologic Simulation on Solid Waste
Disposal Sites (Skl-868)
Percolation and runoff of
precipitation at solid waste disposal
sites create a potential for ground and
surface water contamination with
leachate. The proper landfill site
design and operational approach is to
minimize or eliminate percolation through
the solid waste and thus control the
formation of leachate.
The purpose of this TRD is to provide
an interactive computer program for
simulating the hydrologic characteristics
of a solid and hazardous waste disposal
site operation. A large number of cities
within the U.S. for which five years of
climatic records exist have been put on
tape for easy access and can be used in
lieu of on-site measurements. In
addition, the model stores many default
values of parameter estimates which can
be used when measurements and existing
data files are net available. The user
must supply the geographic location, site
area and hydrologic length, the
characteristics of the final soil and
vegetative cover, and default overrides
where deemed necessary. From minimal
input data, the model can simulate daily,
monthly, and annual runoff, deep
percolation, temperature, soil water, and
evapotranspiration.
The model, which is a modification of
the Soil Conservation Service (SCS)
number method for predicting runoff and
the hydrologic portion of the U.S.
Department of Agriculture, Science and
Education Administration (USDA-SEA)
hydrologic model entitled "Chemicals,
Runoff, and Erosion from Agricultural
Management Systems (CREAMS)," has been
modified to conform to the design
characteristics of solid and hazardous
waste disposal sites. The model takes
engineering, hydrologic, and climatologic
input data in the form of rainfall,
average temperature, solar radiation, and
leaf area indices and characteristics of
cover material and performs a sequential
analysis to derive a water budget
including runoff, percolation, and
evapotranspiration.
The user can specify up to three soil
layers and may also specify a membrane
liner at the base of the cover. The
decreasing effectiveness of the liner is
simulated. The model is designed for use
in a conversational manner, that is, the
user interacts directly with the program
and receives output immediately. No
prior experience with computer
programming is required. All commands
necessary to use the model are presented
in the user's manual. The model can also
be run in batch mode, which requires some
computer programming experience.
Landfill and Surface Impoundment
Performance Evaluation (SH-869)
This TRD describes how to evaluate
the capability of various liner and drain
designs to control leachate release. The
author has allowed for the widely varied
technical backgrounds of his intended
audience by presenting, in full, the
-------
rigorous mathematics involved in reaching
his final equations. Thus, any evaluator
can take full advantage of the manual up
to the level of his own mathematical
proficiency.
This evaluation system considers the
transport of liquids through simple
modular configurations. The important
parts of these modules are identified as
the Liquid Transport Control (LTC)
system, and the Diversion Interface
(DI). The mathematics consider the
efficiency with which the LTC and DI
components convert leachate flow from
vertical to horizontal. The extent of
ponding on a bottom liner is considered
in relation to the position and
effectiveness of the drainage system.
The procedures described in the
document should allow an evaluator to
determine the adequacy of design for:
1) compacted clay liners or
synthetic liners intended to
impede the vertical flow of
fluids,
2) sand or gravel drainage layers
intended to convey liquids
laterally into collection
systems,
3) slopes on such liner systems, and
4) spacings of collection drains.
The evaluation procedure is
independent of the structure of the
regulation it supports. If regulations
specify acceptable rates of release for
leachate, or efficiencies of leachate
collection systems, calculation results
may be used directly to determine if the
landfill is acceptable. However, if the
regulations describe acceptable
environmental impact, results from the
evaluation procedure can be used as input
for further evaluation in other manuals
to determine if the impact of the
landfill is acceptable.
Lining of Waste Impoundment and
Disposal Facilities (SH-870)
This document provides information on
performance, selection, and installation
of specific liners and cover materials
for specific disposal situations, based
upon the current state-of-the-art of
liner technology and other pertinent
technologies. It contains descriptions
of wastes and their effects on linings, a
full description of various natural and
artificial liners; service life and
failure mechanisms; installation problems
and requirements of liner types; costs of
liners and installation; and tests that
are necessary for pre-installation and
monitoring surveys.
Described in the report are
potentially useful lining materials and
their associated technologies, including
remolded and compacted soils and clays,
admixes, polymeric membrane liners,
sprayed-on liners, soil sealants, and
chemisorptive liners. The specific
characteristics of these liners in
service environments with various types
of wastes are discussed, particularly
with respect to compatibility with
wastes, permeability to water and waste,
failure mechanisms, and service life.
Design and construction of waste
facilities with various types of linings
are described, emphasizing installation
of membrane liners and their problems.
Special problems associated with the
design and construction of solid waste
disposal sites are considered. The
subjects of leachate collection above the
liner, gas control, and top-seals are
addressed.
Several processes for selecting a
specific liner or various satisfactory
"liners for a given disposal site are
proposed. These include the use of
compatibility tests, moderate duration
exposure tests, soil conditions tests,
costs, etc. Methods for the preparation
of specifications covering the selected
liners or groups of liners are suggested.
In the appendices data are presented
on a soil classification system; liner
industry organization; selected liner
test methods; liner installation;
leachate collection; soil liner design;
case study analysis and specification for
liner membranes. As new technology is
developed, and as experience in the use
of liners becomes available, the manual
will be upgraded on a chapter-by-chapter
basis.
Management of Hazardous Waste
Leachate (SM-871)
This document has been prepared to
-------
provide guidance for permit officials and
disposal site operators on available
management options for controlling,
treating, and disposing of hazardous
waste leachates. It discusses
considerations 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
pre-treatment of the wastes.
The manual addresses areas that must
be considered in evaluating a leachate
management option: leachate generation
and composition, and available
technologies for treatment of leachate
and hazardous waste.
Leachate composition is examined for
several hazardous waste landfills,
synthetic leachates from lab studies, and
sanitary landfill leachates. The
expected composition of the leachate and
the concentrations of the pollutants in
the leachate are presented along with
possible guidelines for permit writers to
use in determining the relative hazards
of a particular leachate.
A key section enumerates factors
which influence treatment process
selections and provides a suggested
approach for systematically addressing
each. Selected hypothetical and actual
leachate situations are used as examples
for applying the approach to the
selection of appropriate treatment-
processes.
Other sections address monitoring,
safety, contingency plans/emergency
provisions, equipment redundancy/backup,
permits, and surface runoff.
Guide to the Disposal of Chemically
Stabilized and Solidified Waste"
(Ski-87 2)
Stabilization/solidification of
industrial waste is a pretreatment
process designed to insure safe disposal
of wastes containing harmful materials.
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
document addresses the treatment of
hazardous waste for disposal or long term
storage and surveys the current state and
effectiveness of waste treatment
technology. A summary of the major
physical and chemical properties of
treated wastes is presented. A listing
of major suppliers of
stabilization/solidification technology
and a summary of each process is included.
This guide provides the background
information needed for waste generators
and regulatory officials to determine the
testing program and/or product
information necessary for them to make
the best engineering judgments concerning
the long term effectiveness in
site-specific conditions. The user
should have some familiarity with general
soil characteristics, water balance,
climatic conditions, and fundamentals of
leachate generation.
Closure of Hazardous Waste Surface
Impoundments (SW-873)
"Closure of Hazardous Waste Surface
Impoundments" presents a listing of
closure plan and post-closure care
considerations and details for surface
impoundments containing hazardous
wastes. It is written primarily for
staff members in EPA Regional offices or
state regulatory offices, who are charged
with evaluating and approving closure
plans for surface impoundments under
regulations of the Resource Conservation
and Recovery Act of 1976. Methods of
assessing site closure considerations are
documented.
The guide describes and references
the methods, tests, and procedures
involved in closing a site in such a
manner that (a) minimizes the need for
further maintenance, and (b) controls,
minimizes, or eliminates, to the extent
necessary to protect human health and the
environment, post-closure escape of
hazardous waste, waste constituents,
leachate, contaminated rainfall or waste
decomposition products to groundwater,
surface water, or the atmosphere.
Problems that have been overlooked in
abandoned impoundments and have caused
environmental damage are discussed. The
techniques involved are pertinent to
closing an impoundment either by removing
the hazardous waste or by consolidating
-------
the waste on-site and securing the site
as a landfill. Technical criteria for
implementing the closure, specifically
those regarding aspects substantially
different from a landfill, are given.
Relevant literature or procedures are
documented for more in depth review.
Hazardous Waste Land Treatment
(SW-874T
The objectives of "Hazardous Waste
Land Treatment" are to describe current
technology and to provide methods for
evaluating the performance of an
applicant'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. However, the lack of systematic
study or monitoring of many past or
present facilities has limited the
knowledge and data base with regard to
the important parameters and
interactions. Additionally, many
potentially land treatable wastes have
not been tested or have been examined
only under a limited range of
conditions. Therefore,-information is
needed on site selection, waste
characteristics, optimum soil and
climatic conditions for degradation of
organic constituents, application rates
and scheduling, decomposition products,
and the contingency for environmental
contamination. Potential problems
include worker's health and safety,
organic components, water pollution
hazards, and local atmospheric emissions.
Land treatment of hazardous waste has
not been under close scrutiny for a long
enough period of time to acquire numerous
documented failures resulting in
environmental pollution. The potential
for such failures due to the lack of
proper design and/or management is
evident from research conducted on the
land treatment of other nonhazardous
waste materials. For land treatment to
be an effective system, the process must
operate within a narrow range of
parameters. Exceeding them could result
in the uncontrolled release of pollutants
to the environment.
FUTURE TECHNICAL RESOURCE DOCUMENTS
Additional TRDs now being developed
or in planning stages by the Office of
Research and Development:
A listing of these documents is shown
below, along with a brief description and
the project officer's name in parentheses.
1. SOIL-PROPERTIES, CLASSIFICATION
AND HYDRAULIC CONDUCTIVITY TESTING
This report is a compilation of
available laboratory and field testing
methods for the measurement of hydraulic
conductivity (permeability) of soils.
Background information on soil
classification, soil water, and soil
compaction are included along with
descriptions of sixteen methods for
determination of saturated or unsaturated
hydraulic conductivity. This TRD
(SW-925) was published by OSW in March
1984 for public comment. It is being
revised to incorporate public comments
that were received. The revised version
will be published through the Government
Printing Office in late 1985.
(M.H. Roulier)
2.
MANUAL
SOLID WASTE LEACHING PROCEDURES
This is a report 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 by OSW 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
Government Printing Office in late 1985.
(M. H. Roulier)
3. BATCH SOIL PROCEDURE TO DESIGN
CLAY LINERS FOR POLLUTANT REMOVAL
This project is studying the
feasibility of using soi1/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 September
1985. (M. H. Roulier)
4. METHODS FOR THE PREDICTION OF
LEACHATE PLUME MIGRATION AND MIXING
This project has developed a variety
of computer programs for hand-held
calculators, microcomputers, and
macrocomputers. 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 late
1985. (M. H. Roulier)
5. HYDROLOGIC EVALUATION OF
LANDFILL PERFORMANCE (HELP) MODEL
The HELP Model is a modification of
the original waste disposal site
hydrologic model entitled,"Hydrologic
Simulation on Solid Waste Disposal
Sites." This update has incorporated the
two-dimensional aspects of landfill cover
systems, as well as the addition of the
leachate collection system. OSW
published this TRD (SW-84-009 and
SW-84-010) for public comment in two
volumes. 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
comments and results from verification
studies and will be published in early
1986. (D.C. Ammon)
6. DESIGN, CONSTRUCTION,
MAINTENANCE, AND EVALUATION OF CLAY
LINERS FOR HAZARDOUS WASTE FACILITIES
This project 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
May 1985. (M. H. Roulier)
CONCLUSION
The Technical Resource Documents
(TRDs) that are being prepared by the
Land Pollution Control Division (LPCD)
are a series of documents which provide
design, operation, and evaluation
information related to hazardous waste
disposal facilities to assist the
regulated community and the permitting
authorities. Eight documents have been
completed with six more in process by the
Office of Research and Development.
These documents describe state-of-the-art
summaries of technologies and evaluation
techniques determined by the Agency to
constitute good engineering design,
practices and procedures. The TRDs
present the sum total of the body of
information and experience gained by the
Agency over the years on a given topic.
As such, they contain factual summaries
concerning the experience and
effectiveness of design alternatives,
covering what has been found not to work
as well as what has been found to be
effective. The TRDs can be considered as
technical background or development
documents supporting the RCRA guidance
documents and regulations As new
information is developed, the Agency
intends to update each of these documents
so that they reflect the latest
state-of-the-art information.
The project officers can be contacted
by writing or telephoning the: U. S.
EPA, Hazardous Waste Engineering Research
Laboratory, Land Pollution Control
Division, 26 West St. Clair Street,
Cincinnati, Ohio 45268, Telephone (513)
684-7871.
-------
EVALUATION OF CHEMICAL GROUT INJECTION TECHNIQUES
FOR HAZARDOUS WASTE CONTAINMENT
James H. May, Robert J. Larson, Philip G. Malone,
and John A. Boa, Jr.
US Army Engineer Waterways Experiment Station
Vicksburg, MS 39180-0631
ABSTRACT
Field tests were conducted with two chemical grouts to determine if a continuous im-
permeable seal could be developed by injecting grout into coarse-grained soils. Acrylate
and sodium silicate grouts were injected into separate 2 x 4 m test beds of medium sand
at a 30-50 cm depth using a maximum hole spacing of 40 cm (center-to-center) to produce a
pan-like layer of grout. While some silicate grout bulbs coalesced, there were signifi-
cant gaps in the grout layer and shrinkage of grout produced nonuniform cementing. The
acrylate grout flowed to the bottom of the sand bed and did not produce grout bulbs.
In a large field test, 30-percent sodium silicate was injected at 2.44 m depth in
fine sand using injection holes on 1.52 m-centers. Significant problems were noted when
the pods were excavated. The pods were not symmetrical and ungrouted areas were present
between pods. Pockets of coarse sand in the finer sand were not cemented although they
had been filled with grout. Root and rootlet holes in the cemented sand were not sealed.
Conventional injection of chemical grouts did not produce an impervious seal. More
unusual techniques, such as jet grouting, may be needed to obtain a useful seal.
INTRODUCTION
Chemical grouts have been widely
employed in adding physical strength to
foundation materials and reducing the per-
meability of soils ' . The use of chemi-
cal grouts in sealing soils around hazard-
ous waste sites is an extension of existing
techniques; but very specific problems that
relate to grout solidification, durability,
and techniques for grout injection must be
treated when hazardous waste sites are
being sealed. Previous reports from this
work unit ' '" have discussed the prob-
lems of grout setting and the physical
durability of grout after placement.
PURPOSE
The present report discusses problems
associated with the injection of selected
grouts in sand test beds and a larger-scale
field test where chemical grout (sodium
silicate) was injected into an 85-sq m
test area. In both operations the objec-
tive was to evaluate the ability of stand-
ard grout injection techniques to produce a
continuous layer of grouted sand that could
potentially be used as a barrier to waste
movement. None of the test areas contained
hazardous wastes.
APPROACH
The two chemical grouts (acrylate and
sodium silicate) that showed the greatest
potential for successful application based
on their ability to set and to remain unal-
tered in the presence of typical wastes
were selected for further field-scale
injection. The extended testing was de-
signed to determine if the chemical grouts
could be made to form a continuous seal
when injected in sand. Only conventional
injection techniques using grout pumps and
shallow injection points were employed.
-------
Small-Scale Grouting Field Tests
Two types of grout, aerylate and sodi-
um silicate, were selected for small-scale
field testing to see if coalescing grout
bulbs could be produced. These two grouts
were considered as test candidates because
they were relatively easy to control in
terms of gelling times and showed some
ability to set or gel in the presence of
potential contaminants, although displace-
ment of contaminants with clean water may
be necessary to.avoid waste/grout compati-
bility problems
The small-scale tests were performed
by injecting each grout into a medium,
well-sorted sand in a separate bin that
contained a 1-m deep layer of sand. Each
test plot covered an area 2 x 4 m. Grout
was injected at a depth of 30 to 50 cm
in a grid-like pattern (see diagram in
Reference 4) with a maximum spacing of
40 cm between hole centers (Figure 1).
The objective of the test was to pro-
duce a pan-like continuous seal.
The silicate grout used was a 30 per-
cent solution of JM-grade sodium silicate
(technical). The hardener employed was a
proprietary mixture containing calcium
chloride, magnesium chloride, and
dimethylformamide. The chemical grout was
made up using tap water.
The acrylate grout was made up from a
proprietary mixture of acrylate monomer and
methylenebisacrylate. Ammonium persulfate
was used as an initiator. The gel was also
made up with tap water.
Both grouts were injected using a pro-
gressive cavity pump and each injection
hole received approximately 7 liters of
grout. Gel times were established in the
laboratory to allow five minutes per batch
for mixing and grout injection. Each 7-
liter batch was mixed and injected sepa-
rately so that gel time could be checked.
In all cases the test gel time was identi-
cal to or less than the design gel time.
No simulated wastes were present, and no
interference from the test bed material was
observed.
Figure 1. Photo of test bed for small-scale test sealing.
The pipe sections mark the grout injection points.
-------
The test beds were a well-sorted,
medium-grained quartz sand with less than
10 percent pebble-sized material
(Figure 2). The sand was wet to field
capacity. The test beds were saturated but
allowed to drain prior to grout injection.
A 7-cm diameter slotted plastic pipe was
implaced under each bed to allow free
drainage. After grout injection the test
beds were covered with a plastic tarp to
allow the grout to gel and cure without
being washed out by rainwater. During both
grouting operations, water and grout were
observed to be flowing out of the test bed
drains suggesting displaced capillary water
and some of the grout was being lost to the
bottom drains. The grout injection pipe
was withdrawn from the sand after grouting
and in all cases some liquid grout remained
in the injection holes indicating the bot-
tom and side walls of the hole had been
saturated with the grout. In both of the
freshly grouted test beds there was evi-
dence of gel in the surface sand around the
hole; but only the sodium silicate made a
cohesive pod of material around the grout
hole. In the acrylate material thin rub-
bery strands of gel could be observed when
the surface sand was examined, but the in-
jected sand had little cohesion.
Both the silicate and acrylate grouted
beds were tested to determine if a contin-
uous seal had been formed. A shallow (10-
cm deep) trench was dug on the loose sand
over the grouted layer and water containing
a dye (fluorescein) was added. To assure
that the liquid did not overtop the pan-
like layer, only 80-100 liters of water was
placed in the trench and water was added
only when all of the dye had drained
through the bottom of the trench and no
drainage water appeared at the bottom of
the test unit. The times required for the
appearance of dye at the test bed drain and
the quantities of water and dye added were
not significantly different for the grouted
beds when compared to a similar but un-
grouted sand bed. The dye testing indi-
cated that no continuous seal had formed.
The grouted test beds were excavated
and the position and character of the
grouted material was noted. In the sodium
silicate test bed two problems were appar-
ent; gaps existed between adjacent grout
bulbs where water could migrate around
grout bulbs into the drainage layer
(Figure 3) and the center portions of many
of the grout bulbs were not cemented
(Figure 4). The grout bulbs in many cases
were in contact, but enough gaps existed to
allow the rapid movement of dye through the
grouted layer. The lack of cementation in
the central portion of the pods is more
difficult to explain. The phenomenon is
probably caused by shrinkage due to water
loss (syneresis), but the gel is concen-
trating on the outside of the sand mass and
did not shrink to the inside of the mass.
The acrylate grout did not form pods
and very little grout appeared in the upper
sand layer. The major part of the acrylate
grout was found in irregular masses at the
bottom of the test bed (Figure 5). The
grout had not plugged the slotted pipe
drain but was covering portions of the
pipe. The dye could easily pass through
the injected layer into the drain. The
grout had not created a barrier to dye
movement. Test injections of the acrylate
grout indicated that gelling occurred with-
in minutes of injection and inspection of
the sand on freshly grouted holes indicated
gel was present in sand at the surface in
the immediate area of the hole at the time
of injection. The gel did not appear to
adhere to the sand and may have migrated to
the lower part of the test bed. The water
added to the cart during dye testing may
account for some of the grout movement, al-
though the water addition occurred after
gel formation should have been complete.
Large-Scale Grouting Field Tests
Experience with the small-scale sand
test beds was incorporated into a larger
field test. Sodium silicate was used for
grouting and a field test site underlain by
fine sand was selected to minimize problems
of grout shrinkage^ '. To maximize the
probability for producing coalescing grout
pods, all holes were grouted to refusal or
until grout take was 0.76 cu m.
An 85-sq m plot underlain by fine sand
(Figure 6) was made available for this
study on a test area at Fork Polk, LA. Ten
holes were spaced on 1.52-m centers using
the pattern shown in Figure 7. Injection
holes were drilled with a 10.2-cm-diameter
auger to a depth of 2.75 m and the hole was
backfilled with coarse filter sand to a
depth of 2.44 m. A 3.05-m section of 5-cm
plastic pipe was inserted in the boring and
the outside of the borehole was sealed with
pelleted and powdered bentonite. Water was
added after each batch of bentonite to
10
-------
100
90
60
»-^
UJ
3 60
OD
|so
u.
£40
UJ
u
ce
a: 30
20
10
0
6(
U.S. STANDARD SIEVE SPENINO IN INCHES
8 43 2 li 1 1 i
1 1
)0 100
1
1
1 ""
60 10
U.S. STANDARD SIEVE NUMBERS HYDROHETER
34 8 810 16 20 30 40 60 70100140200
•J^
k
1
•^
1
s
1
^
6 1
^
\
1
>
\
1
1
\
\
\
\
;
1
f
l\
ts
1 1
~-e-e.
0.6 0.1
i
i
0
10
20
30 ;
(
40 '.
50
60
70
80
90
inn
0.06 0.01 0.006 0.001
>-
(O
V)
oe
-------
Figure 3. Photo of gaps between adjacent grout bulbs observed in the
small-scale field test.
Figure 4. Photo of grout bulbs formed during small-scale injection. Note
the shells of bulbs indicating the outside layer of sand cemented
and the inside did not.
12
-------
Figure 5. Photo of the sand mass at the bottom of the test bed
cemented with aerylate gel.
13
-------
U.S. STflNDfiRB SIEVE OPENING IN INCHES
U.S. STftNOBRO SIEVE NUMBERS
HYDROMETER
inn 6 4
90
80
£70
ci
UJ
3 60
CD
|so
u.
£40
UJ
o
£30
20
10
0
51
i i
jo ior
3 2 li 1 I i 1
1
1
r T i
SO 10
3 4 fi 8U
T
r
P Ijg 20 30 40 SO 70100140200
' T"'
^1
[>
\
1
\
\
<
1
)
\
\
\
4)
1 l
\
*v
«~
1
" J
0
10
20
30 |
UJ
3
40 >-
ca
Of
UJ
50 OT
cc
o
o
60 ,_
z
UJ
0
70 u5
Q_
80
90
100
5 1 0.5 0.1 0.05 0.01 0.005 0.001
GRflIN SIZE IN MILLIMETERS
COBBLES
GRflVEL
COARSE | FINE
SflND
CORRSE
MEDJUI1 | FINE
SILT OR CLflY
Figure 6. Grain-size distribution of sand under the larger-scale test area.
-------
North-
• Injection point
*2 Edge of grout pod at its maximum diameter
SCALE
. 1 Meter
Figure 7. Layout of grout injection holes at the larger-scale test area.
-------
assure that the bentonite would hydrate
and swell. A 2.5-cm diameter pipe, 4 m
in length with both ends sealed with
tape, was inserted in the 5-cm pipe.
The lower sealed end of the pipe was
pressed into the filter sand. The inner
pipe was cemented in place with a mixture
of 50 percent portland cement and 50 per-
cent mortar sand. A similar portland
cement mixture was used to complete the
seal around the outside of the larger 5-cm
pipe. A 1.5 cm plastic pipe was run down
into the fluid cement/sand mixture and agi-
tated to assure that the thin cement mix-
ture would flow to the points where sealing
was needed.
The cement/sand mix was allowed to
cure for at least 24 hours before grout was
injected. When grout was to be injected,
the top of the 2.5-cm pipe was untaped and
a 1.5-cm plastic pipe with a solid pointed
end was run down inside the 2.5-cm diame-
ter pipe and the lower tape was broken by
driving the point 3 to 5 centimeters in-
to the filter sand. This procedure assured
that the grout had clear access to the fil-
ter sand pocket at the end of the grout
pipe. The water table was 2.94 m below
ground surface. The injection point was 50
cm above the water table when grout was
pumped.
A 30-percent sodium silicate grout
similar to that used in the small-scale
testing was mixed in batches and injected
using a progressive cavity pump. Grout in-
jection continued until refusal was ob-
tained or 760 liters was injected. Refusal
was indicated by a rapid pressure rise in
the grout pipe followed by stalling the
pump or by a "blow-out" of grout to the
surface either through the soil or the in-
jection hole annulus.
The injection holes on the outside of
the plot were grouted first. The two in-
side holes (B2 and B3 in Figure 7) were
grouted last. Only one hole, B4, was not
successfully injected. The grout gelled
prematurely on this boring and plugged the
grout hose. An adjustment in the grout
formulation corrected this problem at sub-
sequent injection points. Samples of grout
were taken from each batch to assure that
the grout gelled. Food coloring was added
to each grout batch to allow solidified
grout to be identified. Color could only
be used to identify a source when fragments
of gel could be recovered intact.
The brown and red sand masked the color of
the gel in the grout pods.
The grout pods were excavated after
approximately 120 days. Figure 7 shows the
distribution of the pods over the test
plot. Three problems were recognized as
the pods were excavated:
a. The grouted sand bodies were often
unsymmetrical and ungrouted areas were
present between pods. A continuous seal
had not formed.
b. Holes larger than those between
sand grains (root holes or rootlet holes)
were not sealed with grout (Figures 8-9).
c. The coarse-grained sand used under
the boreholes to distribute the grout was
not cemented (Figure 10).
CONCLUSIONS
The results of the test grouting and
excavation indicate that the chemical grout
injection did not produce a successful
seal. The anisotropic flow of grout in the
soil prevented the grout pods from meeting
even when large grout takes were observed.
The acrylate grout can flow through sand
without cementing the particles into a co-
hesive mass. The silicate grout could not
seal small openings in soil even though
grout coatings indicated these holes had
initially been filled with gel. The coarse
sand usually contained no silicate grout.
The grout, although initially present,
flowed out of or shrank away from the
coarse sand pockets. Any coarse sand
pocket may represent an unsealable hole
in a silicate grout barrier.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the
assistance of Dennis L. Bean, Larry G.
Caviness, and Steven A. Houston of the
Waterways Experiment Station in this phase
of the project. Herbert Pahren is the EPA
Project Officer for this investigation.
16
-------
Figure 8. Photo of rootlet hole present in solidly cemented sand pod.
Figure 9. Photo of root holes containing hardened grout. The holes
the soil were not sealed.
17
-------
Figure 10. Photo of coarse-grained sand placed at the bottom of the grout-
injection hole. Note that the sand was not cemented although
the adjacent fine sand was cemented.
REFERENCES
1. Bowen, R. 1981. Grouting in Engineer-
ing Practice, 2nd Ed., John Wiley and
Sons, NY, 1981.
2. Karol, R. H. 1982. Chemical Grouts
and Their Properties, Proceedings of
the Conference on Grouting in Geotech-
nical Engineering, Amer. Soc. of Civil
Engineers, NY, pp 359-377.
3. Larson, R. J., and J. H. May. 1983.
Geotechnical Aspects of Bottom Sealing
Existing Hazardous Waste Landfills by
Injection Grouting, Proceedings of the
First Annual Hazardous Materials Man-
agement Conference Tower Conference
Management, Wheaton, IL, pp 513-529.
4. Malone, P. G., J. H. May, and R. J.
Larson. 1984. Development of Methods
for In-Situ Hazardous Waste Stabiliza-
tion by Injection Grouting. Land Dis-
posal of Hazardous Wastes. EPA-600/9-
84-007. US Environmental Protection
Agency, Cincinnati, OH, pp 33-42.
Malone, P. G. 1984. Cementing Reac-
tions in the Hazardous Waste
Environment. Hazardous Wastes and En-
vironmental Emergencies. Hazardous Ma-
terials Control Research Institute,
Silver Spring, MD, pp 299-301.
Bodocsi, A., I. Minkarah, and B. W.
Randolph. 1984. Reactivity of Various
Grouts to Hazardous Wastes and
Leachates. Land Disposal of Hazardous
Wastes. EPA-600/9-84-007. US Environ-
mental Protection Agency, Cincinnati,
OH, pp 43-51.
18
-------
PERMEABLE MATERIALS FOR THE REMOVAL OF POLLUTANTS
FROM HAZARDOUS WASTE LEACHATES
James E. Park
University of Cincinnati
Dept. of Civil and Environmental Engineering
Cincinnati, OH 45221
ABSTRACT
In order to better understand the potential effectiveness of permeable treatment
systems, four readily available, low-cost permeable materials -- limestone, coal, fly
ash, and a soil containing clay -- were tested to determine their ability to remove
organic pollutants from two simulated hazardous waste leachates. The capabilities of
various sequentially ordered layers of these materials were evaluated with respect to
their ability to retain total organic carbon (TOC) and twelve selected priority pollu-
tants. As a result of testing, the most effective ordering of materials was found to
be a layer of fly ash, followed by a layer of coal, followed by a layer of limestone.
INTRODUCTION
As the number of hazardous waste
sites requiring remedial action increases,
technological alternatives that not only
enclose sites, but mitigate contamination
will become increasingly popular. One
such alternative might use permeable mate-
rials in systems designed for the reten-
tion of pollutants from hazardous waste
leachates. These permeable treatment sys-
ems could be designed as vertical treat-
ment walls or horizontal treatment beds.
In the former case, the treatment wall
would partially or totally encircle a site
in a manner similar to a slurry trench,
retaining pollutants as leachates moved
through the wall. In the latter case, the
treatment bed would be used to pretreat
pumped leach ate prior to its release to an
industrial or municipal wastewater treat-
ment plant. These types of permeable
treatment systems are described in the
handbook, Remedial Action at Waste
Disposal Sites (19), but are identified as
conceptual rather than developed, state
-of-the-art technologies. In order to
better understand the utility of the per-
meable treament systems concept, four
readily available, low-cost materials were
tested by measuring their effectiveness in
retaining organic pollutants from two
simulated hazardous waste leachates. The
four materials tested were limestone, fly
ash, and a soil containing clay.
Prior to testing, a review of the
literature indicated that previous work in
the area of pollutant retention by perme-
able materials had been performed by Ful-
ler, et al (4, 5, 6, 7, 8), Korte (13, 14,
15) and Alesii (1) using soils and crushed
limestone: by Griffin, et al (9, 10, 11,
12) emphasizing soils: by Chan, et al (2,
3) and Sheith (16) investigating fly ash,
clays and other materials; and Spoljaric
and Crawford (17, 18) examining green-
sands. These studies focused on in-
organic pollutants and provided a basis
for understanding the retention of metals
and other inorganic toxic pollutants by
soils containing clay, limestone, fly ash,
and greensands. Little information was
found in the literature regarding reten-
tion of organic pollutants by the afore-
mentioned materials.
PURPOSE
The purpose of this study was to test
four readily available, low-cost permeable
materials -- limestone, coal, fly ash, and
a soil containing clay — for their abil-
ity to retain organic pollutants from
19
-------
two simulated hazardous waste leachates.
Various combinations of the materials,
ordered in sequential layers, were tested,
and their interactions with selected or-
ganic pollutants were measured so that the
individual and collective performance
characteristics of the materials could be
measured.
APPROACH
The choice of permeable materials was
based on their estimated ability to retain
pollutant concentrations, as w^ll as their
local availability. Natural or waste ma-
terials were judged to have an apparent
cost advantage over such man-made materi-
als as activated carbon or ion exchange
resins. For this reason, readily avail-
able, low-cost materials were tested.
Coal was obtained from a mine near Stin-
son, KY, limestone from a local dealer,
fly ash from the Miami Fort Power Plant
near Harrison, OH, and soil from northern
Hamilton County near Cincinnati, OH. The
particle size distributions for each of
these materials are shown in Figure 1.
100
80
cc 60
40
20
FLY ASH
LIMESTJ
100
10
1 .1
MM DIAMETER
.01
.001
Figure 1. Particle Size Distribution
The materials were placed in 12,
20.3 cm x 20.3 cm cross-sectional columns.
Layer thicknesses were chosen for compara-
tive purposes of volume, 20 cm being twice
the volume of 10 cm; 6 cm layers were also
used mostly for limestone. The total
height of all columns was 36 cm. The la-
yer orderings used for both experiment
runs are shown in Figure 2. The feed for
each run was common to all columns. Por-
ous stones connected to external sample
ports were placed between each layer and
at the bottom of each column. An illus-
tration of the sampling devices used is
shown in Figure 3. These allowed the TOC
concentration to be monitored daily as the
simulated hazardous waste leachate flowed
through the materials in the 12 columns.
F
L
S
F
L
S
C
L
S
C
L
S
F
C
S
F
C
S
7
C
L
S
8
C
F
L
S
9
S
L
F
10
S
L
C
11
F
12
C
C - COAL
S - SOIL
F - FLY ASH
L - LIMESTONE
Figure 2. Materials Ordering
POROUS STONE
PLEX1QLAS WALL
Figure 3. Sample Device
20
-------
Two different simulated leachates
were used in the test. Since organics
were of primary concern, total organic
carbon (TOC) analysis was used as a mea-
sure of their presence in both Runs #1 and
#2. Gas chromatographic (GC) analysis for
12 spiked priority pollutants was perfor-
med on the effluents from Run #2 to pro-
vide an indication of their relative mo-
bilities. The first simulated leachate
was comprised of experimental solid waste
lysimeter leachate spiked with phenol and
dichlorobenzene at 400 mg/L and 80 mg/L,
respectively. The TOC level for Run #1
began at 2500 mg/L and decreased to 400
mg/L by the end of the run. Average col-
umn feed concentrations based on total
feed volume and daily TOC concentrations
were 115 mg/L and 1596 mg/1 respectively.
The simulated leachate for Run #2 was
prepared daily and consisted of 0.1N
CaS04, 850 mg/L potassium hydrogen phtha-
late (KHP), 100 mg/L toluene and the
priority pollutants listed in Table 1.
The average TOC was 421 mg/1.
TABLE 1. PRIORITY POLLUTANTS IN FEED #2
Bis(2-ethylhexyl) Phthaliate
Di-N-Butyl Phthalate
1,4 Dichlorobenzene
2,4 Dichlorobenzene
Ethylbenzene
Fluoranthene
Isophorone
Pentachlorophenol
Phenanthrene
Phenol
Pyrene
Naphthalene
Actual Cone.
(ug/1)
163
128
192
109
261
129
120
115
135
137
Calculations of the retentive capaci-
ties were made from the breakthrough
curves for each column. Retention was
interpreted as the amount of the pollu-
tants, in this case represented by TOC,
that remained on a given amount of materi-
al (mg/kg dry.wt.). Knowing the total TOC
input and output, the difference is re-
tained by the permeable materials. Typi-
cal examples of breakthrough curves are
shown in Figures 4 and 5. The cumulative
flow volume and the TOC concentration were
normalized to the total flow for each col-
umn and the maximum TOC concentration for
each experiment. The areas between each
curve are proportional to the amount re-
tained by each layer of the column. The
total amount of pollutants retained was
found by summing the rectangular areas
between data points.
1.0
.8
. .6
j
z
J.4
.2
FLY ASH
EFFL.
0 .2 .4 .6 .8 1.0
FLOW VOL.
Figure 4. Normalized Breakthrough Curves,
Run #1, Column 5
1.0
.8
..6
o
z
o
°.4
.2
FLY ASH
EFFL.
.2
.8 1.0
Figure 5.
.4 .6
FLOW VOL.
Normalized Breakthrough Curves,
Run #2, Column 5
21
-------
The overall effectiveness of the or-
derings of materials were compared by
ranking four column results: 1) The num-
ber of days until 50% breakthrough, 2) the
total TOC retained, 3) the linear flow ve-
locity and 4) the calculated permeability
coefficient. The first two results are
measures of retention, while the last two
are indicative of flow. The most effec-
tive column then should have both high
retention and flow.
PROBLEMS ENCOUNTERED
A problem was encountered with the
feed used for the first experimental run.
The decrease of the TOC concentration was
probably caused by a combination of micro-
bial degradation of the organics and vola-
tilization. This caused decreases in con-
centration near the end of the run rather
than the expected increases. This problem
was eliminated in Run #2 when the simula-
ted leachate was spiked daily.
Another problem was the low flow
rates in columns 2 and 4 which both con-
tained 20 cm layers of soil. The clay
component of the soil reduced the flows
through these columns to 200-250 ml/day.
Flows in the other columns ranged from
400-750 ml/day. Although this wasn't a
problem from an experimental standpoint,
the length of the runs, especially Run #2
which lasted 47 days, might cause concern
for commercial operations that require
answers quickly. It should be noted,
however, that the time required for test-
ing is a function of scale and concentra-
tion. Since the concentrations tested
were higher than those expected at most
remedial action sites, the size of the
test apparatus should be reduced to facil-
itate efficient testing time and deter-
mination of permeable material effective-
ness.
RESULTS
The retentive capacity results for
the individual permeable material layers
are shown in Figure 6. The range and
average values for the pollutants retained
during the two simulated leachate runs are
presented in Table 2. Coal retained the
most TOC per unit weight in both runs.
The soil, containing about 20^ clay-sized
particles retained a large amount of pol-
lutants in Run #1 and slightly less in
Run #2. The fly ash retention results
are similar for both runs, indicating
its retentive capacity may not be related
to strong adsorption reactions. Limestone
retained very little of the organic pollu-
tants, although it did precipitate iron
eluted from the coal. Rust colored layers
in the limestone were clearly visible when
a coal layer was located above a limestone
layer.
Batch elution tests were run using
300 ml of 0.1N CaS04 and 50 g of the con-
centration-saturated materials from Run
#2. The elution values (mg/kg dry wt.)
are included in Figure 6, in parentheses.
In the case of fly ash, all the retained
TOC was eluted, supporting the contention
that fly ash is not a strong adsorber.
The coal held as much as half the retained
TOC following elution, while the soil
elution results were mixed, most samples
eluting all the retained TOC with a few
holding up to 30%. Thus, coal was the
only material exhibiting relatively strong
adsorbtion capabilities.
The rate of flow through the layers
influences retention by the permeable ma-
terials. The porosity of the coal layers
allowed rapid movement of TOC. Columns 3
and 7, with 20 cm of coal on the top in
Run #1, provided less retention than the
coal layers in the other columns. Limit-
ing the flow through the coal by a less
permeable layer above it, as in Column 6,
appeared to increase the retentive capa-
bility of coal. The 20 cm layers of soil
in Columns 2 and 4 inhibited flow, while
the fly ash in Columns 1 and 5 only mod-
erated the flow. In order to better
understand the permeability of the materi-
als used, permeameter tests were run and
the results are shown in Table 3 for coal,
fly ash, and soil. Limestone was not
tested in the permeameter since it was a
uniformly distributed granular material
that was not expected to effect liquid
flow rates when in combination with the
other materials.
The results of the GC analyses for
Run #2 are presented in Table 4. The ma-
jority of the occurrences of the twelve
priority pollutants in the effluents were
at lower levels than in the simulated
leachate feed. Generally, even when a
compound was detected at a concentration
greater than 100 ug/1 in a sample, it
22
-------
RUN
ALL VALUES ARE MQ/KG
L= LIMESTONE
C= COAL
F= FLY ASH
S= SOIL
F
144
L o
S
732
F
261
L 87
S
428
C
116
L3U
S
574
C
629
L o
S
208
F
115
C
1115
418
F
147
C
617
so
C
115
L
0
S1033
C
1605
F
412
L
0
S
562
S
358
L
0
F 548
S
660
L
0
°1535
F
372
C
870
8
10 11
12
RUN *2
ALL VALUES ARE MG/KG
( ) = ELUTION
L= LIMESTONE
C= COAL
F= FLY ASH
S= SOIL
F
293
(334)
L 33
(269)
S203
(363)
F319
(323)
L(3&1)
S
172
(379)
C
776
(291)
L 0
(287)
S
145
(341)
°467
(351)
(287)
S
177
(335)
F
360
(350)
C452
(397)
S (88)
F373
(376)
C
427
(459)
S rrr
c
535
(222)
L 0
(262)
s 689
(403)
C 364
(294)
F
377
(421)
L 66
(298)
s 715
(390)
S
345
(356)
L 0
(270)
F 0
(427)
S
339
(441)
L 21
(341)
C £
(544)
F
215
(348)
C
463
(237)
8
10 11
12
Figure 6. Retention and Elution Results, Runs 1 and 2
23
-------
TABLE 2. AVERAGE RETENTION CAPABILITIES
Range(mg/kg) Average(mg/kg)
Run #1 (2500-400 mg/1 TOC)
Coal (1605-115)
Limestone (311-0)
Fly ash (548-115)
Soil (1033-0)
Run #2 (421 mg/1 TOC)
Coal
Limestone
Fly ash
Soil
(776-32)
(98-0)
(377-215)
(715-140)
825
50
286
397
440
32
277
348
TABLE 3. PERMEABILITY RESULTS (cm/sec)
Coal
Fly ash
Soil
1.8
5.3
1.7
- 5.6
- 7.0
x 10-4
x 1C'4
x ID'5
- 3.5
x lO-5
would not be detected in the subsequent
sample. The detected compounds appeared
predominately in discrete cases. Phenan-
threne, fluoranthene, dichlorophenol,
pyrene and pentachlorophenol were either
undetected or appeared at concentrations
less than 20 vg/1. These compounds would
be less likely to migrate than the other
compounds listed in Table 1.
The four criteria used for evaluating
column performance were: 1) the number
of days to 505» breakthrough, 2) the total
TOC retained, 3) the linear flow veloc-
ity, and 4) the calculated permeability
coefficient. As previously mentioned, the
first two results are measures of reten-
tion, while the last two results are indi-
cative of flow. The best performing col-
column then would have both high reten-
tion and high flow . The overall rankings
for performance were found by assigning
numerical values (1 through 12) to each of
the columns for each of the four results.
For example, in Run #1 Column 12 retained
the most TOC so it was given a "1" rank-
ing, while Column 7 retained the least and
received a rank of "12." The four ranked
results were then added and the totals
ranked again, 1 through 12. The orginal
rankings were also multiplied together;
these products werre then ranked. The
additive and multiplicative rankings were
then added together. The overall rank-
ings were determined from these numbers.
The additive and multiplicative methods
were combined in order to afford better
resolution when two or more columns were
closely ranked. The results are presented
in Tables 5 and 6. The overall rankings
are indicative of the responses of the
columns to the simulated leachates used in
the two runs. Since the soil used in the
test hindered flow, especially in Columns
2 and 4, these columns received lower
rankings despite long times to break-
through. Columns 3 and 7, with 20 cm coal
layers, performed poorly in Run #1 due to
high permeability and possible channel-
ling.
Considering the results of Run 12,
the most appropriate layer orderings were
found in Columns 5 and 6, with fly ash
above a coal layer. Ranked next was Col-
umn 3, which did not repeat its perfor-
mance from Run #1. Column 8 also ranked
high, and exhibited the least occurrences
of priority pollutants in its effluent.
Based on the results of Run #2 and indica-
tions from Run fl, the most appropriate
ordering of these materials would be a top
layer of fly ash for flow moderation, fol-
lowed by a coal layer as the primary re-
taining material, then a limestone layer
to adjust pH and precipitate the iron
leached from the coal.
ACKNOWLEDGEMENTS
This work was performed as a part of
the University of Cincinnati's contract
to operate the US EPA's Test and Evalua-
tion (T&E) Facility. The author grateful-
ly thanks Yu-Lin Liu for her considerable
help in running the tests. The Technical
Project Monitor and provider of guidance
was Jon Herrmann.
-------
TABLE 4. PRIORITY POLLUTANT GC ANALYSES RESULTS
Column
1
2
3
4
5
6
7
8
9
10
11
12
Column
1
2
3
4
5
6
7
8
9
10
11
12
Number
Compounds
Samples
Detected 7100 ug/1
30
25
21
13
17
16
20
10
21
14
14
13
Total
Flow (1)
7.18
6.66
6.79
3.93
8.95
7.74
6.47
9.22
6.21
9.23
8.08
8.92
2
1
1
1
1
3
5
0
5
3
4
3
TABLE 5
Total mg
Retained
6377
7839
5709
5247
7299
6103
5177
10493
6567
11683
7900
12274
Compounds >100 ug/1
Dibutylphthalate, Phenol
Phenol
Bis(2-ethylhexyl) Phthalate
Phenol
Bis(2-ethylhexyl) Phthalate
Phenol, Bis(2-ethylhexyl) Phthalate
Phenol, 1,2 Dichlorobenzene, Bis(2-ethyl ...)
Bis(2-ehtylhexyl) Phthalate, Phenol, Isophorone
Dibutylphthalate
Napthalene, Dibutylphthalate, Ethylbenzene
Bis(2-ethyl . . .), Dibutylphthalate, Ethylbenzene
Bis(2-ethyl...), Dibutylphthalate
. RUN #1 COLUMN PERFORMANCE DATA
Linear
Days to 50% Velocity Calculated K
Breakthrough (cm/day) (cm/day)
16 19.3 .78
18 17.2 .73
9 10.2 .96
19 7.5 .41
11 18.8 .98
12 12.4 .85
7 9.2 .91
1L 21.4 1.01
12 27.8 .68
18 23.5 .95
19 9.4 .84
7 22.2 .97
9
Overall
Ranking
8
6
11
10
4
9
12
2
7
1
5
3
TABLE
6. RUN #2 COLUMN PERFORMANCE DATA
Column
1
2
3
4
5
6
7
8
9
10
11
12
Total
Flow (1)
15.86
12.38
25.78
11.86
23.99
24.02
26.12
28.40
14.35
16.12
17.00
29.44
Total mg
Retained
4448
3857
7064
4310
6146
7525
6810
7006
3796
3911
4374
6533
Linear
Days to 5Q% Velocity Calculated K
Breakthrough (cm/day) (cm/day)
26 34.5 .84
39 21.7 .52
30 13.8 1.12
47 8.8 .50
16 60.5 1.47
37 41.2 1.01
19 14.8 1.10
20 16.0 1.19
13 23.3 1.36
14 25.9 1.00
28 31.1 .77
15 8.2 1.23
Overall
Ranking
5
9
3
8
2
1
7
4
11
12
6
10
25
-------
REFERENCES
1. Alesii, B.A. and W.H. Fuller, "The
Mobility of Three Cyanide Forms in
Soils." EPA 600/9-76-015, pp. 213-
223, 1976.
2. Chan, P.C., R. Dresnack, J.W. Lisko-
witz, A. Perna, R. Trattner, "Sor-
bents for Fluride, Metal Finishing,
and Petroleum Sludge Leachate Contam-
inant Control." EPA 600/2-78-024,
1978.
3. Chan, P.C., J.W. Liskowitz, A. Perna,
R. Trattner, "Evaluation of Sorbents
for Industrial Sludge Leachate Treat-
ment." EPA 600/2-80-052, 1980.
4. Fuller, W.H., "Liners of Natural
Porous Materials to Minimize Pollu-
tant Migration." EPA 600/2-81-122,
1981.
5. Fuller, W.H., "Soil Modification to
Minimize Movement of Pollutants from
Solid Waste Operations." Mar. 1980,
CRC Critical Reviews in Environmental
Control (1980a).
6. Fuller, W.H., A. Amoozegar-Fard,
G.E. Carter, "Predicting Movement of
Selected Metals in soils: Applica-
tions to Disposal Problems." EPA
600/9-79023a, pp. 358-374 (1979a).
7. Fuller, W.H., N.E. Korte, E.E.
Niebla, B.A. Alesii, "Contribution of
the Soil to the Migration of Certain
Common Trace Elements." Soil Sci.
122 (4), pp. 223-235 (1976a).
8. Fuller, W.H., C. McCarthy, B.A.
Alesii, E.E. Niebla, "Liners for
Disposal Sites Sites to Retard
Migration of Pollutants." EPA 600/9-
76-015, pp. 112-126 (1976b).
9. Griffin, R.A., R.R. Frost, A.K. Au,
G.D. Robinson, N.F. Shimp, "Attenu-
ation of Pollutants in Municipal
Landfill Leachate by Clay Minerals:
Part II, Heavy Metal Adsorption."
Envir. Geol. Notes. No. 79, 1977.
10. Griffin, R.A., K. Cartwright, N.F.
Shimp, J.D. Steele, R.R. Ruch, W.A.
White, G.M. Hughes, R.H. Gilkeson,
"Attenuation of Pollutants in Munic-
ipal Landfill Leachate by Clay
Minerals: Part I, Column Leaching
and Field Verification."
Envir. Geol. Notes, No. 78 (1976a).
11. Griffin, R.A., R.R. Frost,
N.F. Shimp, "The Effect of pH on the
Removal of Heavy Metals from Leach-
ates by Clay Minerals." EPA 600/9-
12. Griffin, R.A., N.F. Shimp,
J.D. Steele, R.R. Ruch, W.A. White,
G.M. Hughes, "Attenuation of Pollu-
tants in Municipal Landfill Leachate
by Passage Through Clay." ES&T,
V.10, No. 13, pp. 1262-1268
(1976c).
13. Korte, N.E., J.K. Skopp, W.H. Fuller,
E.E. Niebla, B.A. Alesii, "Trace Ele-
ment Movement in Soils: Influent of
Soil Physical and Chemical Proper-
ties." _Soi_l_Sc_i_._ 121(6), pp. 350-359
(1976a).
14. Korte, N.E., J.K. Skopp, E.E. Niebla,
W.H. Fuller, "A Baseline Study on
Trace Metal Elution from Diverse Soil
Types." Water. Air. Soil Poll. 5:
pp. 149-156, 1975.
15. Korte, N.E., W.H. Fuller,
E.E. Niebla, J.K. Skopp,
B.A. Alesii, "Trace Element Migration
in Soils: Description of Attenuated
Ions and Effects of Solution Flux."
EPA 600/9-76-015, pp. 243-258
(1976c).
16. Sheih, M.S., "The Use of Natural Sor-
bents for the Treatment of Industrial
Sludge Leachate." Dissertation, New
Jersey Inst. of Tech., 1979.
17. Spoljaric, N., W.A. Crawford, "Remo-
val of Metals from Laboratory Solu-
tions and Landfill Leachate by Green-
sand Filters." Univ. of Delaware,
Delaware Geol. Survey, Rept. of
Invest. No. 32 (1979a).
18. Spoljaric, N., W.A. Crawford, Removal
of Contaminants From Landfill Leach-
ates by Filtration Through Glaucona-
ted Greensands." Envir. Geol., V.2,
No. 6, pp. 359-363 (1979b).
19. USEPA, Handbook for Remedial Action
at Waste Disposal Sites. EPA 625/6-
82-006, 1982.
26
-------
MECHANISMS OF CONTAMINANT MIGRATION THROUGH A CLAY BARRIER-
CASE STUDY, WILSONVILLE, ILLINOIS
R. A. Griffin, B. L. Herzog, T. M. Johnson, W. J. Morse,
R. E. Hughes, S. F. J. Chou, and L. R. Follmer
Illinois State Geological Survey
Champaign, Illinois 61820
ABSTRACT
Routine monitoring of a hazardous waste disposal facility at Wilsonville, Illinois
revealed that organic contaminants were migrating 100-1000 times faster than pre-
dicted. A detailed hydrogeologic investigation of the site revealed that the local
groundwater flow and gradient were dominated by a 45-foot high coal mine refuse pile
that causes radial flow from the pile through the trenches. Measurements of hydraulic
head indicated that the vertical flow was much less significant than the lateral com-
ponent of groundwater flow. Results of hydraulic conductivity tests and geologic
investigations lead to the conclusion that the higher-than-predicted migration rates
could be accounted for by the differences between laboratory and field measurements of
hydraulic conductivities. These differences appear to be primarily due to the inability
of small laboratory specimens, whether "undisturbed" or recompacted, to simulate the
flow through relatively large joints or partings present in natural materials.
The distribution of contamination in the groundwater was clearly related to the
flow patterns. Time series sampling of monitoring wells in slowly recharging formations
showed that concentrations of volatile organics varied widely through the recovery
period.
INTRODUCTION
Laboratory studies have shown that
organic chemicals can increase the hy-
draulic conductivity of compacted clay
soils; several field studies have docu-
mented instances where field-measured
hydraulic conductivity of compacted clay
liners exceeded the laboratory-measured
values used to design the facility.
These and other studies have resulted in
concern about the potential failure of
clay barriers to control release of
leachate from landfills and surface
impoundments.
Interactions between clay, percola-
ting water, organic solvents, and
contaminants can result in rates of
migration greater than expected.
Anderson, Brown, and Green (2), Acar,
Oliveri, and Field (1), and Foreman and
Daniel (7) have reported that hydraulic
conductivities of soils and clay could
be greatly increased when exposed to
various organic solvents. Bridley,
Wiewiora, and Wiewiora (3) and Griffin
et al . (10), found that clay minerals
that expand in water can undergo rela-
tive collapse when exposed to organic
solvents. Murray and Quirk (15) have
shown that dry, pressed pellets of non-
expanding clays will expand when
solvated by various solvents in direct
proportion to the dielectric constant of
the solvent. These studies support a
mechanism for producing higher hydraulic
conductivities in which clay materials
exposed to water, which has a relatively
high dielectric constant, shrink when
exposed to organic solvents with a lower
constant.
Failure of clay barriers may also
result from measurement errors during
design or construction. Daniel (5)
27
-------
reported four case studies of clay-lined
impoundments where unexpectedly high
leakage rates were traced to under-
estimation of hydraulic conductivity
because laboratory measurements were
used; field-measured hydraulic con-
ductivities were found to be much
greater.
Failure of some clay barriers to
prevent leachate migration may result
from natural as well as man-induced con-
ditions. For example, evaluators of
potential landfill sites in unconsoli-
dated glacial deposits often overlook
the existing natural joint system.
Contaminated aqueous wastes and organic
solvents can move along these routes.
Additionally, potential clay-solvent
interactions that can enhance these
existing natural joints or planar
structures have usually been ignored
during landfill and surface impoundment
siting and design.
The case study reported here began
as an outgrowth of a 1982 court order to
exhume and remove the hazardous wastes
buried at a disposal facility at
Wilsonville, Illinois. The 130-acre
landfill operation was a trench-and-fil 1
procedure that basically relied on
natural attenuation of contaminants by a
clay-containing till deposit native to
the site. Routine monitoring of the
site revealed that organic contaminants
were migrating 100 to 1000 times faster
than predicted. The Illinois State
Geological Survey, supported by the U.S.
Environmental Protection Agency, the
Illinois Environmental Protection
Agency, and the site owner, SCA
Services, Inc., began a study to deter-
mine the cause or causes of contaminant
migration at the facility. Two obvious
questions were posed: (1) Why were
these organic contaminants migrating
faster than predicted, and (2) what were
the implications for land disposal of
similar wastes at other sites?
The background events, overall site
characteristics, and project description
have been published previously (9, 13,
and 21). The geology of the site has
been described by Follmer (6) and by
Griffin et al. (10).
PURPOSE
The purpose, objectives, and
approach for the project have been
described in detail in earlier pub-
lications (9 and 13).
This paper describes hydrogeologic
conditions at the site and how they
affect the distribution of contaminants
in the groundwater. To determine the
mechanism responsible for the higher-
than-expected migration rates of organic
contaminants at the site, hydraulic
conductivity was measured by several
methods in the field and in the labor-
atory, and the effect of joints in the
till on the measured values was examin-
ed. The results of these measurements
are particularly significant because
relatively few data have been published
that compare laboratory- and field-
determined hydraulic conductivity values
from actual disposal sites.
This paper also discusses results
of time series sampling of volatile
organics from monitoring wells finished
in fine-grained earth materials. This
sampling work was conducted to determine
the effect of sample collection proce-
dures on the amounts of volatile
organics detected in water samples.
RESULTS
The work described in this paper is
currently in progress. The results pre-
sented should be considered as tentative
and subject to possible reinterpretation.
HydrogeolOjjic/Geochemical Studies
Hydrogeologic and geochemical in-
vestigations of the site have involved
the completion of 17 piezometer/monitor-
ing well nests (nests A to K, profiles V
and W in Figure 1), with two to nine
piezometers and monitoring wells per
nest. The details of the design, con-
struction, and experimental plan were
published previously (9, 10, and 13).
The piezometers were used initially
for in-situ hydraulic conductivity tests
at various depths. After the water
levels stabilized, these piezometers
were used to establish the long-term
piezometric surface and, in turn, the
28
-------
Coal mine cleaning refuse
(Gob Pile)
— Water table (m), April 1984
Figure 1. Site map showing: water table elevation (m) in April, 1984;
locations of trenches, wells and cross section
hydraulic gradient and flow across the
site. The elevation of the water table
measured in all wells and piezometers at
the site, including SCA wells in April,
1984, is shown in Figure 1. The effect
of the coal refuse pile on the shallow
groundwater flow pattern is evident. A
groundwater mound is present beneath the
coal refuse pile, resulting in shallow
groundwater flow in a radial pattern
toward trench area B to the west and
trench area A to the south. Some dis-
posal trenches in each area were appar-
ently excavated below the water table;
however, the trenches may have remained
relatively dry during the disposal oper-
ation due to the very low hydraulic con-
ductivity of the surrounding glacial
till.
Figure 2 is a cross section of the
site along the line shown on Figure 1.
This cross section shows the sequence of
geologic materials in this area; these
materials have been described in detail
by Follmer (6) and Griffin et al. (10).
The cross section also shows the posi-
tion of the water table and equipoten-
tial lines as measured in April, 1984.
Measurements of hydraulic head in wells
at different depths in each cluster of
wells in Profile V and at Nest I indi-
cated a downward vertical hydraulic
gradient and groundwater flow. However,
these were much less significant than
the lateral component of groundwater
flow dominant in each geologic unit
because of the lateral hydraulic
gradient.
Hydraulic Conductivity Tests
The determination of hydraulic
conductivity at the site was an
29
-------
ft
640
630^
620
c
g
610-
600-
590-
580J
Deona Loess
Roxana Silt
Vandalia ablation
Trench Area B
till, stiff, .clayey
Basal Vandalia Till, altered, jointed
Basal Vandalia Till
unaltered
-185
-180 » »
-625- Equipotential line (ft)
0 100ft
0 30m
Figure 2. Cross section from Profile V through Trench Area B to the gob pile
showing equipotential lines of groundwater flow.
essential part of this project. There-
fore, several methods were used to
measure it.
Laboratory values of hydraulic con-
ductivity were determined from tests of
undisturbed and recompacted soil cores,
using a Harvard-type miniature permea-
meter. This method is described in
detail in Herzog and Morse (11). Addi-
tional values were available from engin-
eering reports about the site (12 and
8). Field analyses included slug tests
analyzed by the method of Cooper, Brede-
hoeft, and Papadopulos (4) and by a
method proposed by Nguyen and Pinder
(16). Recovery tests were performed on
monitoring wells that were pumped for
the purpose of well development, using
the technique described by Todd (22).
Field tests were performed on both
vertical and angle holes. In addition
to the vertical piezometers installed at
each nest shown in Figure 1, angle pie-
zometers were also installed at nests A,
B, E and K to attempt to measure the in-
fluence of vertical joints on hydraulic
conductivity.
Angle holes were drilled at a 45°
angle. The cased interval was back-
filled by pumping grout from the bottom
of the casing to the ground surface
rather than from the intervals of
expanding concrete and bentonite des-
cribed in earlier reports (9 and 13).
All other design and construction
details were identical to those
described for the vertical piezo-
meters. Construction details of the
monitoring wells and vertical
piezometers are shown in Figure 3.
Figure 4 illustrates that the angle
holes have a much greater chance of
intersecting vertical fractures because
the holes are longer and cover a greater
lateral area.
Slug tests were originally designed
to be analyzed by the method proposed by
Cooper, Bredehoeft and Papadopulos (4),
and the type curves published by
Papadopulos, Bredehoeft and Cooper
(18). This method was chosen because it
assumed boundary conditions that could
reasonably be met in the field.
30
-------
Monitoring well
Open-hole piezometer
T
: ^ expanding „
cement
, metal or
pvc casing
2%-in. ID
PVC casing
sand- bentonite
" slurry "
^ expanding ^
cement
"!
: «- sand pack
< i— well screen
Shelby tube hole —
»
•j
-»
r
Figure 3. Construction details for mon-
itoring wells and vertical
piezometers.
All slug test data were reanalyzed
using the method of Nguyen and Pinder
(16). The method is a direct calcula-
tion method that uses data obtained
during the early part of the slug test
(usually within the first half-hour).
Our slug tests were not originally de-
signed for analysis by this method, and
many of the data were less than ideal
for analysis by this method because few
early measurements had been taken. Add-
itional error may have resulted from
using discrete measurements obtained by
steel tape or electronic water level
meter instead of having continuously
recorded water levels.
The Nguyen and Pinder method is de-
signed to consider problems involving
partially penetrating wells in aquifers
for which the short-term effects of a
water table or leakage from a confining
layer can be ignored. It appears to be
appropriate for analyzing tests in
materials of moderate to low hydraulic
conductivity. The boundary conditions
(except for full aquifer penetration)
and parameters to be measured are the
same as in the Cooper, Bredehoeft, and
Papadopulos method.
The third field method used was the
recovery test (22). The recovery test,
often used after an aquifer pumping
test, applies to unsteady radial flow in
a confined aquifer, if the well fully
penetrates the aquifer and well storage
is negligible. For this study, the re-
covery test was used to analyze all
monitoring wells that were pumped for
development.
| Wall of open hole ' /rs £
Figure 4. Parameters to be measured for slug test analysis,
31
-------
Table 1 summarizes all the labora-
tory and field hydraulic conductivity
data collected to date. Data for many
individual measurements have been re-
ported previously (11). Field data are
presented as the geometric means because
hydraulic conductivity values in field
soils are generally acknowledged to be
log-normally distributed (17 and 19), so
arithmetic averaging would not be appro-
priate (14). For some formations, the
data had a range of about three orders
of magnitude. Because of the hetero-
genity of glacial tills, this variabil-
ity is not unexpected.
The number of values averaged
ranged from 1 for the Banner Formation-
bedrock contact and some tests on the
Sangamon soil, to 13 for recovery tests
in the weathered, basal Vandalia Till.
The laboratory values are reported as
ranges because they include data report-
ed by consultants who did not report
individual values (8 and 12).
As reported in Table 1, the range
of values determined in the laboratory
for tests of each type of sample, dis-
turbed and undisturbed, is small. The
undisturbed samples yielded higher
values of hydraulic conductivity than
the disturbed samples for soft sedi-
ments, but the reverse is true for the
hard tills.
Whereas the field data showed a
wide range of values, the geometric
means calculated from the data from the
three methods were nearly equal. The
most divergent value was found in the
Sangamon Soil Bt horizon for which there
was only one slug test site and one re-
covery test site and these were widely
separated (1/2 mile apart). The unalt-
ered basal Vandalia Till also shows
disagreement between slug tests and
recovery tests. For this unit, the
higher mean value from the recovery
tests can be explained by noting that
the recovery tests were performed on
monitoring wells and that three of the
nine wells used for recovery test
analysis were finished in zones that
included a sand lens in their screened
interval, thus yielding higher values
than from wells or piezometers finished
in the matrix material .
On the basis of these tests, it
appears that the various laboratory
methods used in this study can produce
comparable values of hydraulic conduc-
tivity for a given geologic material.
The various field techniques tested at
Wilsonville also yielded similar
values. However, in general, laboratory
methods produced much lower values than
field methods. Laboratory and field
results were similar only between undis-
turbed laboratory samples and slug tests
of the Sangamon Bt horizon.
Laboratory results were generally
less variable than field results. This
may have occurred because fewer samples
were run, but is probably due to a
natural bias toward selecting cohesive
samples for undisturbed analysis; crumb-
ling or fractured samples were not
chosen for testing, thus leading to more
uniform results and bias toward lower
values. Recompaction of samples also
yields more uniform results because the
original structure, including any
natural joints present, are destroyed
during the sample preparation steps.
This suggests that laboratory measure-
ments of hydraulic conductivity should
not be substituted for field measure-
ments unless caution is used in their
interpretation.
The biggest disadvantage of the
Cooper, Bredehoeft, and Papadopulos
method is the length of time necessary
to obtain results for materials that
have a very low hydraulic conductiv-
ity. Some tests in this study ran for
more than a year before water levels
stabilized. Therefore, error may be
introduced by seasonal changes in
regional water levels and by barometric
fluctuations. A second problem results
from the assumption that wells fully
penetrate the aquifer.
Recovery tests seem a reasonable
choice for monitoring wells, but it may
be difficult to meet the assumption of a
steady pumping rate, especially when the
formation has a very low hydraulic con-
ductivity. Another problem with the
recovery method is its basis on the as-
sumption that the aquifer is fully
penetrated by the well screen—an
assumption that is not usually valid.
Constructing monitoring wells to meet
32
-------
this latter condition usually reduces
their effectiveness as monitoring
wells. However, the test can be
performed relatively quickly as part of
a well development program to give
additional information on the geologic
unit that the well monitors.
The Nguyen and Pinder method offers
several advantages over the other field
methods. This method was designed for
partially penetrating wells--the most
common type of well construction. In
addition, the test usually takes less
than a day to run; therefore, it is not
influenced by seasonal water level
changes, and barometric fluctuations are
minimized. The biggest disadvantage
appears to be that small errors in
water-level measurements will affect the
results of this method more than those
from the other two field methods.
However, these errors can be minimized
by using a sensitive, continuous water-
level recorder.
Geometric mean values of hydraulic
conductivity from the angle tests were
greater than the corresponding values
from the vertical tests for both types
of slug test analyses. This was not
necessarily true for all piezometer
pairs. The mean difference ranged from
less than a factor of two for the abla-
tion zone to more than an order of
magnitude for lower basal till. This
difference was expected because angle
holes should intersect more of the pre-
dominantly vertical fractures (Figure
4). Although these results are inter-
esting, they are not conclusive because
the theory for analyzing angle piezo-
meters has not yet been fully developed
and the statistical significance of the
differences between results for vertical
and angle-holes has not been determined.
Although the Nguyen and Pinder
method for the analysis of data from
slug tests seems to be the best of the
three field methods tested, the recovery
test and the Cooper, Bredehoeft, and
Papadopulos method of slug test analysis
should not be abandoned. The recovery
test is the only one of these tests that
can be performed on individual monitor-
ing wells after pumping. The Cooper,
Bredehoeft, and Papadopulos method could
be used to supplement the Nguyen and
Pinder method because data that cannot
reasonably be analyzed using one slug
test method may be analyzed by the
other.
Contaminant Migration
The hydraulic conductivity tests
could explain the rate of contaminant
migration. However, it is equally
important to determine the pattern of
contamination. Therefore, a separate
set of monitoring wells was constructed
to obtain water samples for chemical
analyses at each of the nests and pro-
files (Figure 1). Cores from these
borings were analyzed chemically for
organic and inorganic constituents. The
results, described in Griffin et al .
(10), indicated that the zones of high-
est contaminant concentration were in
the soft ablation till and the weath-
ered, jointed basal till (Figure 5).
The results of in-situ field tests
(Table 1) indicated that the hydraulic
conductivities of the soft ablation zone
and weathered basal till are greater
than those of the underlying or over-
lying units; this probably accounts for
the higher concentrations in the soil
and apparently higher migration rates in
this zone.
The extent of contaminant migration
in the groundwater flow system at pro-
file V is illustrated in Figure 5, which
shows the concentrations of trichloro-
ethylene (yg/L) in groundwater in
September, 1984. The distribution of
contamination in this cross section is
clearly related to the measured ground-
water flow patterns in Figure 2. How-
ever, the highest levels of contamina-
tion were found in wells finished in
sand lenses within the upper part of the
unweathered, basal till zone. Contam-
inants from Trench Area B introduced
into the soft ablation till and jointed
basal till must be able to move downward
into the unweathered basal till through
interconnected joints or sand lenses.
The extent of contaminant migration in
this area is significantly affected by
the strong lateral component of ground-
water flow away from the trench area.
33
-------
Table 1. Summary of Hydraulic Conductivity Data
Geometric Means of Field Data
MATERIAL
Peoria Loess
Peoria Loess/gob
Contact
Vandalia Till Ablation
Zone, Sangamon Bt
Horizon
Vandalia Till
Lower Ablation
Zone
Base of Ablation
Zone/top of
weathered Vandalia
Weathered basal
Vandalia Till
Unaltered basal
Vandalia Till
Ranges of Laboratory Data
Recompacted Undisturbed
Samples Samples
Orientation (cm/sec) (cm/sec)
Vertical 3.3xlO~8 - 3.6xlO~8 7.5xlO~8 - l.lxlO"6
Vertical
Vertical 6xlO~10 - 1 .2xlO~9 1.6xlO"7
45° angle
Vertical 2.6xlO~8 - 3.5xlCf8 5.8xlO~7 - 1.9xlO"6
45° angle
Vertical
Vertical 1.2xlO~8 - 7.4xlO~8
45° angle
Vertical 8.9xlO"9 - 2.7xlO"8 3.3xlO"9 - 5.1xl(T9
45° angle
Slug Tests
Cooper, et.al .
Method
(cm/sec)
1.3xlO"7
3.8xlO"5
7.5xlO"5
2.4x10'*!
9.5x10"°
l.SxlO'7-
2.5x10""
Nguyen & Pinder
Method
(cm/ sec)
1.2xlO"7
6.4xlO"7
1.9xlO"5
2.4xlO"5
2.3x10"^
1.2xlO"b
8.4x10"?
9.4x10"
Recovery
Test
(cm/ sec)
4.8xlO"5
2.0xlO"6
1.6xlO"5
1.3xlO"6
4.0xlO"6
6.8xlO"7
Vandalia Till/Banner
Formation contact
Vertical
9.9x10
-7
Banner Formation/
bedrock contact
Vertical
1.1x10
-7
1.7x10
-7
-------
ft
640H
630-
620
SW
-190
c
o
610-
600-
590-
580J
'eona Loess
Roxana Silt
KI85
-180
Rtisal Vandalia Till, altered, jointed
A(§>-' Basal Vandalia Till,
%6'C unaltered
I Trichloroethylene
(Sept., 1984)
_ 10
Figure 5. Cross section from Profile V through Trench Area B to the gob pile
showing distribution and Trichloroethylene (ppb) in groundwater.
Volatile Organic Sampling in Fine-
Grained Material
A protocol has not been established
for collecting water samples from
monitoring wells finished in fine-
grained materials with slow recharge
rates for analysis of the water's
volatile organic content. Schuller,
Gibb, and Griffin (20) have recommended
evacuation of 4 to 6 well volumes to
purge the well of stagnant water prior
to collection of samples. However, this
is not possible for wells with very slow
recharge rates; most wells at the
Wilsonville site could be pumped down to
the top of the well screen after remov-
ing less than 2 well volumes. Thus, the
question arises as to what impact the
well recovery time has on the chemical
composition of water samples collected
for analysis.
To address this question, sets of
time-series water samples were collected
at several wells. Representative re-
sults obtained for o-xylene from three
wells are shown in Figure 6. The re-
sults clearly indicate that changes in
o-xylene concentration occur as a
function of well recovery time.
10
8-
6-
4-
2-
O-XYLENE, WELL V1D
10 20 30 40
Well recovery time (hours)
50
Figure 6. Volatile Organic Concentra-
tion vs. Time
35
-------
Samples were collected using a
double check valve Teflon* bailer with a
bottom draining device (Timco Mfy.,
Prairie Du Sac, Wisconsin)*. Samples
were drained into 40-mL volatile organic
analysis (VGA) vials with a Teflon^-
faced septum and a screw cap. The vials
were placed on ice immediately after
collection and transported to the lab-
oratory within 48 hours of collection.
Samples collected before purging
the well of stagnant water (time 0)
contained o-xylene concentrations that
were below detection limits at two of
the three wells for which data are shown
in Figure 6. This is probably due to
volatilization losses of the compound
resulting from the water standing in the
well. After the initial samples were
collected, the well was pumped down to
the top of the 2-foot well screen with a
bladder-pump; the pump was then removed
from the well. Samples were then col-
lected by bailer at intervals of
approximately 1, 2, 4, 6, 24, 31, 48,
124, 290, and 460 hours. Maximum
concentrations were obtained after
approximately 2-8 hours of recharge.
Concentrations fell to low values
between 24 and 48 hours of recharge.
Concentrations of o-xylene in all
samples collected after 124, 290, and
460 hours were below detection limits.
We interpret the data to indicate
that fresh formation water containing o-
xylene enters the well during the
initial hours of recharge after the well
is pumped down. The concentration of o-
xylene in the water column is then
apparently depleted over time due to
volatilization losses. The sampling
protocol can affect the results obtained
from slowly recharging wells relative to
volatile organic concentrations.
SUMMARY AND CONCLUSIONS
This paper has presented results of
three phases of the investigation of the
Wilsonville hazardous waste disposal
site: determination of the hydraulic
*Mention of trade names or commercial
products does not constitute endorsement
or recommendation for use.
conductivity of materials underlying the
site, evaluation of distribution of con-
taminants, and testing of experimental
protocols for sampling of volatile
organic compounds. The hydraulic con-
ductivity portion of the study is
essentially completed. Results of
laboratory tests of hydraulic conduc-
tivity were found to be consistently
lower than those from in-situ field
tests for most materials at this site.
This is thought to be due to the fact
that laboratory tests do not generally
indicate the effects of macrostructures
such as joints and sand lenses. In-situ
tests using angle holes were used to
assess the significance of vertical
joints. The in-situ hydraulic conduc-
tivity tests and the geologic investiga-
tion both revealed that the upper
portion of the glacial till underlying
the trenches was highly jointed. The
differences between predicted and actual
migration rates, therefore, can be ex-
plained by the fact that only laboratory
determined values of hydraulic conduc-
tivity were used for the predictions.
Groundwater sampling revealed that
the highest degree of contamination was
within the unaltered, basal Vandalia
Till. The depth of contaminant migra-
tion was greater than expected based on
measurements of hydraulic gradient and,
as such, the pathway is uncertain. The
pathway may involve migration through
joints and along shear planes, basically
controlled by the fracture network.
Flow may also have occurred through sand
lenses and through the fine-grained
matrix.
Time-series sampling of monitoring
wells prior to, during and after pumping
revealed that o-xylene concentrations
reached a maximum after 2 to 8 hours of
recharge to the well. Additional stud-
ies of the behavior of other compounds,
currently underway, are necessary before
a protocol can be proposed for sampling
of volatile organics in fine-grained,
low-permeability materials.
ACKNOWLEDGMENTS
The authors wish to acknowledge
partial support of this project by SCA
Chemical Services, Inc., Wilsonville,
Illinois; the Illinois Environmental
36
-------
Protection Agency; and the U.S. Envir-
onmental Protection Agency, Cincinnati,
OH, under cooperative Agreement No.
R810442-01; Dr. Michael Roulier is the
U.S. EPA project officer.
The authors also gratefully acknow-
ledge the contributions of C. J. Stohr,
D. N. Cote, W. J. Su, P. B. DuMontelle,
and K. Cartwright to this manuscript.
REFERENCES
1. Acar, Y. B., I. Olivieri, and S. D.
Field. 1984. Organic fluid ef-
fects on the structural stability
of compacted kaolinite. _In_ Pro-
ceedings of the Tenth Annual Re-
search Symposium, Ft. Mitchell, KY,
April 3-5, 1984; U.S. Environmental
Protection Agency, Cincinnati, OH,
EPA-600/9-84-007.
2. Anderson, D., K. W. Brown, and J.
Greene. 1982. Effect of organic
fluids on the permeability of clay
soil liners. _In_ Land Disposal of
Hazardous Waste, EPA-600/9-82-002,
U.S. Environmental Protection
Agency, Cincinnati, OH, pp!79-190.
3. Brindley, G. W., K. Wiewiora, and
A. Wiewiora. 1969. Intracrystal-
line swelling of montmorillonite in
some water-organic mixtures (clay-
organic studies. XVII): Amer.
Mineralogist, V. 54, pp!635-1644.
4. Cooper, H. H., J. D. Bredehoeft,
and I. S. Papadopulos. 1967. "Re-
sponse of a finite-diameter well to
an instantaneous charge of water,"
Water Resources Research. 3:1,
pp263-269.
5. Daniel, D. E. 1984. Predicting
hydraulic conductivity of clay
liners. ASCE Journal of Geotech-
nical Engineering, 110:2, pp285-
300.
Follmer, L. R. 1984. Soil - an
uncertain medium for waste dispos-
al . Seventh Annual Madison Waste
Conference Proceedings, University
of Wisconsin-Extension, Madison,
WI, pp296-311.
7. Forman, D. E. and D. E. Daniel.
1984. Effect of hydraulic gradient
and method of testing on the hy-
draulic conductivity of compacted
clay to water, methanol , and hep-
tane. U.S. Environmental Protec-
tion Agency, Cincinnati, OH, EPA-
600/9-84-007.
8. Geoengineering, Inc. 1982. Pro-
posed remedial action plan, Earth-
line Site, Wilsonville, Illinois,
SCA Chemical Services, Inc.,
Boston, MA, 22pp with appendices.
9. Griffin, R. A., K. Cartwright, P.
B. DuMontelle, L. R. Follmer, C. J.
Stohr, T. M. Johnson, M. M. Killey,
R. E. Hughes, B. L. Herzog, and W.
J. Morse. 1983. Investigation of
clay soil behavior and migration of
industrial chemicals at Wilson-
ville, Illinois. U.S. Environmen-
tal Protection Agency, Cincinnati,
OH, EPA-600/9-83-018, pp70-79.
10. Griffin, R. A., R. E. Hughes, L. R.
Follmer, C. J. Stohr, W. J. Morse,
T. M. Johnson, J. K. Bartz, J. D.
Steele, K. Cartwright, M. M.
Killey, and P. B. DuMontelle.
1984. Migration of industrial
chemicals and soil-waste inter-
actions at Wilsonville, Illinois.
U.S. Environmental Protection
Agency, Cincinnati, OH. EPA-600/9-
84-007.
11. Herzog, B. L and W. J. Morse.
1984. A comparison of laboratory
and field determined values of
hydraulic conductivity at a waste
disposal site. Seventh Annual
Madison Waste Conference Proceed-
ings. University of Wisconsin-
Extension, Madison, WI, pp30-52.
12. John Mathes and Associates.
1976. Report of subsurface inves-
tigations: proposed site of
industrial residue management
facility, Earthline Corporation,
Wilsonville, Illinois. 38pp.
37
-------
13. Johnson, T. M., R. A. Griffin,
K.Cartwright, L. R. Follmer, B. L.
Herzog and W. J. Morse. 1983.
Hydrogeologic investigations of
failure mechanisms and migration of
organic chemicals at Wilsonville,
Illinois. National Water Well
Association Third National Sympos-
ium and Exposition on Aquifer
Restoration and Ground Water Moni-
toring, May 25-27, 1983, at
Columbus, OH.
14. Keisling, T. C., J. M. Davidson, D.
L. Weeks and R. D. Morrison. 1977.
Precision with which selected soil
parameters can be estimated. Soil
Science. 124:4, pp241-248.
15. Murray, R. S. and J. P. Quick.
1982. The physical swelling of
clays in solvents; Soil Sci. Soc.
Am. J., V. 46, pp865-868.
16. Nguyen, V. and G. F. Pinder.
1984. Direct calculation of
aquifer parameters in slug test
analysis In J. Rosensheim and G.
D. Bennet feds.), Groundwater
Hydraulics . American Geophysical
Union Water Resources Monograph 9.
pp222-239.
17. Nielsen, D. R., J. W. Biggar and K.
T. Erh. 1973. Spatial variability
of field-measured soil-water
properties. Hilgardia, 42:7,
pp215-259.
18. Papadopulos, S. S., J. D.
Bredehoeft and H. H. Cooper.
1973. On the Analysis of 'Slug
Test' Data, Water Resources
Research, 9:4, pp!087-1089.
19. Rogowski, A. S. 1972. Watershed
physics: soil variability
criteria. Water Resources
Research, 8:4, ppl015-1023.
20. Schuller, R. M., J. P. Gibb and R.
A. Griffin. 1981. Recommended
sampling procedures for monitoring
wells. Ground Water Monitoring
Review 1:42-46.
21. Stohr, C. J. 1983. Applications
of close-range photogrammetry for
geologic investigations during the
exhumation of a hazardous-waste
disposal site. Technical Papers of
the 49th Annual Meeting of the
American Society of Photogrammetry,
March 13-18, 1983, at Washington,
D.C.
22. Todd, D. K. 1980. Groundwater
Hydrology (2nd edition). John
Wiley and Sons, 535pp.
38
-------
RECLAMATION AND REDEVELOPMENT OF UNCONTROLLED HAZARDOUS WASTE SITES
Garrie L. Kingsbury and Robert M. Ray
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
ABSTRACT
There are numerous cases in the United States where uncontrolled dumping or
industrial spills have contaminated properties with hazardous materials (now more than
18,000 sites inventoried by U.S. EPA). Since many of these properties are in prime urban
locations, issues surrounding the reclamation and redevelopment of contaminated properties
have assumed national importance. The principal objective of this study has been to
document with case studies relationships between site remediation methods, cleanness
criteria, and redevelopment land uses.
After extensive interviews with Federal and State officials in all 50 states, 12
uncontrolled hazardous waste sites were selected for detailed study. For each of these
sites, remedial actions have been undertaken with some upgraded redevelopment of the
property in mind. Redevelopments include single- and multi-family residential,
recreational, commercial, institutional, and light industrial land uses.
Two distinctly different types of redevelopment efforts were encountered--
public-initiated projects and developer-initiated projects. In the case of
public-initiated projects (for example, most Superfund sites), immediate concerns for
community health are paramount, and site reuse, if any, tends to be incidental to site
cleanup.
In the case of developer-initiated projects, the developer attempts to recover site
cleanup costs through resale of the property. Thus, he simply diverts into cleanup
operations money that would otherwise be used to purchase uncontaminated land. The
economic feasibility of a developer-initiated project may depend directly on the standard
of cleanness required of a site for a particular redevelopment type. Since property
decontamination standards and guidelines have not been formulated for most situations,
some confusion exists and hence, developers generally view contaminated site reclamation/
redevelopment projects as risky ventures.
INTRODUCTION
In May 1983, RTI began a study to
identify and document specific instances
where uncontrolled hazardous waite sites
had been cleaned up and redeveloped for
some other, upgraded land use. One focus
of the study involved careful analysis of
the various standards and criteria for
cleanup that have been or could be
employed in these remediation/
redevelopment efforts. It is anticipated
that through this ongoing effort a more
complete understanding of relationships
between site reuse, remediation, and
cleanup criteria will emerge to benefit
reclamation/reuse efforts.
BACKGROUND AND PURPOSE
Prior to 1976, few states had regula-
tory programs for land disposal of hazard-
ous wastes. However, national awareness
of hazardous waste problems increased
dramatically in the mid-to-late 1970's as
it became evident that mismanagement of
hazardous wastes could lead to the release
of toxic materials into the soil, water,
and air. Congress responded to the
concern over uncontrolled hazardous wastes
by enacting the Comprehensive Environ-
mental Response, Compensation, and
Liability Act of 1980 (CERCLA), better
known as Superfund. Under the CERCLA
program, 538 sites with hazardous waste
39
-------
contamination are included as of October
1984 in the Superfund National Priority
List (NPL) with an additional 238 sites
proposed for inclusion (U.S. EPA, 1984a
and b). The extent of the problem is much
larger, however. To date, the U.S.
Environmental Protection Agency (EPA) has
inventoried more than 18,000 uncontrolled
hazardous waste sites.
In view of the large number of uncon-
trolled hazardous waste sites in the U.S.
and the effort required to reclaim this
land for productive use, issues related to
uncontrolled hazardous waste site
remediation and redevelopment are of
national significance. However, because
site remediation and reuse are relatively
new public concerns, a great deal of
information has not previously been
assembled to describe instances where site
redevelopment has occurred following site
cleanup. This study attempts to make a
beginning in this new area.
APPROACH
The project consists of three compon-
ents: (1) identifying suitable sites for
analysis, (2) preparing detailed case
studies on selected sites, and (3)
determining what criteria exist for guid-
ing the extent of site cleanup that is
necessary. Through all three of these
study components, the emphasis is on the
functional relationships among site
remediation methods, cleanup criteria, and
redevelopment objectives.
To identify sites that might serve as
case studies for remediation and reuse, we
conducted telephone and personal inter-
views with Federal and State officials
involved with uncontrolled hazardous waste
site cleanups within the U.S. In addition
to contacting environmental officials in
all 50 States, we have solicited
information from each of the ten EPA
Regional Offices and each of the ten EPA
Regional Superfund Offices. Inquiries
have also been made to the U.S. Army Toxic
and Hazardous Materials Agency (USATHAMA),
the U.S. Army Corps of Engineers, and the
Regional Offices of the U.S. Department of
Housing and Urban Development. This
approach has been highly successful and
has led to the identification of sites for
detailed study. A carefully designed
literature search of 13 online data bases
was also conducted, but proved to be of
limited value, since there is rarely any
mention of site redevelopment in journal
and newspaper articles dealing with
hazardous waste sites.
In many cases, we were able to
obtain copies of pertinent project
records, local newspaper articles, and
various types of government and contractor
reports pertaining to specific sites that
were identified through telephone
contacts. These materials were reviewed
carefully to decide if a particular site
was appropriate for further study.
Several sites were identified as
candidates for detailed case studies. An
effort was then made to assemble complete
files on these sites to describe site
location, ownership, and special charac-
teristics; land use history and redevelop-
ment objectives; the nature of the
contamination; the remediation actions
taken or planned; and the specific
criteria used to guide the cleanup effort.
This detailed information gathering
required additional telephone interviews
and, in some cases, site visits. Site
visits proved to be highly effective in
assembling the background documentation
for several sites.
We discussed the "how-clean-is-
clean?" issue with numerous environmental
officials involved in site remediation and
reviewed the criteria that were applied
for the sites chosen as case studies for
the project. In addition to the criteria
based on Federal guidelines for air and
water, we sought other types of quantita-
tive guidance, particularly standards
applicable to soils, that could be applied
to site cleanup. These various standards
and criteria were compiled to facilitate
comparison with the types and levels of
contaminants encountered at hazardous
waste sites. Information for this work
was culled from many sources including
various Regional and Program Offices
within EPA, the U.S. Army Medical
Bioengineering Research and Development
Laboratory, the Centers for Disease Con-
trol, and the British Department of the
Environment.
RESULTS OF THE STUDY
Site Redevelopment Case Studies
During the initial phase of the
study, 12 uncontrolled hazardous waste
sites located in seven states have been
addressed as detailed case studies to
document land reuse. Additional sites
where reuse is being planned have been
identified for further investigation.
40
-------
Case study sites include former DOD
properties, a defunct coal gasification
site, abandoned chemical recovery and drum
recycling facilities, a former steel mill,
munitions and fertilizer manufacturing
sites, and several uncontrolled dump
sites. Land reuses at these sites include
industrial parks, a marina, a recreation
park, a hotel and convention complex,
single family residences, a public school,
residential condominiums, a housing
complex for handicapped and elderly, a
neighborhood playground, and State offices
and facilities. Documentation of these
redevelopment case studies has been the
major focus of the project. Brief
descriptions of the sites that were
examined in detail are provided in
Table 1.
Six of the land reuse case studies
are located in California. The remedia-
tion/redevelopment efforts at all the
California sites that we investigated were
initiated by developers seeking to upgrade
the land use. Concerns at these sites
pertained mainly to potential exposure of
persons who might live or work at the site
following redevelopment. In most cases,
material and soil that was determined by
the California Department of Health
Services (CDHS) to be hazardous was
removed to a permitted disposal facility.
All of the sites that we investigated in
California are located near large
metropolitan areas.
In Hercules, CA, the former site of
the Hercules Powder Works which manufac-
tured dynamite and other munitions from
1912 until 1963, three cases of successful
redevelopment were examined. A
residential subdivision (single family
homes) with a public school was developed
as Bayside Village on the southernmost
portion of the former Hercules property
following a very stringent remediation
effort in 1981. Cleanup operations on a
second tract of the property were
completed in 1983 by Bio-Rad Laboratories,
and an industrial park is currently being
developed there. Another tract of 50
acres was also cleaned up in 1983 to make
way for the residential condominiums known
as Hercules Village.
Residential condominiums also have
been developed at former uncontrolled
hazardous waste sites in Huntington Beach
and Yorba Linda, CA. Contamination at
these sites stemmed from dumping of
refinery wastes including both acid and
alkali sludges. The removal operations at
these sites were complicated by extensive
foul odors from sulfur compounds that were
released from the petroleum waste during
excavation. In the interest of expediency
because of the odor problem, the extent of
the removal that was required was left to
the discretion of the CDHS representative
onsite during the excavation. All waste
material and contaminated soil were
removed to a landfill permitted to receive
hazardous waste.
In South San Francisco, CA, a former
steel mill and fabrication plant site has
been redeveloped as "The Gateway", a hotel
and convention center complex, by Homart
Development Company. The remediation
agreement stipulated that the location of
the contaminated soils be clearly
designated on a site map and that these
areas not be excavated or substantially
disturbed in the future without CDHS
approval. A deed restriction was
negotiated as a way of enforcing these
provisions over time.
The Dade County Transit Authority has
plans for a maintenance facility at the
former Miami Drum Services site in Miami,
Florida. The contamination resulting from
the drum recycling operation caused major
concern because the Biscayne Aquifer which
supplies the drinking water for Dade
County lies only one meter below the
natural ground surface at the site. The
extent of the initial soil excavation was
based on the following criteria:
• Soils obviously contaminated as
indicated by the total metals
analysis were removed. If
levels were in excess of ten
times "minimum criteria" for
groundwater, the soils were
generally considered to be
contaminated.
t Soils with highly colored, oily
deposits as indicated by visual
inspection of corings were
removed.
• Known locations that received
runoff from the drum-washing
operations were excavated.
The criterion for removal based on
ten times "minimum criteria" refers to
guidelines established by the State of
Florida for groundwater. These guidelines
41
-------
are consistent with the National Primary
Drinking Water Standards for barium,
cadmium, chromium, lead, silver, and
selenium, but are more stringent for
arsenic and mercury. Although total metal
concentrations in the soil were used as
guidance in the initial excavations, final
excavations were guided by the results of
chemical tests together with engineering
and scientific judgment.
The former Chemical Metals Indus-
tries, Inc., in Baltimore, Maryland, also
presented an immediate potential hazard.
In this case, the major concerns included
imminent threat of fire or explosion in
the residential neighborhood due to the
chemical incompatibilities of the
materials present and to the potential
posed by runoff from the site.
Remediation and redevelopment at the
Kapkowski Road site in Newark, New Jersey,
is just getting started. Pockets of
PCB-laden oil are being eliminated through
a series of oil recovery wells. The site
was used from the mid-1950's as an
uncontrolled dump site for solid refuse
and waste oil. The property, adjacent to
the Newark Airport, is a prime location
for development as an industrial park.
The extent of the remediation effort that
will be performed there will be determined
when excavation begins for building.
During construction if soils are
encountered that contain more than 5 ppm
of PCB's, the contaminated material will
be removed to a permitted disposal
facility.
At Plattsburgh, New York, a recrea-
tion park now occupies a site where a coal
gas generating plant was operated from
1896 until 1960, most recently by the New
York State Electric and Gas Corporation
(NYSEG). Large quantities of coal tars
stored in unlined ponds resulted in the
contamination. The site remediation
consisted of containment onsite and a
cement-bentonite partial cutoff wall to
arrest any further migration of the
contamination into the Saranac River. In
allowing the material to remain onsite,
the New York State Department of Environ-
mental Conservation (NYSDEC) has imposed
certain restrictions to development.
Additional restrictions are also under
consideration. These include the
following:
• Sale of the lands on which the
containment cell was constructed
is prohibited by NYSDEC.
• No structures or other
activities may be placed or
performed on the containment
cell that could result in
rupture to the Hypalon cap.
• All trees and shrubs will be
maintained at a distance from
the slurry walls such that their
mature drip line will not
intersect the slurry wall.
t All construction on or near the
cement-bentonite partial cutoff
wall and/or groundwater
collection system must have
prior engineering approval of
NYSEG.
Among the U.S. Department of Defense
(DOD) sites where remediation and re-
development have been undertaken is the
Frankford Arsenal site located in eastern
Philadel- phia on the Delaware River. For
more than 150 years, this 110 acre site
was associ- ated with Federal munitions
research, development and production.
When the U.S. Army decided to excess the
facility in 1976, USATHAMA assumed
responsibility for the site cleanup to
satisfy the require- ments of the General
Services Adminis- tration prior to sale of
the property to private developers. A
large portion of the old arsenal has been
sold to a development consortium and will
probably be used by multiple tenants for
light industry. The property closest to
the water is intended for use as a
regional marina and park to be built by
the Pennsylvania State Fish Commission.
Projected for completion in 1986,
development of this 18-acre facility is
expected to cost $3 million.
General Observations
From the record searches and inter-
views, several findings of a general
nature emerged. One observation was
simply the small number of Superfund-
assisted sites that have become available
for redevelopment. Many uncontrolled
hazardous waste sites do not pose an
immediate threat to public health or
welfare as long as they are left undis-
turbed and thus do not meet the criteria
to qualify them as NPL sites. As a
result, cleanup must be initiated by the
local government or by the private sector
if site remediation is to be accomplished.
Where remediation costs are high,
42
-------
remediation/reuse will be undertaken by
the private sector only in areas where
land values are so high that cleanup costs
can be recovered through rents or resale.
Across the U.S., several contaminated
DOD sites have been remediated and turned
back to the private sector during the past
few years. Efforts to restore hazardous
DOD or former DOD properties contaminated
by hazardous wastes, unexploded ordnance,
and unsafe or unsightly debris will expand
under the Defense Environmental
Restoration Program (DERP) authorized by
Public Law 98-212. DERP covers both
active installations and formerly used DOD
properties.
A central issue for the planning of
any site redevelopment is the criteria to
be used in determining the extent of
cleanup that is to be required. Accept-
able concentration limits to establish the
extent of cleanup that is necessary to
protect public health and welfare have not
been determined for most toxic substances
of concern. A recent GAO report notes
that "Superfund provides that long-term
remedies be cost-effective, but no
standards exist that specify to what
extent sites must be cleaned up to effect
permanent remedy" (GAO, 1984).
The "how-clean-is-clean?" question is
often posed in relation to the National
Contingency Plan under CERCLA. Although
EPA has received considerable comment
regarding the need for allowable levels of
release from sites following remediation,
the Agency has maintained that the flexi-
bility in the current approach is appro-
priate because of the site-specific nature
of most problems and the need to move
ahead in an expeditious fashion.
Almost all States have problems deal-
ing with the "how-clean-is-clean?" issue.
Since different uses may imply a need for
different cleanup criteria, this type of
judgment must be made on a case-by-case
basis. Residential use is generally felt
to require the most stringent cleanup.
Recognized guidelines pertaining to
acceptable pollutant levels include air
quality standards, occupational exposure
guidelines, drinking water standards, and
water quality criteria. Although devel-
oped for other purposes, these established
sets of guidelines are frequently relied
upon as criteria for cleanup.
California appears to be ahead of
other states with respect to policy and
legislation guiding the redevelopment of
contaminated land. Within CDHS, a system
has evolved during the last 4 years for
regulating the cleanup and redevelopment
of abandoned industrial sites and waste
disposal areas.
Assembly Bill 2370 (AB 2370) is an
important feature in the California
program for hazardous waste site
remediation. This law authorizes CDHS to
impose deed restrictions on any tract of
land (and surrounding properties) that
poses a significant threat to the public
health. Though seldom carried to its
fullest extent (which requires a public
hearing), the existence of AB 2370 gives
CDHS much voice in the cleanup of Califor-
nia hazardous waste sites to ensure that
adequate cleanup levels are achieved.
California has also developed
criteria to explicitly define hazardous
wastes. The guidelines contained in the
Draft California Assessment Manual (CAM),
referred to as the CAM Standards, have
been revised several times but are now
officially filed, effective October 27,
1984, as formal definitions of "hazardous"
and "extremely hazardous" wastes. The CAM
Standards were developed to enable a
generator to determine if the waste he
produces must be managed as a hazardous
material. The standards are also applied
as criteria to establish the extent of
remediation necessary to insure that a
site is clean enough to be redeveloped for
an upgraded use.
CONCLUSIONS
The first significant finding of our
study is that there are many instances in
the United States where uncontrolled
hazardous waste sites have been or soon
will be redeveloped for some upgraded land
use. This trend is expected to increase
in the future as more sites are remedi-
ated. Sites that are redeveloped follow-
ing hazardous waste cleanups are not
easily documented through references to
the open literature, for several reasons.
First, the brief history of Federally-
supported waste site redevelopment efforts
in this country has not yet resulted in a
long list of completed projects. Second,
the substantial costs and investment risks
associated with hazardous waste site
cleanup operations appear to have
discouraged developers from attempting to
reclaim many contaminated sites. Third,
the delays associated with decisionmaking
may stifle redevelopment projects where
-------
hazardous wastes are involved. When
developers do become involved with an
uncontrolled hazardous waste site, they
try to perform the necessary remediation
to the satisfaction of the regional and
local authorities with little publicity if
possible.
The uncontrolled waste site develop-
ment projects that we have found appear to
be of two distinctly different types. The
first type may be termed the developer-
initiated redevelopment effort. Such
cases of site cleanup and redevelopment
occur in large metropolitan regions and
other areas where the locational advan-
tages of a site alone are so great that
cleanup costs can be recovered through
future resale of the remediated property.
In such cases, the decision to remediate
and redevelop a specific site is made in
the private sector, and the public sector
simply regulates and certifies the cleanup
process. Such was the case for all of the
examples of site redevelopment we identi-
fied in California.
The second type of hazardous waste
site redevelopment project is the
public-initiated project where reuse of
the land is clearly secondary in impor-
tance to the site remediation that is
required for public health and safety.
This appears to be the case, for example,
with almost all sites on the NPL. For
most public-initiated cleanup operations,
remediation activities are so complicated
and costly that the economic value of the
site following cleanup only partially (if
at all) justifies the cleanup operation.
Where a remediated site passes into public
ownership, reuse will probably be deter-
mined by the specific property needs of
the governing body at that particular
location and point in time.
It is important to note that these
two types of hazardous waste site redevel-
opments result from two entirely different
motivating forces. In the first case
(developer-initiated cleanups), the
developer is simply responding to land
market forces and diverting into cleanup
operations dollars that would otherwise be
used to purchase uncontaminated land. In
the second case (public-initiated
cleanups), the redevelopment decision is
made in the public sector, and there is no
explicit requirement that cleanup costs be
recovered through future use of the
property.
Most of the redeveloped sites that we
have identified are of the developer-
initiated type. Only two sites in reuse
were identified that involved Superfund.
The complexity of the legal process in
dealing with site cleanup is not conducive
to deliberate redevelopment efforts. Even
in cases where emergency remedial response
actions have been completed, it appears
that the site may remain in receivership
and go unused for long periods of time
(typically several years) while the courts
decide cost recovery and/or property
ownership issues.
California appears to lead other
States in the formulation and enactment of
legislation and regulations pertaining to
the cleanup and redevelopment of proper-
ties contaminated with hazardous waste.
With adoption of the CAM Standards,
California has begun to define quantita-
tively what is meant by "hazardous waste
contamination." Their program for guiding
redevelopment of contaminated land appears
to be the most advanced State program in
the Nation.
Based on our investigation to date,
we believe that planning efforts involving
uncontrolled hazardous waste site mitiga-
tion and reuse can benefit from the
learning experiences of others involved in
similar activities. Therefore, an ongoing
effort to assemble the type of information
provided through this study could provide
a valuable source of information for
Federal, State, and local authorities.
Such information would also be of great
value to developers in the private sector
who, having more knowledge of actual
procedures required to decontaminate prop-
erties, might be persuaded to get involved
in such activities.
ACKNOWLEDGMENTS
The authors wish to acknowledge RTI
Economist, Mr. Richard Harper, for his
assistance in information gathering during
the initial phase of the project. We also
acknowledge EPA staff members—Project
Officer, Ms. Norma Lewis; Mr. Don Sanning,
Program Manager for Remedial Action
Investigation; and Ms. Naomi Barkley, who
has recently replaced Ms. Lewis as the
official Project Officer—for their
guidance and helpful comments throughout
the study. We are very grateful for the
cooperation shown by the various local,
State, and Federal officials who have
44
-------
answered our questions and provided
information on specific sites.
REFERENCES
California Department of Health Services.
1983. Initial statement of reasons
for proposed regulations, criteria
for identification of hazardous and
extremely hazardous wastes.
September 1983.
GAO: General Accounting Office, 1984.
EPA's efforts to clean up three
hazardous waste sites. Report by the
U.S. General Accounting Office to the
Chairman, Subcommittee on Commerce,
Transportation and Tourism, House
Committee on Energy and Commerce.
GAO/RCED-84-91. June 7, 1984.
Ray, R. M., and G. L. Kingsbury, 1984.
Feasibility of site usage following
hazardous waste remedial actions.
Interim Report. Prepared for U.S.
EPA by Research Triangle Institute,
Research Triangle Park, N.C., EPA
Contract No. 68-03-31490-23-1.
November 1984.
U.S. EPA, 1984a. Amendment to National
Oil and Hazardous Substance
Contingency Plan; National Priorities
List. Final Rule. 40 CFR, Part 300.
Federal Register. Vol. 49., No. 185,
September 21, 1984, p. 37073.
U.S. EPA, 1984b. Amendment to National
Oil and Hazardous Substance
Contingency Plan: The National
Priorities List. Proposed Rule. 40
CFR, Part 300. Federal Register,
Vol. 49, N. 200, October 15, 1984, p.
40320.
45
-------
TABLE 1. REDEVELOPMENT SITES SELECTED FOR CASE STUDIES'
Site and Reuse Description
Nature of Contamination
Remediation
Bio-Rad Laboratories
175 acres
Hercules, CA
Reuse: Industrial Park
Hercules Village
50 acres developed by
D & S Company
Hercules, CA
Reuse: Residential condominiums
Bayside Village
100 acres
Developed by Citation Builders
Hercules, CA
Reuse: Single-family residences
and public school
Bolsa Chica
12.5 acres
Mola Development Corporation
Huntington Beach, CA
Cost of Cleanup: >$5 million
Reuse: Residential Condominiums
(224 units)
The Gateway
117 acres
Developed by Homart Develop-
ment Company
South San Francisco, CA
Reuse: Hotel and Convention
Center
Formerly used for TNT production by
Hercules Powder Works; numerous
deteriorating metal drums and soil
contaminated with heavy metals (Pb,
Zn, Cu, Cd) and explosives.
Formerly used as wastewater treat-
ment area by Hercules Powder Works;
contaminated with As, Pb, Cr, Zn,
DNT, and DNB.
Formerly Hercules Powder Works
property. Surface and subsurface
soil contained Pb at levels as
high as 530 ppm.
Site of uncontrolled dumping of
refinery sludges (>l,529 cu m)
prior to 1963; permitted Class III
landfill from 1963. Contaminated
with acidic petroleum wastes
(aliphatic and aromatic hydro-
carbons, amines, thiophenes);
methane and odor.
Formerly a steel mill and fabri-
cation plant using coal-fired
open hearth furnace. Soils contam-
inated by heavy metals, oils, and
acids deposited over 75 years.
Localized high concentrations of As,
Cr, Zn, Cu, Ni, low pH, and PCB's.
Drums removed; contaminated soils exca-
vated. More than 1,835 cu m (>2,400 cu
yd) of soil and debris removed to a per-
mitted hazardous waste disposal facility.
Almost 7,646 cu m (10,000 cu yd) of
soil removed to a permitted hazardous
waste disposal facility; some residues of
DNT and DNB left onsite and covered by
more than 3 m (10-12 ft) of clean fill in
area restricted from residential develop-
ment; 2-year monitoring program.
All contaminated soil removed.
All 45,873 cu m (60,000 cu yd) contami-
nated material removed to a permitted
hazardous waste disposal facility.
Due to large volume, majority of contami-
nated materials contained onsite and
capped. More than 700 cu m of hazardous
material removed to offsite disposal
facility.
-------
TABLE 1. (Continued)
Site and Reuse Description
Nature of Contamination
Remediation
Kellog Terrace
21 acres
Gfeller Development Company
Yorba Linda, CA
Reuse: Residential Condominiums
(224 units)
Annapolis Road Sites
Baltimore, MD
Cost of Cleanup: >$32S,000
Reuse: Maryland Department of
Health Office;
Neighborhood Park
Miami Drum Site
1 acre
Miami, FL
Cost of Cleanup: $1.6 million
Reuse: Dade County Transit
Authority Maintenance Facility
Kapkowski Road Site
West Area--134 acres
East Area--lll acres
Elizabeth, NJ
Reuse: Elizabeth Industrial Park
by the Port Authority of New
York and New Jersey
Site of sand and gravel excavation
prior to 1940; uncontrolled dumping
of refinery waste and used drilling
muds in 1940's. Contaminated from
Pb, As, and aliphatic and aromatic
hydrocarbons.
Sites used by Chemical Metals Indus-
tries, Inc. for recovery of precious
metals from waste chemical solutions
and printed circuit boards. Large
number of leaking drums (corrosive
liquids, ammonia- or cyanide-bearing
materials, organic solvents); storage
tanks, unstable powdered and solid
zirconium metal; high levels of Cd
in soil. Contaminated groundwater.
Used for 15 years by Miami Drum Ser-
vices for drum recycling; 400 to 500
used drums left onsite. Surface and
subsurface soils and groundwater
contaminated from spills—phenols,
heavy metals, oil and grease, and
pesticides.
Uncontrolled dump site from mid-
1950's to 1972. Contamination from
waste oil containing PCBs up to
4,800 ppm. Between 1 and 3 million
gallons of oil floating on water
table and adhering to materials in
refuse, primarily in the west area.
Approximately 7,034 cu m (9,200 cu yd) of
contaminated soil and fill removed to a
permitted hazardous waste disposal
facility.
More than 1,500 plastic and metal drums
removed. Waste oil, water and gasoline
pumped from underground tanks. Walls
decontaminated by sandblasting. Approxi-
mately 91 metric tons (100 tons) of con-
taminated soil and debris and 175 drums of
solids and sludges removed to a permitted
hazardous waste disposal facility. Above-
ground tanks emptied and removed. Zircon-
ium metal removed.
Hazardous debris and contaminated soil
(7,335 cu m) removed to permitted hazard-
ous waste disposal facility. Marginally
contaminated soils left in place. Ground-
water treated onsite.
Pump out free floating oil from pockets;
leave refuse in place. Soils with PCB's
greater than 5 ppm to be disposed to a
permitted facility if such soil is discov-
ered during excavation for building
foundations.
-------
TABLE 1. (Continued)
Site and Reuse Description
Nature of Contamination
Remediation
New York State Electric and Gas
Corporation (NYSEG) site
11 acres
Plattsburgh, NY
Reuse: Recreation Park on
the Saranac River
Frankford Arsenal Site
110 acres to be developed by
Shetland Properties; 18 acres
to be developed by Pennsyl-
vania State Fish Commission
Philadephia, PA
Cost of cleanup: $8 million
Reuse: Regional Park and
Marina; Light Industrial
Development
The Courtyard
Developed by Vermont Associates
Winooski, VT
Cost of Cleanup: $15,000
Reuse: Housing Project for the
Handicapped and Elderly
Formerly a coal gasification plant
operated by NYSEG. Soil and sedi-
ment from the Saranac River contam-
inated with coal tar leached from
unlined coal tar ponds.
Formerly used by the U.S. Army
for munitions manufacture and
research development of armaments.
Low-level radiological contamina-
tion in buildings, deposits of
explos i ve/pyrotechni c res i dues,
unexploded ordnance, inorganic
chemical residues in buildings and
underground waste discharge system.
Warehouse used for storage of
industrial chemicals; formerly
occupied by a silk-screening firm.
Approximately 5.66 cu m (200 cu
ft) of solid chemical wastes
beneath wood flooring; traces of
organic solvents, oil.
Subsurface migration of tar from original
disposal ponds arrested by soil-bentonite
slurry wall. Contaminated sediment from
river bed excavated and removed to con-
tainment area surrounded by slurry wall.
Containment area capped with Hypalon
liner. Cement-bentonite cutoff wall con-
structed to prevent migration of
uncontained coal tar into the river.
Explosives residues destroyed. Radiologi-
cal and heavy metal contamination removed
from the site.
Underground tanks emptied; 84 metric tons
(93 tons) of solid inorganic waste and
contaminated wood flooring removed to a
permitted hazardous waste disposal facil-
ity. Vapor barrier installed and covered
by 4 inches of concrete.
Information summarized from case studies in Ray and Kingsbury (1984),
-------
REMEDIAL ACTION MODELING OF THE WESTERN PROCESSING
AND OCCIDENTAL-LATHROP SITES
by
David B. Watson
Fredrick W. Bond
Stuart M. Brown
ABSTRACT
Data from the Western Processing and Occidental-Lathrop hazardous waste sites were
used to demonstrate, and to the extent possible, confirm, the utilization of numerical
computer models for the evaluation of remedial actions at hazardous waste sites. Ground-
water flow and contaminant transport models of the sites were developed and calibrated
against existing data. The Western Processing model was used to assess and rank possible
remedial actions for the site.
The Occidental-Lathrop site model was used to evaluate the effectiveness of an in-
stalled ground-water extraction, treatment, and reinjection system. Model results helped
to confirm the use and limitations of numerical models for the evaluation of hazardous
waste sites.
INTRODUCTION
Proper design and operation of
remedial action systems at hazardous waste
sites is crucial in terms of human
health/environmental risk reduction,
costs, design life, and probability/
consequences of failure. Assessment of
remedial action alternatives and
evaluation of the effectiveness of
installed systems are important to. insure
proper design and efficient operation.
Techniques for evaluation of remedial
action systems include desk-top procedures
for screening remedial action alter-
natives, and complex, numerical computer
models for detailed site evaluation.
Detailed evaluation of two sites has been
conducted with numerical computer models:
1) the Western Processing (WP) hazardous
waste site located south of Seattle,
Washington, within the city limit of Kent
(Figure 1); and 2) the Occidental
Chemical Company (OCC) hazardous waste
site located south of Lathrop, California,
in the Central Valley (Figure 2).
1 CH2M-Hill, Redding, CA
2 Battelle Pacific Northwest
Laboratory, Richland, WA
3 CH2M-Hill, Bellevue, WA
1, 3 formerly with Anderson-Nichols &
Palo Alto, CA
Co.
The WP facility was in operation from
1957 through 1982, and consisted of a
solvent recycling plant, a fertilizer
plant, bulk storage tanks, drum storage
areas, and large above-ground storage
lagoons for liquid wastes, cooling water,
and process water (EPA, 1983). Inspec-
tions made in May 1982 by the EPA and
Washington State Department of Ecology
(WDOE) indicated that drums and bulk
storage tanks were in poor condition and
were leaking, and numerous areas of
spilled material were observed. Twenty-
six of the priority pollutants were found
in soil and ground-water samples taken
from the site. As a result of these
findings, operations at the site ceased in
1982. Initial mitigative measures were
implemented (limited drum inventory and
storage, limited excavation and removal,
etc.) and studies were initiated to
determine a permanent remedial action for
the site.
From 1953 until the present, the
Lathrop plant has produced a wide range of
agricultural products, primarily pesti-
cides and fertilizers. Historically,
liquid process wastes were disposed of in
several unlined surface impoundments and
ditches, and solid pesticide wastes were
buried in an area known as the boneyard.
-------
$ 196TH 5TBEET
Ln
O
PLASTIC COVERED
PILE OF
CONTAMINATED SOIL
EMPTY DRUMS
FULL DRUMS BATT"Y CHI«
LAGOONS FILLED KITH
CONTAH1 HATED VATER
STEEL HILL FLUE DUST
CONTAINING HEAVY METALS
NOT TO SCALE
POLLUTED POND
ASPHALT
,- 1
I
"1
KA1NVAHK
ran
Kiur
nm
row
DRUMS OF ZINC CHLORIDE
Figure 1. Western Processing Site Map.
Figure 2. Occidental Chemical Company, Lathrop,
Site Map (adapted from Canonie, 1981)
-------
The possibility of ground-water contam-
ination was first brought to the
attention of the California Regional Water
Quality Control Board (WQCB) in 1978. In
March 1979 the WQCB issued a cease and
desist order directing an immediate end to
the discharges on the site and compliance
with a site assessment and clean-up sched-
ule because of the public health threat
from 1, 2 dlbromochloropropane (DBCP) con-
taminated ground water (U.S. EPA, 1984).
At this time, both DBCP and sulfate had
been found in private wells used for
drinking water. Subsequently, excavation
and clean-up of the site began under the
review of the EPA National Enforcement
Investigation Center (NEIC) and the WQCB.
Buried pesticide wastes and contaminated
soil were removed and a clay cap
installed. Fifty-four monitoring wells
and an extraction, treatment, and
re-injection system were installed for the
long-term monitoring and remedial clean-up
of the aquifer beneath the OCC site. The
remedial system consists of: 1) five wells
extracting water from the upper aquifer;
2) a reverse pulse granular activated
carbon system to treat the contaminated
extract; and 3) two wells injecting
treated water into a lower briny aquifer.
PURPOSE
The first objective of this study was
to demonstrate, and to the extent
possible, confirm, the use of models for
remedial action evaluation. The second
objective was to provide technical infor-
mation on the performance of remedial
action technologies at actual sites that
will improve the evaluation and decision
process at those sites. The objective of
the Western Processing study was to pro-
vide technical assistance to EPA Region X
in designing and evaluating remedial
action alternatives for the uncontrolled
hazardous waste site. The primary objec-
tive of the OCC study was to demonstrate
the use of numerical computer models for
the evaluation of the effectiveness of an
active remedial action system. A secon-
dary objective was to provide technical
information to the WQCB and California
Department of Health Services.
APPROACH
Conceptual models were developed for
the WP and OCC sites based on the avail-
able hydrogeologic data. The conceptual
models formed the framework for developing
the flow and transport numerical models.
Next, a finite element grid was developed
and the appropriate data on stratigraphy,
boundary and initial conditions, hydraulic
conductivities, and hydraulic stress were
input to the code. The Finite Element
Three-Dimensional Groundwater (FE3DGW)
code (Gupta et al., 1979) was used to
simulate ground-water flow in a localized
area around both sites.
Contaminant transport was simulated
using the Three-dimensional Coupled Fluid,
Energy, and Solute Transport (CFEST) code
(Gupta et al., 1982). CFEST is an exten-
sion of FE3DGW in that it uses the same
hydrologic flow data structure and finite
element grid. In addition, CFEST includes
the necessary parameters to couple contam-
inant transport with ground-water flow.
WP Site
The top 100 ft of the water table
aquifer was modeled at Western Processing.
This section was further divided into four
horizontal layers to allow for a more
detailed description of horizontal and
vertical hydraulic conductivity.
The difference between predicted and
measured hydraulic potentials was mini-
mized in the flow calibration process by
adjusting vertical and horizontal hydrau-
lic conductivities, flow to a creek and
drainage ditch within the study area, and
boundary conditions.
Trichloroethylene (TCE) was chosen
for simulation in the transport model be-
cause TCE has been measured in high con-
centrations, and it is one of the most
ubiquitous contaminants at the site. The
difference between predicted and measured
TCE concentrations was minimized in the
calibration process by adjusting the re-
tardation factor, infiltration rates, and
source strengths.
Once the model was calibrated, it was
used to predict the effectiveness of six
proposed remedial action alternatives: 1)
no-action; 2) source removal; 3) cap; 4)
source removal combined with a cap; 5) up-
gradient slurry wall combined with a cap;
and 6) pump and treat. For each of the
remedial actions simulated, the flow por-
tion of the model predicted changes in the
flow field and volumes of water removed,
while the transport portion of the model
predicted the mass of contaminant removed
and average concentrations up to 25 years
into the future.
51
-------
OCC Site
The aquifer at the OCC site consists
of 250 feet of interbedded clay, silt,
sand and gravel (COM, 1981). Upper,
middle, and lower aquifer zones were
simulated, and five material types were
defined based on the areal distribution of
percent sand and gravel reported in well
driller's logs.
The flow model was calibrated by
minimizing the difference between pre-
dicted and measured potentials through the
adjustment of material properties, boun-
dary conditions, and the quantity and
location of agricultural pumpage.
Because of its toxicity, mobility and
persistence, DCBP has been the target
chemical of the remedial action at the OCC
site. Initial plume distribution for the
the transport model was estimated from
measurements of DBCP concentrations in
samples taken from the permanent moni-
toring wells during October 1982.
The transport model was then used to:
1) compare measured and predicted plume
concentrations; 2) compare measured and
predicted DBCP concentrations at the re-
medial extraction wells; and 3) predict
long-term effectiveness of the remedial
system.
PROBLEMS ENCOUNTERED
At both sites the flow model cali-
bration was hindered by a lack of hydro-
logic data. Extensive hydrogeologic data
exists for the 13-acre Western Processing
Site, but very little data exists for the
area surrounding the site, and for depths
greater than 30 feet below the surface.
At the OCC, site agricultural pumpage to
the west and south of the site was the
most important variable in the flow
calibration process. However, pumping
schedules at these wells were not avail-
able. Estimation of recharge at two waste
water discharge ponds and estimates of
infiltration rates due to irrigation were
also problems for the OCC site.
At both sites there was a lack of
data regarding transport parameters and
operation history. Little or no detailed
data exists regarding when and where con-
tamination occurred, and the quantity and
types of chemicals present. For Western
Processing, no site-specific data were
available regarding partitioning co-
efficients and dispersivities. This lack
of data resulted in assumptions regarding
contaminant transport. Temporal and
spatial variability in DBCP concen-
trations, measured at the monitoring
wells, made detailed mapping of the plume
difficult for the OCC site.
RESULTS
WP Site
A ground-water flow and contaminant
transport model of the Western Processing
Site was developed and calibrated. The
calibration effort was based on comparison
of predicted and observed hydraulic poten-
tials and TCE concentration in the creek
within the study area.
The model results showed that the
creek has been and will continue to be the
primary discharge point for TCE. By 1983,
almost half the TCE estimated to have
entered the flow system during site oper-
ation had exited to the creek. Over the
next 25 years, the model predicted, for
the no-action case, that about 60% of the
TCE remaining in the system in 1983 will
have exited to the creek.
The results of the remedial action
simulations are summarized in Figure 3.
Of the remedial actions simulated, pump
and treat removed the most TCE from the
flow system. Of the passive actions,
source removal with or without a cap was
the most effective. None of the simulated
actions prevented TCE from discharging to
the creek, but the pump and treat and
source removal actions were most effective
in reducing the amount. The capping and
slurry wall options simulated were in-
effective in terms of significantly
reducing the movement of TCE to the creek.
OCC Site
The calibration of the flow model was
based on comparison of predicted and
measured hydraulic potentials. The match
between predicted and measured hydraulic
potentials was very good for the October
1980 to September 1982 period. For the
October 1982 to February 1984 period, the
comparison was not as good as the earlier
period, however, it was determined further
calibration of the flow model would have
little affect on the transport model
results.
October 1983 DBCP concentrations were
generally under-predicted in the center of
the plume by a factor of approximately
52
-------
Ul
OJ
VI
£
-------
1.5. Concentrations on the fringe of the
plume (location of 10 ug/1 contour) were
predicted with reasonable accuracy. A
comparison of predicted and measured
percent plume reduction for October 1983
in the shallow, middle and deep zones is
presented in Table 1. On the whole, plume
reduction was under-predicted, especially
in the middle zone where most of the
remedial extraction wells are screened.
Table 1. MEASURED AND PREDICTED
Zone
Shallow
Middle
Deep
DBCP PLUME REDUCTION
Measured
20%
50%
20%
Predicted
5%
15%
20%
Composite influent concentrations
to the treatment plant were greatly under-
predicted at early times and slightly
over-predicted at latter times. Figure 4
shows the computed and observed concen-
trations at each extraction well. Con-
centrations are slightly over-predicted
for EW-1, 2, and 4 and greatly under-
predicted for EW-3 and 5.
The difference between predicted and
measured results (presented in Table 1 and
Figure 4) are probably caused by hetero-
geneities such as sand and gravel lenses
that are not represented by the transport
model and not completely quantified by the
site data, and by spatial variability in
the plume. DBCP is probably concentrated
and transported in preferred pathways such
as sand and gravel lenses or layers. An
extraction well near one of these pre-
ferred pathways will probably withdraw
much of its water from this sand and
gravel layer, especially at the beginning
of withdrawal. Therefore, concentrations
measured in these wells should initially
be greater than concentrations measured at
the unpumped monitoring wells. Reports
that extraction well concentrations were
much greater than expected during start-up
of the remedial system supports this con-
tention (Roy F. Weston Consultants, 1984).
Because the plume distribution utilized in
the model was estimated from monitoring
well data, the transport model under-
predicted DBCP concentrations at several
of the extraction wells in contact with a
preferred pathway (EW-3, and 5). Spatial
variability in the plume made it im-
possible to estimate the concentration of
DBCP at any individual extraction well.
Plume reduction was under-predicted
because the DBCP is concentrated in the
preferred pathways and not spread through-
out the entire aquifer as modeled.
Long-term model predictions show
substantial DBCP remaining in the aquifer
after ten and twenty years. Stagnation of
the plume occurs after ten years sug-
gesting that the extraction wells may have
to be moved for complete capture of the
plume. Significant movement of the plume
off-site was not predicted. However,
irrigation wells south and west of the
site were predicted to capture some of the
plume. Long-term model results may change
if the model is altered to represent
transport of DBCP through preferred
pathways.
CONCLUSIONS
Results from the two modeling studies
show that numerical computer models are
helpful for the evaluation of remedial
actions at uncontrolled hazardous waste
sites and analysis of the effectiveness of
installed remedial systems, but they have
limited ability to accurately quantify
contaminant transport especially at com-
plex sites with many unknowns concerning
the temporal and spatial attributes of
transport. However, the modeling of both
sites helped to identify data deficiencies
and improve the understanding of contam-
inant transport at the sites. Modeling of
the WP site facilitated the ranking of
remedial action alternatives. Results
from the OCC study shows the importance of
re-evaluating the conceptual model and
recalibrating the numerical model as more
data becomes available.
ACKNOWLEDGEMENTS
The authors wish to thank the staff
at EPA Hazardous Waste Engineering Re-
search Laboratory, particularly the
Technical Project Monitor, Douglas C.
Ammon, for supporting this project. This
work was performed by Anderson-Nichols &
Co., Inc. under Work Assignment No. 21,
EPA Contract Number 68-03-3116. The
Project Officer was Lee A. Mulkey at the
EPA Environmental Research Laboratory in
Athens, Georgia. The authors wish to ac-
knowledge the support they have received
from Mr. Fred Wolf, EPA Region X Office,
in making available data and reports
regarding the Western Processing Site.
54
-------
Obi.rv.«
OV-1 One.
Ob»«rv«tf
CW-2 Cone.
Ob..rv.<
EW-4 Cent.
Ob»«rv«d
EW-5 Cent.
DBCP Concentroton
A
v
-------
Data for the Occidental Chemical Company
Study were made available from Mr. Howard
Hatayarna, California Department of Health
Services and Mr. Thomas Pinkos, Water
Quality Control Board.
REFERENCES
Camp, Dresser, and McKee. Mathematical
Modeling of Ground Water Flow and
Chemical Transport in the Vicinity of
Occidental Chemical Company, Lathrop,
California. Submitted to the
Occidental Chemical Company, Lathrop,
California, July 1981.
Canonie Environmental Services Corpor-
ation. Soil Mitigation Measures
Western Storage Area. Submitted to the
Occidental Chemical Company, Lathrop,
California, July, 1981.
Gupta, S.K., C.R. Cole and F.W. Bond.
Methodology for Release Consequence
Analysis — Part III, Finite-Element
Three-Dimensional Ground Water (FE3DGW)
Flow Model. Formulation Program
Listing and User's Manual, PNL-2939,
Pacific Northwest Laboratory, Richland,
Washington, 1979.
Gupta, S.K., C.T. Kincaid, P.R. Meyer,
C.A. Newbill and C.R. Cole. A
Multi-Dimensional Finite Element Code
for the Analysis of Coupled Fluid,
Energy and Solute Transport (CFEST).
PNL-4260, Pacific Northwest Laboratory,
Richland, Washington, 1982.
Roy F. Weston Consultants. Groundwater
Remedial Program, Lathrop, California
1983 Annual Report, Prepared by Roy F.
Weston, Inc., West Chester,
Pennsylvania for Occidental Chemical
Corp. 1984.
U.S. EPA. Case Studies 1-23: Remedial
Response at Hazardous Waste Sites. EPA
540/2-84-002B, Cincinnati, Ohio, 1984.
U.S. EPA. Investigation of Soil and Water
Contamination at Western Processing,
King County, Washington. Environmental
Services Division, EPA Region X,
Seattle, Washington, 1983.
56
-------
DEMONSTRATION OF A GEOGRAPHIC INFORMATION SYSTEM FOR HAZARDOUS UASTE SITE ANALYSES
Margrit von Braun, P.E.
College of Engineering
University of Idaho
Moscow, Idaho 83843
ABSTRACT
This paper investigates the applicability of a Geographic Information System (GIS)
to the study of a hazardous waste site. Existing data, including aerial photographs,
maps, and site study reports, were digitized to create a spatial data base for an active
Superfund site. These "computer maps" were then analyzed by cartographic modeling
techniques to assess current site conditions and prospective remedial actions. Existing
site drainage basins were identified and characterized by whether stormwater ran off-site
or infiltrated on-site. Wastes were inventoried within each basin and stormwater/waste-
budgets were developed for each drainage basin. Additional "computer maps" were then
constructed to represent remedial action scenarios. Stormwater/waste-contact budgets
were developed for various combinations of berms, dikes, and asphalt covers. These
alternatives were evaluated according to the estimated reductions in waste contacted
runoff and infiltration. Excavation volume estimates were developed for the above-grade
waste types and successive estimates were accomplished for removing contaminated soils
down to the groundwater table.
The application of a GIS to hazardous waste site investigations offers significant
advantages. In addition to the novel analyses demonstrated in this report, a GIS can
provide for more efficient data storage, availability, and utilization than currently
practiced in most site investigations. Several improvements in EPA data management
capabilities and practices, however, will be required. Specific recommendations are
offered. Phase II of this project involves a workshop instructing EPA personnel and
associates in the use of a GIS at a hazardous waste site.
INTRODUCTION
There has been a tremendous in-
crease in both the volume of information
collected and the types of studies and
analyses conducted at waste sites
regulated under the Resource Con-
servation and Recovery Act (RCRA) and
the Comprehensive Environmental
Response, Compensation and Liability
Act (CERCLA or "Superfund"). The
procedures employed at both RCRA and
Superfund sites often require several
years and involve many agencies and
contractors collecting and analyzing
data for different purposes. Unfortu-
nately, as these projects move forward,
there may be abrupt changes in the
personnel and the level of attention
addressed to a site. New investigators
may be unaware of unpublished data
sources collected earlier in a project.
Many studies have a short-term focus and
the data collected are inappropriate or
do not meet quality control standards.
Much information is lost because few data
management systems are flexible enough to
accept, inventory and format particular
types of data (such as maps or aerial
photographs) collected at hazardous waste
sites. These problems result in inef-
ficient use of resources and often
expensive duplicative efforts to restore
or recollect lost or inadequate informa-
tion.
The use of Geographic Information
Systems (GIS) can provide a framework for
resolving several of these difficulties.
A GIS is a computer-based system capable
of assimilating and inventorying
spatially-oriented data from a number of
57
-------
sources in a single location-specific
or "computer map". It is also capable
of manipulating those data in a variety
of processes called cartographic model-
ing. Data from several sources can be
concurrently analyzed without redevelop
ing information collected in earlier
phases of a project. The system can
provide for the storage of almost all
types of available data, facilitate the
evaluation of data from outside sources,
and aid in the design of study plans that
consider the long term value of any
information collected.
controlled and uncontrolled
hazardous waste sites
3. Create a spatial data base for
the demonstration site
4. Perform analyses to assess current
site conditions and prospective
remedial actions
APPROACH
Selection of Geographic Information
System
This paper reports on Phase I of a
two-phase project to investigate the ap-
plicability of a GIS in the management
and analysis of a hazardous waste site.
In Phase I a spatial data base was de-
veloped for an existing hazardous waste
site. Cartographic modeling and geo-
statistical analyses techniques were ap-
plied to the data base to demonstrate the
flexibility and capabilities of a GIS.
In Phase II of the project, a GIS
workshop for EPA project personnel and
associates will be developed. A
lecture/exercise format will be used to
apply the GIS to specific hazardous waste
site problems, including those demonstrat-
ed in Phase I. EPA's National Computer
Center (NCC) computer at Research
Triangle Park, North Carolina will be
utilized. The three-day workshop will
follow this Research Symposium in
Cincinnati, Ohio in May 1985.
PURPOSE
The purpose of Phase I of this pro-
ject was to demonstrate the applicability
of a Geographic Information System (GIS)
to a hazardous waste site. Specific tasks
were:
1. Select a demonstration GIS
suitable for analysis and
instruction purposes
2. Select a demonstration site re-
presentative of the problems and
practices encountered at both
The Map Analysis Package ("MAP"),
a grid-based GIS available from the Yale
School of Forestry and Environmental
Sciences, was selected for use in this
project (1). "MAP" is an academically
oriented GIS developed primarily for
educational and instructional purposes.
It is an English-based "user friendly"
system that employs most of the carto-
graphic analysis techniques used in
environmental and natural resource
research. It is available in the public
domain and is advanced in its analytical
capabilities in comparison to many com-
mercial GIS1.
The "MAP" grid cell structure re-
sembles a matrix, consisting of a
specific number of rows and columns.
Each row-column element or "cell" con-
tains a value representing some dimension
or an attribute of that location. It is
this value assigned to a cell that allows
algebraic manipulation of computer maps.
The manipulation of a cell's value in one
map, based on either the value of its
neighbors or on that same cell's at-
tribute in another map, is called "map
algebra" (2). Map algebra is analagous
to traditional algebra except individual
maps are treated as the independent and
dependent variables.
Figures 1 and 2 show how maps may be
combined using traditional algebraic
operators to produce new maps. In Figure
1, the "point characteristics" of the
maps are utilized. The value of a
particular cell in one map is combined
with the corresponding values in the same
cell of other maps to produce a new map.
58
-------
MAPS
FIGURE 1 A new map, D, is produced
by simultaneously considering
values of existing maps, A, B,
& C, on a cell-by-cell basis.
Figure 2 represents how "neighborhood
characteristics" are utilized. Using
Boolean algebra in a procedure analagous
to traditional map overlay or successsive
subtraction techniques, the "neighbor-
hood" of a specific sample location may
be characterized. Logical combinations
of intersections and unions are utilized.
MAPS
DECISIONS
WASTE PILES
DEPTH TO
GROUNDWATER
EXISTING DATA
NEW
SAMPLE SITES
PRESENCE OF
WASTE
SHALLOW
GROUNDWATER
NO DATA
GOOD SAMPLE
LOCATION
FIGURE 2 A new map, "New Sample Sites",
is created by considering the
neighborhood of a cell on
existing map overlays.
These figures demonstrate the concept of
map manipulation. The actual techniques
used for the manipulation vary from
simple addition to movement across
friction or weighted surfaces. Detailed
descriptions of the techniques may be
found in Tomlin (1).
Selection of Demonstration Site
Western Processing, a Superfund site
in Kent, Washington, was selected as a
demonstration site representative of the
problems, management practices, and data
quality encountered at both controlled
and uncontrolled hazardous waste pro-
jects. The 13-acre site was operated as
an Army 70-mm battery and as a privately
owned waste recycling and storage
facility until 1982. A variety of
hazardous wastes had been stored on the
site. EPA conducted a detailed site in-
vestigation in 1982 during which over
30,000 data points from 30 monitoring
wells, 50 soil sample sites and 28 air
sample sites were collected. Eighty-
seven priority pollutants, including 21
known and 28 suspected carcinogens were
found on or near the site (3).
The EPA and the State of Washington
began remedial actions in 1983 utilizing
berms and impervious surfaces to prevent
imminent spillage of materials and to
reduce runoff of contact stormwater. The
EPA has supported several research in-
vestigations of alternate cleanup
strategies at this site. They include
building decontamination, drum encapsula-
tion, control material weathering,
groundwater control and revegetation
studies.
Creation of Spatial Data Base
Aerial photographs, maps, slides,
reports and data summaries obtained from
the EPA Region X Office, EPA Laboratories
and contractors were digitized to produce
base maps of the site (4). A scale of
225 ft2 per grid cell was used for each
map. The maps, which were 58 columns
wide and 116 rows long, and contained
6728 cells equivalent to 34.8 acres, are
demonstrated in Figures 3 through 5. A
perimeter around the 13-acre site allowed
for consideration of surrounding off-site
conditions.
59
-------
The base maps were combined and
manipulated to create new or "derived"
maps which became part of the spatial
data base for the site. For example, an
overlay of the PROPERTY LINES, CREEK,
BUILDINGS and SITE ROADS base maps yields
a new map of the SITE FEATURES. Maps
resulting from the manipulation of other
maps are distinguished from base maps in
that they require no additional data
input or digitization but instead are
derived from information already con-
tained in the data base.
Demonstration Analyses
Three demonstration analyses
utilizing the completed data base are
discussed. These analyses address
on-site surface characteristics, however
these techniques could be applied to
off-site groundwater and air pollution
problems as well. A brief description
of each follows:
1. Identification of Original Site
Characteristics
Cartographic dispersion techniques
were applied to the original site
topography map to identify twenty-seven
unique drainage basins for surface
runoff waters. Figure 3 illustrates the
results of these analyses.
Basins were categorized by whether
they drained off-site or ponded and in-
filtrated on-site. An overall stormwater
budget was developed and the basins were
ranked by their contribution to site
runoff or site infiltration as shown in
Tables 1 and 2.
Cartographic inventory methods were
then used to characterize the type and
volume of wastes in each basin. These
results were combined to identify those
basins where major stormwater/waste
contact occurred and how and where those
waters eventually left the site. Figure
4 shows the wastes within the major
basins. Resulting inventories for the
major basins are summarized in Table 3.
TABLE 1. DESCRIPTION OF SITE BASINS
PONDING ON-SITE
% of
Total
Basin Area Site % of
Number Drained Drainage On-site
Rank (Fig.3) (acres) Area Ponding
1
2
3
4
4
5
5
6
6
6
6
7
8
9
16
23
24
20
21
15
18
11
12
14
17
10
22
9
2.0
1.0
.7
.6
.6
.5
.5
.4
.4
.4
.4
.3
.1
.2
13
6
4
4
4
3
3
3
3
3
3
2
<1
1
24
12
8
7
7
6
6
5
5
5
5
4
3
2
53% OF TOTAL SITE RUNOFF
PONDS ON-SITE
TABLE 2. DESCRIPTION OF SITE BASINS
DRAINING OFF-SITE
Basin Area % of
Number Ultimate Drained Off-site
Rank (Fig.3) Destination(acres) Runoff
1
2
3
4
4
5
6
7
7
7
8
9
1
19
7
3
6
8
25
2
7
27
26
5
Creek
West
Site Ditch
East
Site Ditch
West
South
North
North
South
South
Site Ditch
1.6
1.1
1.0
.8
.8
.5
.4
.3
.3
.3
.2
<.l
22
15
14
11
11
7
5
4
4
4
3
1
47% OF TOTAL SITE RUNOFF
DRAINS OFFSITE
60
-------
TABLE 3. FEATURES IDENTIFIED WITHIN
MAJOR DRAINAGE BASINS
0
N
S
I
T
E
0
F
F
S
I
T
E
No. Cel
taining
Basin # Wastes (1 cell
16 Flue Dust
Barrels
Plating Tanks
Storage Tanks
23 Barrels
Lead & Wood Chips
Debris
24 Barrels
Debris
1 Fuel Tanks
19 Flue Dust
Barrels
7 Gypsum Pile
Is Con-
Waste
=225 ft2)
5
21
1
13
17
32
17
56
6
1
44
5
47
This application illustrates the
quick use of the initial data base in
characterizing site impacts. Both
specialized cartographic dispersion
modeling techniques and procedures
mimicing traditional map reading
(overlay and inventory) were utilized.
These types of results can be used in
the design of sampling strategies,
preparing contractor directives, or in
developing and evaluating remedial
action proposals.
2. Evaluation of Stormwater Control
Remedial Alternatives
This analysis also involves
stormwater budgets, but illustrates
repetitive cartographic modeling
techniques. The input topography map
was modified to reflect proposed
remedial actions. Berms were "built" on
the topography map to stop site runoff
and the new ponds and infiltration
basins were identified and quantified.
Areas were then "paved" co prevent
infiltration and route uncontaminated
stormwater offsite. Several com-
binations of berming and ponding were
evaluated according to their effect on
the site's stormwater budget. Figure 5
illustrates one cartographic "con-
struction" of remedial activities.
Figure 6 compares the final stormwater
disposition of all alternatives
considered.
FIGURE 3 Drainage basins and ponded
areas (in white) are identified
to characterize the Hestern
Processing site status in
July 1983.
61
-------
-------
-------
project would greatly facilitate data
management efforts and could be ac-
complished with a modest expenditure of
resources. Specific recommendations
are:
* Require all site studies to
conform to a master site map ard all
data to be collected in a location-
specific format
* Adopt standards of quality for
survey design and data collection
* Upgrade hardware and interactive
computing capabilities
These requirements offer advantages
in quality control, quality assurance,
and efficiency regardless of the data
analyses or management system employed.
RESULTS
The example analyses completed in
this study successfully demonstrated
many of the capabilities and usefulness
of a GIS within the limitations of this
data base. These particular analyses
illustrate the GIS' ability to:
* Accept information from several
type of data bases
* Initially characterize the site
with traditional map reading techniques
* Use cartographic modeling to
evaluate stormwater control remedial
alternatives
* Perform traditional engineering
calculations in estimating excavation
volumes and site slopes
The function of a GIS in hazardous
waste site application can be summarized
in four general areas:
* Inventory and storage of site
data in a common data format
* Data analyses either i) using
the GIS or ii) preparing data for
outside processing and returning results
to the GIS
* Development and integration of
external data bases
* Preparation of reports or
special studies
This project did not demonstrate
some aspects of the GIS important to
sophisticated spatial data management
and analyses — the use of the GIS as an
input-output device for other analytical
systems. Many environmental models and
statistical analysis packages require
input data to be submitted in grid-based
format. Sophisticated hydrologic
models, the EPA UNAMAP series of air
quality models, spatial interpolation
routines such as kriging, and several
statistical techniques operate with
cartesian coordinate systems. The GIS
can be used to develop input data for
these routines and to accept, manipulate
and further analyze results obtained
from external programs. For example, the
results of groundwater models could be
analyzed in the context of local land
use characteristics. A GIS can also be
used to develop input information for
external hardware devices. Data can be
prepared for plotters and printers and
as input to graphics display packages.
The GIS1 ability to store, combine, and
manipulate the results of various site
studies for output to external models
and for display purposes could add a new
dimension to hazardous waste site
analyses.
In addition, continuing advances in
digital cartography will greatly facili-
tate the application of GIS' in the next
few years. Institutional data bases are
being developed for direct input to a
GIS. Many government agencies have
developed GIS capabilities and several
are digitizing their data files on a
routine basis. Digital information is
available for topography, roads and
highways, utilities, soils, surface
hydrology, vegetation cover, geology,
and landuse. With these types of data a
GIS can make rapid assessments of site
suitability with regard to environmental
conditions, surrounding land uses, and
number of people affected by hazardous
waste facilities. GIS1 can be used to
quickly assess site characteristics for
disposal locations under RCRA authority.
The development of standardized data
bases by the USGS and the US Bureau of
Census and the explosion in information
and data processing capabilities will
make it likely that GIS applications
will become commonplace in the next
decade.
ACKNOWLEDGEMENTS
The research described in this
paper was supported by the U.S. En-
vironmental Protection Agency, Land
Pollution Control Division, Hazardous
Waste Engineering Research Laboratory
(HWERL), Cincinnati, Ohio under contract
64
-------
#68-03-3210 to the University of
Cincinnati (UC) Center Hill Solid and
Hazardous Waste Research Facility. The
work was performed under Work Assignment
No. 5 of the contract which provided
technical support systems for evaluating
hazardous waste disposal technology.
Sincere appreciation goes to the
staff members of the Land Pollution
Control Division at HWERL, Cincinnati,
Ohio, EPA Region X, Seattle, Washington,
the Environmental Monitoring Systems
Laboratory (EMSL), Las Vegas, Nevada, and
the UC Center Hill Facility, Cincinnati,
Ohio who provided guidance, data and
facilities.
Special thanks go to the following
individuals: Douglas C. Ammon, HWERL
Technical Project Monitor, for his
support and advice in the conception and
implementation of this project; John
Barich, EPA Region X, for his special
interest in the project and insights
about the Western Processing site; and
Martha Lambert, University of Cincinnati,
for her support and assistance at the
Center Hill Facility.
REFERENCES
1. Tomlin, C. Dana. 1980. The Map
Analysis Package . Yale School of
Forestry and Environmental Studies,
New Haven, CT.
2. Berry, Joseph K. and C. Dana
Tomlin. 1984 Geographic
Information Analysis Workbook.
Yale University School of Forestry
and Environmental Studies, New
Haven, CT.
3. U.S. Environmental Protection
Agency. 1983. Investigation of
Soil and Water Contamination at
Western Processing, King County,
Washington; Parts 1 and 2. 1982.
EPA 910/9-83-104a and 104b,
Environmental Services Division,
Seattle, Washington.
4. von Braun, Margrit. 1984.
Demonstration of_ a Geographic
Information System iji the Analysis
of a^ Hazardous Waste Site. (Draft
report^Prepared for the
Environmental Protection Agency
Hazardous Waste Engineering Research
Laboratory, Cincinnati, Ohio.
65
-------
QUALITY CONTROL OF HYDRAULIC CONDUCTIVITY AND
BENTONITE CONTENT DURING SOIL/BENTONITE
CUTOFF WALL CONSTRUCTION
Matthew J. Barvenik, William E. Hadge, Donald T. Goldberg
Goldberg-Zoino & Associates, Inc.
320 Needham Street, Newton, Massachusetts 02164
ABSTRACT
Soi 1/bentoni t e backfilled cutoff walls (slurry walls) have been used over the past
several years to contain and isolate hazardous wastes. While may cutoff walls have been
installed, their effectiveness and quality of construction have not been studied or
documented to any extend. This paper reports on the first phase of testing at the Gilson
Road site, Nashua, New Hampshire. The objective of this phase was to evaluate quality
control measures for the determination of bentonite content and hydraulic conductivity.
This paper specifically addresses methylene blue titration for measurement of bentonite
content and the API fixed ring permeameter for evaluation of hydraulic conductivity. The
accuracy, applicability and practicability of these tests as field quality control
measures are discussed.
INTRODUCTION
Soi 1/bentonite backfilled cutoff
walls (slurry walls) have successfully
been used as an adjunct to pumping for
excavation dewatering and general
groundwater control for over forty years.
Specifications governing this work
typically fell under the dewatering
section of construction documents and, as
such, were performance oriented. In many
cases, success was judged solely on the
ability of the barrier to limit flows to
quantities that could be reasonably pumped.
More recently, soil/bentonite (S/B) cutoff
technology has been gaining acceptance as
a tool for remediation of US EPA National
Priority List (NPL) sites, as well as
private industrial sites. For these more
critical applications, leakage rages which
were considered insignificant now
represent unacceptable performance with
respect to hazardous waste containment.
The successful construction of a
hazardous waste cutoff wall for
containment is still largely dependent on
the quality of the contractor's
workmanship and as such, specifications
remain performance oriented. However, in
recognition of the difficulties involved
in performance verification and the
consequences of detecting substandard
construction only after wall completion,
contract documents are beginning to
include minimum design standards. These
center around minimum bentonite and fines
content (see Ayres j^t al. 1983). This
allows quality control (QC) testing to be
executed in the field during construction
as a check on attainment of the minimum
standards specified (Figure 1).
Quality control testing for the
Gilson Road (NPL site) cutoff wall was
consciously planned from design through
specification preparation and then
executed in the field to effectively guide
the contractor towards a successful S/B
cutoff wall installation. New S/B
backfill QC techniques, including
methylene blue titration and API fixed
ring permeation, were utilized for
bentonite content and hydraulic
conductivity determinations, respectively.
This work demonstrated the applicability
and practicability of these new methods
for field control of S/B backfills and
thus led the US EPA to fund research into
the refinement of the procedures and
evaluation of their precision and accuracy.
This work represents Phase I of a larger
66
-------
LJ
O
o:
u
15 20 25 30 35 40 45
PERCENT -*200 SIEVE
468
PERCENT BENTONITE
10
0 0.4 0* 1.2 1.6 20
HYDRAULIC CONDUCTIVITY x O'Tcm/sec
Figure 1. Quality Control Data Obtained
During Construction of the
Gilson Road Cutoff
effort funded by a cooperative agreement
between the New Hampshire Water Supply &
Pollution Control Commission (NHWSPCC) and
US EPA which also includes post-
construction barrier continuity
verification via piezocone techniques,
undisturbed sampling/lab testing, and pump
testing. Subsequent phases of the
research are currently ongoing with the
overall objective of producing the basis
for general QC and verification testing
guidelines for S/B cutoff walls.
EVALUATION OF BACKFILL HYDRAULIC
CONDUCTIVITY
Fixed ring equipment was used to
provide hydraulic conductivity QC data
during construction of the Gilson Road
cutoff wall. This inexpensive, easily
operated equipment allowed field testing
of numerous samples prior to permanent
incorporation of the backfill into the
cutoff. The one day availability of
hydraulic conductivity data allowed real
time quality control of the construction
process. In light of the high degree of
usefulness and practicability of the fixed
ring hydraulic conductivity testing as
shown on the Gilson Road project, a R&D
testing program was undertaken to:
1. Assess the actual accuracy and
precision of the test
2. Based on No. 1 above, evaluate the
validity of the QC data collected on
the Gilson Road project, and
3. Refine and standardize the test
procedure for incorporation in
general guidelines for QC testing of
S/B cutoff walls.
API Fixed Ring Test
Various equipment configurations
based on fixed ring confinement of a test
sample for evaluation of hydraulic
conductivity have been in use for many
years. This equipment is discussed in the
U.S. Army Corps manual (1970) and Lambe
(1951). In most cases, however, the
equipment is meant for laboratory use and
not designed to withstand the
"environment" associated with field
testing. In an effort to standardize
procedures around equipment which would
survive repeated field use and also be
readily obtainable, the conventional
67
-------
American Petroleum Institute (API) filter
press was adapted for use as a fixed ring
permeameter (see Figure 2). As a filter
press, this equipment is already routinely
used on cutoff wall projects to test the
filter cake forming capacity of bentonite
slurries used for trench stabilization.
In basic terms, API hydraulic
conductivity testing involves placing a
sample in the cell, covering it with
water, and forcing the water through the
sample using pneumatic pressure. Once
consolidation of the sample is complete,
the steady state hydraulic conductivity is
computed based on standard Darcy's Law
relationships assuming 100 percent
saturation. For typical S/B backfill
samples, approximate values of hydraulic
conductivity can be obtained in several
hours and final values in less than 24
hours. As such, the procedure is ideal
for real time quality control of the
construction process. A detailed
description of the standardized test
procedure, which directly reflects the
results of research efforts performed for
this project, will be presented in a
future report.
DIAL GAUGE
17
COMPRESSED
AIR SOURCE
BASE CAP
STEEL MESH SCREEN
FILTRATE TUBE
SUPPORT ROD
GRADUATED CYLINDER
Figure 2. API Hydraulic Conductivity Apparatus
68
-------
API Test Precision
The first level of investigation was
directed towards determining if equipment
and procedures as simple as the API based
QC testing performed on the Gilson Road
project would provide repeatable hydraulic
conductivity values. The degree to which
the precision of a test procedure can be
established, however, is highly dependent
on how identical the test specimens are
prepared. This is problematic for API
hydraulic conductivity testing in that a
single specimen cannot be set up more than
once. Therefore, multiple specimens are
required to evaluate the degree of
variability in the entire procedure, from
set-up through permeation. Unfortunately,
the coarseness of granular based backfills
such as used on the Gilson Road project
can result in a somewhat heterogeneous
sample from which the individual aliquots
are taken for the comparison testing.
Therefore, uniform mixing, to the extent
practical in the laboratory, was
emphasized during sample preparation.
However, it is likely that some degree of
heterogeneity still existed among the
comparison samples.
Five batches of the Gilson Road
backfill were selected from record samples
obtained during construction. Selection
emphasized encompassing the full range of
gradations and hydraulic conductivities
encountered during construction of the
actual cutoff wall. After thorough
remixing in the laboratory, three
"identical" aliquots of each sample were
tested at a gradient of 40 (driving
pressure of 3 psi). The results are
presented in Table 1. The data indicate a
worst case precision of +57 percent within
one standard deviation of the mean value
for any group of three tests. Although
this initially appears as a large degree
of scatter, it is actually quite small
when compared to the five orders of
magnitude (100,000 times) reduction in
permeability of the initial in-situ soils.
Further inspection of the data also
reveals that in all five groups of tests,
TABLE 1
API HYDRAULIC CONDUCTIVITY TEST; PRECISION SERIES
HYDRAULIC CONDUCTIVITY X 10-7 in on/sec
STD. DEV.
SAMPLE Individual Splits One Standard MEAN
of Each Sample Mean Deviation in %
Sta. 0405
Sta. 7+75
Sta. 14+75
Sta. 20+25
Sta. 35+00
0.48
0.38
0.46
4.02
4.02
4.44
0.80
1.86
0.79
0.63
0.25
0.25
0.69
0.28
0.38
0.44 +0.05 11%
4.16 +0.24 6%
1.15 +0.61 53%
0.38 +0.22 57%
0.45 +0.21 47%
69
-------
at least two of the three tests in each
group agreed to +10 percent within one
standard deviation of the mean. Hence,
some of the scatter reflected in the worst
case precision of ;+5 7 percent may be
attributable to the effect of less than
identical samples as previously explained.
This hypothesis is supported by field data
obtained during subsequent QC testing for
other cutoff walls utilizing primarily
clayey mixes where the probability of
obtaining identical samples is
significantly enhanced.
API Test Accuracy
To evaluate the accuracy of a
specific test procedure, it is imperative
to know the "true" value of the parameter
of interest for the specimen being tested.
This is not possible for hydraulic
conductivity due to its dependence on the
method of measurement used. The value of
hydraulic conductivity adopted as the
"absolute" value to which all other values
are compared in this work was derived via
triaxial testing. Triaxial testing,
however, can also yield a wide range of
values for the same S/B backfill sample.
This variability stems primarily from the
lack of standardization of physical
testing parameters such as state of
stress, degree of saturation, gradient,
etc., to which the sample is highly
sensitive (see Schulze et al., 1984). The
accuracy evaluations made herein are
therefore based on "absolute" triaxial
values in which the physical testing
parameters were standardized to model
in-situ conditions at the top of the
cutoff wall (upper 5 to 10 feet). As
such, these "absolute" values of hydraulic
conductivity would be conservative (high)
relative to those expected for the same
sample tested under conditions simulating
greater depths or incomplete saturation.
The philosophy of modeling the worst case
condition of "top of wall" recognizes the
indeterminacy of the final vertical
location of a specific batch of backfill
represented by a QC sample taken prior to
backfill placement.
During construction of the Gilson
Road cutoff wall, ten samples of the
backfill were obtained for triaxial
testing. These samples were split and one
portion was also field tested in the API
apparatus as part of the QC program.
Therefore, to the degree that the samples
taken were homogeneous, the value of
hydraulic conductivity was determined by
both triaxial and API procedures on
"identical" specimens. These data are
presented in Figure 3. The data indicate
that the API values match the triaxial
values to within +_66 percent as a worst
case. This error band not only
encompasses the accuracy of the API test
but also incorporates errors associated
with heterogeneity between "identical"
samples and precisional errors inherent in
both the API and triaxial procedures. As
such, the data indicate a very good level
of agreement between the simple field API
test and the more sophisticated laboratory
triaxial test.
Field Use of the API Test
Based on the research presented
herein, and the field data from actual
cutoff wall projects undertaken to date,
the API test appears to yield reliable and
accurate values of hydraulic conductivity.
However, it is considered prudent to run a
limited number of triaxial tests for each
project to verify the site specific
accuracy of the API procedure. The
initial check between API and triaxial
data should be made during the design
phase of the project when backfill
component selection and proportioning is
established based on laboratory testing.
This design testing should then be
followed by triaxial testing of actual
backfill samples obtained during
construction.
Although a high degree of additional
backfill mixing is required for the
API/triaxial comparison samples, mixing of
field samples taken for QC testing in
excess of that achieved during
construction is highly undesirable. This
directly reflects the dependence of
hydraulic conductivity on the degree of
backfill component blending to a scale
commensurate with that of interstitial
pore size. Data from triaxial testing of
post-construction, undisturbed samples
indicate that up to an order of magnitude
reduction in hydraulic conductivity can
occur if the originally tested undisturbed
samples are mixed thoroughly and retested.
Therefore, samples for QC should be taken
and placed in the test chamber with as
little additional mixing as possible while
still removing air voids. Failure to
observe this procedural "detail" can
70
-------
u
o>
X
£
u
U
Q
Z
o
u
o
oc
o
HYDRAULIC COND.x I0"'cm./$ec.
O.I 0.5
2.0
3.0
TRIAXIAL HYDRAULIC CONDUCTIVITY x I0"7cm./sec.
Figure 3. Comparison of API vs. Triaxial Hydraulic Conductivity
4.0
result in artificially low values of
hydraulic conductivity. It also follows
directly from this discussion that both
the API and triaxial data sets obtained
for test comparison, do not qualify as QC
data. The sole purpose of these tests is
to demonstrate the site specific validity
of the API procedure.
The gradient (driving pressure) used
during API QC testing should be that value
which yields a hydraulic conductivity
closest to that derived from triaxial
testing. (The gradient used for the
triaxial testing is commensurate with "top
of wall" stresses as previously discussed).
It is emphasized that significantly higher
gradients should not be used in an effort
to increase the flow per unit time so as
to make effluent volume measurements
easier. High gradient results in
consolidation of the sample to levels no
longer commensurate with "top of wall"
stresses. As such, artificially low
values of hydraulic conductivity can be
obtained. Furthermore, high gradient can
lead to piping and thus unrealistically
high values of hydraulic conductivity not
representative of the future performance
of the wall.
EVALUATION OF BACKFILL BENTONITE
CONTENT
In the past, field measurements of
bulk quantities of backfill soils and
bentonite have been used to determine
bentonite percent by applying basic
weight/volume relationships. This method
is approximate due to the limited accuracy
of the bulk weight/volume estimates, and
optimistically assumes the bentonite is
evenly mixed throughout the backfill.
Thus, bulk measurements can be used as a
71
-------
gross check on the percentage of bentonite
in the backfill, but are inadequate for
quantitative field control. As an
alternative, the use of conventional
hydrometer test procedures have been
attempted for bentonite determination.
Inherent limitations in the test with
respect to accuracy, and completion time,
however, restrict its use for backfill QC
testing. A laboratory testing program was
therefore carried out to investigate the
applicability of the quicker and
potentially more accurate methylene blue
test for determination of percent
bentonite in S/B backfill.
Methylene Blue Titration
In a suitable aqueous environment,
such as an electrolytic clay-water
suspension, clay materials are associated
with naturally occurring absorbed cations
of a fixed total charge (typically sodium,
calcium, magnesium, or potassium). These
cations are either retained by the clay or
may be exchanged for other cations
subsequently introduced into solution.
The degree to which this exchange may
occur, referred to as cation exchange
capacity, increases with the amount and
activity of the clay in suspension. The
determination of bentonite content
therefore relies on the proportionality
between the amount of methylene blue
cations exchanged and the amount of
bentonite in the backfill sample tested.
The quantity of methylene blue cations
exchanged is determined by titration.
The titration procedure involves the
incremental addition of methylene blue
solution to a suspension consisting of the
backfill sample and a predetermined volume
of distilled water and dispersant until
the cation exchange capacity of the
suspension is exceeded (end point).
Methylene blue consumption at the end
point is then used to determine the dry
weight of bentonite in the sample via a
pre-established correlation. Although the
titration volume actually corresponds to
the dry weight of bentonite, the data will
be presented herein as an equivalent
bentonite percent in the sample. This
format is used in recognition of the
ultimate objective of standardizing this
procedure for QC of S/B backfill mixes.
The methylene blue test is well
suited for quality control of S/B
backfills due to the relatively short
testing time (approximately 30 minutes per
test), low cost, and minimal laboratory
requirements. A detailed description of
the standardized test procedure, which
directly reflects the results of research
efforts performed for this project, will
be presented in a future report.
Testing Program
Evaluation of the accuracy and
precision of methylene blue titration when
used as a QC test for S/B backfill was
separated into four distinct phases.
Phasing was implemented to progressively
assess the impact attributable to
increasingly more complicated matrices
accompanying the bentonite. These
matrices encompassed the simplest case of
distilled water only to more realistic
combinations approximating actual backfill
samples including natural soils with
non-bentonite related cation exchange
capacity and interstitial water
contaminated with hazardous wastes. The
applicability of linear proportionality,
matrixing effects and superposition were
thus investigated. Testing accuracy was
assessed through comparisons of measured
bentonite content with that actually added
to the laboratory prepared samples. The
four phases were:
• Bentonite in a distilled water
matrix;
• Bentonite in an Ottawa sand and
distilled water matrix;
• Bentonite in a contaminated water
matrix;
• Bentonite in an off-site borrow,
trench spoil and distilled water
matrix.
Distilled Water Matrix (Phase I)
Bentonite in a matrix of distilled
water and dispersant was chosen for the
first phase of testing to eliminate
complicating reactions which might
otherwise be induced by additional
backfill components. This, the simplest
possible matrix, was used to assess the
proportionality and inherent accuracy/
precision of the methylene blue titration
for bentonite content. Three to four
aliquots each of ten slurries containing
72
-------
increasing bentonite amounts from 0.0 to
0.9'grams were prepared. The bentonite
amounts selected correspond to 10 gram
backfill samples containing 0 to 9 percent
bentonite by dry weight. Standardization
of a 10 gram sample size reflects
sampling, testing, and typical backfill
gradation considerations. The same
bentonite was used throughout the testing
and conformed to the American Petroleum
Institute's API grade, minimum 91 barrel
yield.
For each of the 36 individual tests,
the volume of methylene blue solution
added to reach the endpoint of titration
(completion of cation exchange) was
plotted versus the known percent bentonite
referenced to a 10-gram sample (see
Figure 4). The trend portrayed by the
data not only demonstrates proportionality
but also indicates a linear relationship
between methylene blue titration volume
and percent bentonite. Linear regression
analysis performed on the data yields a
line referred to as the "bentonite
calibration curve" which fits the data set
to a high degree as reflected by a
coefficient of determination equal to
0.999. Hence, given a slurry sample
containing an unknown amount of bentonite,
it would be expected that the titration
volume in conjunction with the previously
established calibration curve would yield
the correct percent bentonite to within
+0.2 percent with a 95 percent degree of
confidence.
The bentonite/distilled water samples
used to establish the basic relationship
between titration volume and percent
bentonite excluded all other possible
backfill components. This "best case"
condition is therefore expected to yield
UJ
d
UJ
>
X
UJ
234567
% BENTONITE IN DISTILLED WATER
Figure 4. Titration for Bentonite Content in Distilled Water Matrix
73
-------
the highest degree of accuracy. The curve
defined by this data is the baseline
calibration against which the results of
subsequent tests on more realistic
backfill samples are compared and testing
accuracy assessed.
Ottawa Sand/Distilled Water Matrix
Phase 2)
For the second phase of testing,
known amounts of bentonite were added to a
matrix composed of Ottawa sand, distilled
water, and dispersant. These samples were
one degree more complex than those used
for Phase I testing and also one step
closer to a typical backfill sample.
Ottawa sand was selected in that it is
almost entirely quartz with no clay
minerals present. As would be expected
based on cation exchange principles,
quartz, and thus the Ottawa sand, has no
exchange capacity. The specific objective
of this testing phase was therefore to
100
determine whether the presence of backfill
components having no cation exchange
capacity would interfere with the
bentonite/methylene blue cation exchange.
This interference, referred to as
"matrixing effect," would include any
physical, chemical, or electrical
phenomena which alter the surface activity
of the bentonite clay mineral.
Three identical samples were prepared
individually at each of five bentonite
contents; 0, 2, 4, 6, and 8 percent. Each
sample contained appropriate dry weights
of Ottawa sand and bentonite to yield
10 gram "backfill" samples with the
specified bentonite percentage. These
samples were added to the standardized
volume of distilled water/dispersant and
slurried for testing. For each of the
15 tests, the methylene blue titration
volume at end point was plotted versus the
known bentonite percentage (Figure 5).
The previously determined calibration
BENTONITE CALIBRATION CURVE
95% CONFIDENCE INTERVAL
I 2345678
% BENTONITE IN OTTAWA SAND/DISTILLED WATER
Figure 5. Titration for Bentonite Content in
Ottawa Sand/Distilled Water Matrix
74
-------
curve and 95 percent confidence interval
(from Figure 4) are also presented with
the data for comparison. The data
indicate that, within the +0.2 percent
error band established via the 95 percent
confidence interval, the Ottawa sand has
no effect on the methylene blue/bentonite
cation exchange process. Therefore,
matrixing problems are not probable due to
the presence of quartz based backfill
components with no cation exchange
capacity.
Contaminated Water Matrix (Phase 3)
During methylene blue testing,
backfill samples are, in all cases, mixed
with distilled water to create a
suspension which is then titrated. Hence,
the water used during the testing is
always contaminant free. However,
contaminants may be added to the
suspension via the backfill sample itself,
particularly where the cutoff wall is
excavated through contaminated portions of
the aquifer. To assess the possible
matrixing effects of these contaminants on
the accuracy of the methylene blue test, a
worst case scenario was utilized for
sample preparation. In addition to
incorporation of contaminants via the
backfill, contaminated water was used to
form the test suspension instead of
distilled water.
Contaminated water from three
different sources on the Gilson Road site
were obtained for this phase of the work.
The first two, designated well water and
stream water, exhibited significant
hardness and were sources actually used
during construction for slurry makeup
water. The third, and most contaminated
source, was leachate containing volatile
organic solvents taken from monitoring
installations on the site. Backfill
samples were prepared at 0, 2, 5, and
8 percent bentonite in a manner similar to
Phase I testing. Two aliquots were tested
for each bentonite content using the
stream and well water (total of
16 samples). One series was performed
using leachate at the four bentonite
contents (4 samples). The data, presented
in Figure 6, indicate that the
contaminants tested had no significant
effect on the test results. All but one
of the individual data points fall within
+0.2 percent of the original bentonite
calibration curve. As such, the initial
calibration accuracy of +0.2 percent with
a 95 percent confidence level was
maintained in spite of the Gilson Road
contaminants introduced into the matrix
tested.
For volatile organic contaminants,
noninterference would be expected inasmuch
as they are non-ionic and therefore should
not exhibit a high cation exchange
capacity. In addition, the backfill
sample is dryed in an oven at 105°C and
the backfill/water/dispersant matrix is
boiled prior to titration. The
temperature involved in the above
procedural steps would be expected to
remove the volatiles. For other sites,
particularly.those containing highly ionic
contaminants, the cation exchange
potential of the species present should be
compared to that of methylene blue with
respect to affinity for the bentonite clay
mineral. In cases where high non-volatile
solvent concentrations exist, reduced
solvent (water) polarity effects, as they
relate to methylene blue solubility,
should also be investigated. This can be
done on a theoretical basis or by
utilizing the simple testing procedure
described herein. It should be recognized
however that this testing procedure can
yield "false positives" in that the
quantity of contaminants introduced into
the matrix far exceeds that which is
likely to be added via an actual backfill
sample.
Gilson Road Backfill Matrix (Phase 4)
The fourth phase of testing employed
prepared samples duplicating the backfill
used for the Gilson Road project. The
Gilson Road backfill consisted of
approximately equal proportions of trench
spoil (sand and gravel) and off-site
borrow (sandy silt) with 3 to 5 percent
bentonite added. The objectives of this
series of tests were to 1) evaluate the
impact of a cation exchange capacity
attributable to non-bentonite backfill
components on the determination of percent
bentonite in the backfill; 2) determine if
the bentonite calibration could be
corrected to account for non-bentonite
cation exchange capacity utilizing
superposition; and 3) based on the above,
evaluate the accuracy of the methylene
blue test results previously obtained for
the Gilson Road project.
75
-------
• WELL WATER
STREAM WATER
LEACHATE
BENTONITE CALIBRATION
CURVE
95% CONFIDENCE INTERVAL
234567
% BENTONITE IN CONTAMINATED WATER
Figure 6. Titration for Bentonite Content
in Contaminated Water Matrix
The on-site soils (trench spoil
component of the backfill) used for
testing were obtained during a
post-construction boring program. The ten
samples were mixed with distilled water
and dispersant in a manner similar to
Phase II testing except that no bentonite
was added. The on-site soils exhibited
average cation exchange capacities
equivalent to 0.2 percent bentonite with a
range of 0.1 to 0.4 percent. Nine samples
of the off-site borrow were also selected
from record samples taken during
construction. Eight of these samples
yielded bentonite equivalents of 0.6 to
0.7 percent with one sample yielding
0.4 percent. This preliminary testing for
the exchange capacity of individual
non-bentonite backfill components provided
a basis for prediction of the bias
expected relative to the bentonite
calibration previously developed. An
anticipated offset equivalent to
0.4 percent bentonite was derived using a
weighted average of the individual
exchange capacities based on the relative
amounts of borrow and trench spoil in the
backfill.
Fifteen splits of backfill were
prepared containing equal proportions of
the on-site soil and the off-site borrow.
Bentonite was added to the resulting
mixture (a silty sand and gravel) to
obtain three samples at each of 5
bentonite contents; 1, 3, 5, 7, and
9 percent by dry weight. The samples were
titrated with methylene blue solution
until endpoint and the data plotted on
Figure 7. Once again the data exhibited a
linear trend, but, as expected, also
demonstrated bias reflecting greater
exchange capacity than attributable to the
previously predicted 0.4 percent
76
-------
non-ben ton ite derived off-set on the
original calibration curve yielded a
"corrected calibrated curve" presented as
a dashed line on Figure 7. The original
bentonite calibration curve developed
during Phase I testing is also presented
for comparison.
Although the corrected calibration
does not exactly match the best fit line
through the actual data, it does agree to
+0.3 percent bentonite. These data there-
fore support the validity of superposition
in that the small discrepancy obtained
corresponds to inaccuracies incurred by
averaging the data upon which the
corrected calibration curve was based.
This error band represents less than +^6
percent of the 5 percent specified
bentonite and verifies that the data
amassed during the cutoff wall
construction at Gilson Road was in fact
sufficiently accurate for field QC use.
Field Use of the Methylene Blue Test
The laboratory program described
herein was used to assess the accuracy of
the methylene blue test and was therefore
structured to measure the unknown quantity
of methylene blue required to titrate
backfill samples with known bentonite
contents. The data thus developed
demonstrated the utility and accuracy of
the test for field QC in cases similar to
the Gilson Road project. During actual
use for QC in the field however, the
methylene blue consumption would be
measured and used to determine the unknown
bentonite content in the backfill via
pre-established calibrations. Thus, the
testing methodology used in field
applications initially would follow a
logic similar to the laboratory program
for establishment of a calibration curve,
but thereafter utilizes methylene blue
consumption as the independent variable
BENTONITE CALIBRATION CURVE
CORRECTED FOR BORROW
a SPOIL
BENTONITE
CALIBRATION
CURVE
BENTONITE EQUIVALENT
IN BORROW 8 SPOIL
(0.4%)
I 234 5678
% BENTONITE IN BORROW/TRENCH SPOIL/DISTILLED WATER
Figure 7. Titration for Bentonite Content in
Gilson Road Backfill Matrix
77
-------
and bentonite content as the dependent
variable. Implementation of the methylene
blue test for quality control therefore
consists of the following procedure:
1. Establish a bentonite calibration
curve representing the relationship
between methylene blue consumption
and bentonite content for
bentonite/distilled water suspensions.
The procedure would be similar to the
Phase I testing described herein
except that the methylene blue
consumption would be plotted on the
x-axis as the independent variable.
2. Determine if contaminants are likely
to be present in the backfill samples
and evaluate the probable effect on
the calibration.
3. Determine average cation exchange
capacity (methylene blue consumption)
of the non-bentonite backfill soils
combined in the same proportions they
will be mixed in the field. Use the
bentonite calibration curve to
determine the bentonite equivalent
for this volume.
4. Offset the initially determined
bentonite calibration curve to
account for the bentonite equivalent
of the non-bent onite soils and/or
contaminants to establish a field
calibration curve. Note that
multiple field calibration curves may
be required if significant variation
in the offset values are anticipated
due to variations inherent in the
site geology or construction
procedures.
5. Perform the methylene blue test on
backfill samples and enter the field
calibration curve with the volume of
methylene blue consumed at the end
point to determine bentonite content.
CONCLUSIONS
The work presented herein was
directed towards verification of the
precision, accuracy, and practicality of
two new tests for QC of S/B backfills.
This is part of a larger endeavor focused
on the development of standardized QC
guidelines concordant with a strategy of
real time field testing to ensure
attainment of an effective barrier during
its construction rather than resorting to
post-construction performance verification
based litigation. As based on this work,
it can be concluded that the methylene
blue test for percent bentonite and the
API hydraulic conductivity test both
satisfy the four basic criteria required
of quality control testing:
• They are both inexpensive and thus
allow a statistically significant
number of tests to be performed.
• They are s-imple thus lending the
procedures to execution in the field
rather than a remote laboratory.
• They are quick and therefore can
provide the timely data necessary for
real time quality control and,
• They are sufficiently precise and
accurate to yield a meaningful
representation of the expected
post-construction performance.
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 U.S. EPA. The authors would
therefore like to express appreciation to
Michael Donahue, Thomas Roy, and Paul
Heirtzler of the State of New Hampshire
and Stephen James, Walter Grube, and John
Herrmann of the U.S. EPA for their
continuing cooperation and assistance. In
addition, the authors would like to
acknowledge Katrina Bikis and Donald
Schulze without whose research
contributions this paper would not have
been possible.
REFERENCES
1 . Ay res , J .
Barvenik ,
Cutoff
2.
Specif icat
E., D.C. Lager, and M.J.
The First EPA Superfund
Wall: Design and
ions. Presented at the
Third National Symposium on Aquifer
Restoration and Groundwater
Monitoring (1983).
Lambe, T.W., Soil Testing for
Engineers
John Wiley & Sons, Inc.,
78
-------
Schulze, D., M.J. Barvenik, and J.E. Monitoring, Sponsored by NWWA,
Ayres , Design of the Soil/Bentoni te Columbus, Ohio, May 1984.
Backfill Mix for the First EPA
Superfund Cutoff Wall, Presented at 4. U.S. Army Corps of Engineers Manual,
the Fourth National Symposium on Laboratory Soils Testing, EM
Aquifer Restoration and Groundwater 1110-2-1906, November 1970.
79
-------
CONTROL OF FUGITIVE DUST EMISSIONS
AT HAZARDOUS WASTE CLEANUP SITES
Keith D. Rosbury
PEI Associates, Inc.
14062 Denver West Parkway
Golden, Colorado
Stephen C. James
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, Ohio
ABSTRACT
Three field studies were performed to determine the effectiveness of dust control
strategies at hazardous waste sites. A total of 12 dust suppressants were tested to
determine the effectiveness of fugitive dust control against wind erosion from exposed
areas. Based on a tracer sampling protocol, the suppressants were 100 percent effective
for 1 to 4 weeks after application, with declining control efficiencies thereafter. The
second wind erosion field study was an evaluation of the effectiveness of windscreens and
windscreen/dust suppressant combinations in controlling fugitive dust from storage piles.
Based on 82 tests, the windscreen showed no consistent significant level of control of
particles _<10 micrometers. The wind erosion emission rate of particles _<10 micrometers is
fairly constant at windspeeds above a threshold of about a 7 mph average. The windscreen
may control larger particles more effectively, but these particles are less of a threat
because they exceed the inhalable size and are not carried long distances in the wind.
Chemical stabilizers were superior in performance to the windscreen. The third field
study, investigating 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. Control
efficiencies of 41 to 77 percent were measured for area spraying and spray curtains. In
general, chemical dust suppressants are often wastes or by-products themselves. Before
applying a dust suppressant, the user should analyze the dust suppressant for toxic sub-
stances and test for reactivity between the dust suppressant and the contaminated soil.
INTRODUCTION Contaminated soil can be re-
entrained in the air by three basic mecha-
Spills, waste disposal, and various nisms:
industrial operations can result in the
contamination of land surfaces with toxic • wind erosion
chemicals. Soil particles from these • Reentrainment by moving vehicles
areas can be entrained into the air, (rubber-tired or tracked-vehi-
transported offsite via the wind, and cles) on soil or paved/unpaved
result in human exposure by direct inhala- roads
tion. Indirect exposure can result when • Active cleanup (movement of soil
particulates are deposited in agricultural by dozers, loading by front-end
fields, pastures, or waterways and enter loaders, etc.)
the human food chain.
80
-------
These three mechanisms can act individual-
ly or in any combination.
PURPOSE
Very few data are available to assist
in developing dust control programs at
cleanup sites. Field testing of dust
control effectiveness has been limited
largely to vehicle reentrainment of dust
from paved and unpaved roads. The intent
of this project was:
• To perform field demonstrations
of several products to determine
their effectiveness for control-
ling dust.
* To prepare a handbook on state-
of-the-art methods of dust
control.
Three field studies were performed.
The first involved the testing of 12 dust
suppressants to determine how effectively
they control fugitive dust resulting from
wind erosion of exposed areas. The second
wind-erosion 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, investiga-
ting 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.
A Dust Control Handbook was prepared
which included discussions of the fol-
lowing for each of the three basic reen-
trainment mechanisms:
Identification of dust producing
points
Principles of control
Product listing by name, ad-
dress, telephone number, dilu-
tion, application rate, basic
application method, and cost
Detailed application procedures
Product effectiveness
APPROACH
Control of Wind Erosion Dust Emissions
From Exposed Areas
Tracer sampling was chosen as the
methodology most closely paralleling the
requirements of this study. A tracer was
applied to the soil surface before a
commercial dust suppressant product was
applied to the plot. Any tracer-laden
particles found later in the ambient air
around the test plot indicated a failure
in the integrity of the crust formed by
the dust suppressant.
The test plots were located on a
small farm rear Cincinnati, Ohio. The
soil type in the test plots was clay and
quite uniform from plot to plot. The
plots were spaced several hundred feet
apart to eliminate cross contamination of
the plots by the tracer. The plots were
prepared by removing vegetation from 50 ft
x 50 ft areas with a bulldozer and grading
the plot to make it smooth. Application
of zinc sulfate (water-soluble), the
selected tracer, resulted in a surface
concentration of about 300 ppm. After
application of the tracer, the dust-con-
trol products were applied according to
the manufacturer's recommendations. Table
1 describes each test plot. No control
plots (tracer contaminated plots with no
dust suppressant) were established.
Emissions from the control plot would have
likely contaminated other test plots at
the site, resulting in a false indication
of dust suppressant failure.
Particulates being removed from the
plot by wind movement were sampled with
saltation catchers and surface vacuum
sampling. Four catchers (36-inch plastic
tubes with a 2-inch-wide vertical slot
that samples at 12- to 30-inch height
intervals) were placed at each plot; they
were oriented at the midpoint of each
side, with the sampling slot facing the
plot. Samples from the saltation catchers
were taken about once each week during the
test period and were analyzed for zinc by
atomic absorption spectroscopy. An area
of about 35 ft was vacuumed each week
from each test plot and placed onto a
glass fiber filter. The material was
removed from the filter and also analyzed
for zinc.
81
-------
TABLE 1. EXPOSED AREA TEST PLOTS
Name
Soil Seal
AMSCO-RES 4281
Supac 8 NP
Flambinder
Genaqua
Curasol
M166 & M1667
Coherex CRF
Sherman Process
(no grass seed)
Sherman Process
(with grass seed)
Terra Tack I
Material
cost/acre,
dollars
$1450
2350
4840
62
484
273
891
847
800
800
68
Formulation
Latex Acrylic Copolymer
Carboxylated styrene-
Butadiene Copolymer
Nonwoven Geotextile
Lignosulfonate
Vinyl Acetate Resin
Synthetic Resin
Latex
Petroleum resin
Straw mulch with
emulsified asphalt
Straw mulch with
emulsified asphalt
Vegetable Gum
Applica-
tion con-
centration*
3%
20%
8 oz/yd2
111 0.5
10%
3%
7%(M166) +
0.2%(M167)
25%
MA
MA
0.3%
Application
rate
1.0 gal/yd2
0.6 gal/yd2
12-ft rolls
gal/yd*
0.2 gal/yd2
0.3 gal/yd2
0.5 gal/yd2
0.5 gal/yd2
NA
MA
1.4 gal/yd2
NA = net applicable.
* Percent compound in aqueous solution, unless otherwise indicated.
Control of Hind Erosion Dust Emissions
From Storage Piles
This field study was designed to
demonstrate the effectiveness of a wind-
screen, alone or in combination with other
control measures, in reducing dust emis-
sions due to wind erosion from an inactive
waste storage pile. The windscreen was 8
feet high ard 75 feet in length with a
porosity of 50 percent. Control measures
tested in combination with the windscreen
were pile watering and chemical dust
suppressants (Curasol and Genaqua).
Concurrent upwind/downwind aerosol/wind-
speed measurements were made with real-
time data retrieval around a storage pile
protected with a windscreen.
The waste storage pile tested was not
contaminated. It was assumed that the
hazardous material would be integrally
bound to the soil and therefore would be
emitted at a rate directly proportional to
the loss rate of soil from the pile. The
soil selected was a very fine shredded
topsoil. The desired size of the pile was
determined to be large enough for accurate
simulation of wind erosion action from a
temporary waste storage pile, yet not so
large that building the pile consumed a
major portion of the study's resources.
This "reasonable" size was estimated to be
about 8 ft high with an elliptical base 25
ft by 20 ft, or about 100 tons of shredded
topsoil.
All atmospheric particulate concen-
trations were measured with RAM-1 contin-
uous aerosol monitors. These instruments
emit a pulsed light across a continuous
flowing sample airstream and sense the
amount of light scattering with a silicon
detector. The RAM-1's are stated to have
an upper particle size range of 20-ym
diameter, but are generally considered to
have a 50 percent collection efficiency
upper cut point of about 10 ym. Windspeed
and wind direction were measured with a
Met One system that included six windspeed
sensors and one direction vane. The
anemometer sensors had an accuracy of
±0.25 mph; the wind direction vane had an
accuracy of ±5 degrees.
Both the RAM-1's and wind instruments
were connected through a translator and an
analog-to-digital converter to an onsite
Apple He computer, which collected the
individual signals during testing; dis-
played the readings continuously; averaged
them over 1-minute, 5-minute, and 1-hour
time periods; and stored the averages for
later retrieval. All the calibration
values could be changed by inputting new
values to the computer.
Both the upwind sampling stand and
the main downwind stand had RAM-1's and
82
-------
windspeed sensors mounted at 3.1 and at
6.2 feet heights (Figure 1). The upwind
stand also had a wird direction vane
mourted at a 6.2 feet height. The third
sampling stand, which wes alternated
between a location beside the main down-
wind stand and a distance farther down-
wind, had a RAM-1 and windspeed sensor
only at a 3.1 feet height. A sixth wind-
speed sensor was located at the upwind
edge of the pile at a height of 3.1 feet.
Chemical 1 was Curasol. It was first
applied at a dilution of 4.0 percent at a
rate of 0.6 gal/yd2. On the next day,
Curasol was applied at thp same dilution
at 1.0 gal/yd2. Testing occured over a
period of 0 to 3 days since application.
Chemical 2 was Genaqua. It was applied et
a dilution of 10 percent at a rate of 0.90
gal/yd2. Testing was conducted over a
period of 0 to 1 days since application.
Watering occured at the rate of 0.9 gal/yd
at the beginning of each test day. Test-
ing occured at 0 to 6 hours since applica-
tion.
WIND
JIRECTION
UPWIND SAMPLING STAND
100 FEET
75 FT. WINDSCREEN
Control of Dust Emissions From Active Site
Cleanup
The operation selected for testing
consisted of a front-end loader (FEL) and
dump truck combination. The FEL scraped
material from the surface, turned, travel-
ed to the dump truck, and dumped its load
into the truck. This activity simulates
the most common method of loading contam-
inated soil into trucks for offsite dis-
posal .
The exposure profiling method was
used to sample the dust emissions downwind
of the operation. This method employs a
tower with multiple profiling heads to
perform simultaneous multipoint isokinetic
sampling over the cross section of the
plume. The primary sampling instruments
were profiling heads utilizing the stacked
filter concept. The sampling head inter-
nally fractionates the dust sample by
particle size. Three profiling towers
were used. One tower was located upwind
of the operation to measure the background
concentration not attributable to source
• RAM-1 AND WINDSPEED SENSOR
AT 3.1 AND 6.2 FEET HEIGHTS
* WIND DIRECTION SENSOR
D RAM-1 AND WINDSPEED SENSOR AT
3.1 FEET HEIGHT
O WINDSPEED SENSOR AT 3.1 FEET
HEIGHT
20-80 FEET
20 X 25 FT., 8 FT. HIGH
30 FT.
t
I
•D
Figure 1. Windscreen sampling configuration.
83
-------
operation. A similar sampling tower was
located downwind of the PEL scraping and
traveling path to measure the scraping
emissions. The other tower which was
equipped with a horizontal crossarm, was
located downwind of the dump truck to
measure the dump cycle emissions. On each
sampling day, both an uncontrolled and a
controlled test were run. The control
efficiency is merely the difference in
emission rate per unit of activity between
the uncontrolled srd controlled tests.
Four control measures were tested.
Control Measure 1 consisted of spraying
the active working area of the PEL and
dump truck with water. At those few
hazardous waste sites where dust control
measures are currently in use, this is the
control being used. A portable 200-gallon
tank, pump, generator, hose, and nozzle
were used to spray the working area prior
to the controlled test and again during
the test to allow the field team to note
drying of the surface and visible emis-
sions. Watering amounts averaged 0.9
gal/yd2 of surface area disturbed. More
water was used on the travel paths of the
PEL and dump truck than on the scraping
area. These active travel areas dried
much faster than the scraping area. It
was noted throughout the testing that only
the top 1 or 2 inches of the surface were
dry. Below this dry crust, the soil was
very damp.
For Control Measure 2, application
procedures were identical to those used in
the application of plain water; however, a
surfactant (Johnson-Marsh Compound MR) was
added to the water to form a 1:1000 dilu-
tion of surfactant to water. Somewhat
less watering was needed for these con-
trolled tests. Application of the water/
surfactant mixture averaged about 0.75
gallon/yd2 of surface area disturbed.
The concept behind Control Measure 3
was to create a flat spray or curtain that
would totally cover the top of the truck
box. The material is moistened as it
passes through the spray. More important-
ly, the dust that is pushed upward from
the bottom of the truck box by air turbu-
lence is caught in the spray. A frame was
built to support a line of six spray
nozzles on each side of the length of the
truck bed. The truck was then backed into
position between the standards and the
spray was turned on.
Because the spray heads were in
constant operation throughout the test
period, the amount of surfactant required
averaged about 1.5 gal/yd3 of material.
In retrospect, the spray probably should
have been operated only during the actual
dump. This would have entailed the use of
substantially less liquid. Although not
tested, it is believed that control values
would have been similar to those with the
continuous spray.
For Control Measure 4, the spray
system used for the water curtain was
modified to produce a foam spray. Four
foam-spray rozzles were used in this
array, one at each corner of the truck
bed. A mixture of water and surfactant,
approximately 30:1 dilution, was used to
generate a stream of foam. The system was
only turned on during the dump. About 0.4
gal/yd3 of mixture was used.
PROBLEMS ENCOUNTERED
Two problems were encountered with
the exposed area study. Initially, vacuum
surface samples of the test plots were
taken in addition to the saltation catcher
samples. The vacuum samples showed very
high concentrations of the tracer, which
led the investigators to believe that the
vacuum negative pressure was lifting the
dust-suppressant-formed crust. These data
were not used. The second problem encoun-
tered was the rapid growth of vegetation
on the initially bare plots. The plant
stems destroy the crust, as explained in
the results section.
The windscreen/storage pile emissions
testing program went very well. Use of
samplers with a particle size cut-off
greater than 10 micrometers would have
assisted in some of the data interpreta-
tion, however.
The active cleanup testing was conti-
nually hampered by wet weather, which kept
conditions at the test site fairly moist.
Additional engineering design refinements
-------
to both the surfactant and foam spray cur-
tains probably would have resulted in
•improved control efficiencies.
RESULTS
Control of Wind Erosion Dust Emissions
Erom Exposed Areas
Table 2 lists the results of the
saltation sample analyses. Values over 75
ppm (the background concentration of zinc
in the soil tested in each plot) indicate
failure of the dust suppressant.
The data indicate integrity of the
crust of all plots except Plots 5 and 8 on
June 11 . Time since application varied
from 17 to 31 days. Two weeks later, or
June 25, only Plot 7 remained near the
background level of zinc in the soil.
Plot 11, to which the dust suppressant and
tracer were applied later in the test
period, was also effective after 30 days.
No reapplications were tested; however, it
is reasonable to assume that control after
reapplication would be slightly greater
than with the initial application.
A problem with all plots was the
rapid growth of vegetation. All plots
were stripped of vegetation before ap-
plication of the tracer and dust-suppres-
sant. The rationale for this action was
that a dust suppressant spray could not
form a dirt crust in the presence of vege-
tation stems, because the dust suppressant
would not uniformly pass through the
vegetation and reach the soil. When
vegetation did grow, it penetrated the
crust, and a small dirt pile was evident
around each stem. Testing of these dust
piles indicated large quantities of the
tracer material (>200 ppm). When this
finding is applied to hazardous waste
sites, it is apparent that vegetation must
be controlled if the crust is to stay
intact. This can be done by spraying
preemergent weed control on the plot
before applying of the dust suppressant.
An alternate procedure to eliminating
vegetative growth would be to encourage
it. Products are available that are
temporary soil binders impregnated with
grass seed. While grass was beginning to
grow, the same weed problem just described
would occur. Assuming a thick stand of
grass did grow, control would probably not
be 100 percent, as some loose dirt would
always be present between grass stems.
Chemical dust suppressants sprayed on
thick grass stands would not be effective
because the suppressant would stick to the
stems and little would penetrate to the
soil.
Control of Wind Erosion Dust Emissions
From Storage Piles
/> total of 82 tests were taken—47
with the pile surface dry, 14 with water-
ing, and 21 with chemicals applied.
Hourly average windspeeds varied from 4.1
to 24.1 mph. The distance of the screen
from the pile dictated the amount of
windspeed reduction observed at the pile:
TABLE 2. SELECTED SALTATION SAMPLER RESULTS (ppm)
Test
plot
1
2
3
4
5
6
7
8
9
10
11
Date
establ ished
5/24/84
5/11/84
5/18/84
5/25/84
5/18/84
5/24/84
5/24/84
5/24/84
7/10/84
7/10/84
7/31/84
6-11
78
71
42
69
91
75
64
113
NS
NS
NS
6-25
97
114
163
105
172
111
76
162
NS
NS
NS
Sample date
7-13
111
121
125
116
152
130
96
100
NS
NS
NS
(1984)
7-30
155
322
<65
166
58
97
152
118
NS
NS
NS
8-22
183
172
252
140
141
167
160
206
513
157
48
8-31
170
204
120
119
217
128
122
214
132
103
72
NS = no samples.
85
-------
Screen-to-pile
dist.. ft.
20 (2.5 screen heights)
50 (6.3 screen heights)
80 (10 screen heights)
Av. windspeed,
% of incoming
30
50
P5
These reductions were fairly independent
of incoming windspeed. They were a1so
consistent with reductions reported pre-
viously in the literature.
The screen did reduce windspeeds by
the amount anticipated, but this did not
result in commensurate reductions in
particulate concentrations coming from the
pile. Without the windscreen in place,
net concentrations from the pile ranged
from 2 to 34 yg/m3, averaging 18 pg/m3, at
the 6.2 foot height. Reductions in net
particulate concentrations attributed to
the windscreen are shown in Table 3. On
the dry pile, the screen was able to
reduce wind erosion slightly if it was
placed close to the pile. The 2 meter
samplers at 50 or 80 ft. distance appeared
to be out of the area sheltered by the
screen, and possibly in an area of wind-
screen induced wind shear.
Taken in total, the results indicate
that the windscreen did not produce con-
sistent reductions in wind erosion. An
explanation for the windscreen's perfor-
mance, developed after review of all the
data from this study, is that wind erosion
emission rates in the less than 10 micro-
meter size range are fairly constant at
windspeeds above the threshold of about 7
mph (hourly average). The additional
emissions associated with high wind ero-
sion losses at high windspeeds are larger
particles that are not detected by the
RAM-l's. The windscreen may control
larger particles more effectively, but
these particles are less of a threat
because they exceed the inhalable size and
are not carried long distances in the
w'nd.
In tests with different pile-to-wind-
screen distances, it was found that a
distance of 2.5 screen heights was super-
ior to a distance of 6.3 to 10.0 screen
heights.
Chemical dust suppressants without
the windscreen were very effective in
reducing wind erosion over the four days
since the application was tested. The
tests of the windscreen in combination
with watering or chemical dust suppres-
sants inexplicably showed that the screen
was counterproductive.
In terms of dust control for an
inactive hazardous waste storage pile,
this study showed that chemical dust
suppressants provided the greatest and
most consistent reduction in wind erosion.
No firm evidence was uncovered to indicate
that windscreens provided any incremental
control beyond that of the chemicals when
the two controls were applied together.
Based on requirements for putting the
controls in place and their time of ef-
fectiveness, experience in this study also
shewed that chemical dust suppressants
were preferable to either windscreens or
watering, even if the piles must be re-
stabilized or a daily or weekly basis
because of intermittent activity. One
person using simple equipment could spray
TABLE 3. REDUCTION IN NET PARTICULATE CONCENTRATIONS
ATTRIBUTED TO A WINDSCREEN
Screen
dist., ft.
Surface condition
Est. reduction
due to screen, yg/m3
3.1 ft. ht.
0 6.2 ft. ht.
20
50 or 80
20 or 50
20 or 50
dry
dry
watered
chemicals applied
3 to 4
3 to 4
negative
-2
0 to 5
-8 to -10
neaative
-3 to -6
86
-------
the pile with chemicals in 20 minutes or
less. The windscreen took two persons
several hours for initial installation and
two persons 30 minutes to move the screen
when the wind shifted. A shift in wind
direction would require no change to the
chemically stabilized pile. No informa-
tion w?s obtained during this study on the
durability of the chemical crust, except
that chemical 1 was on the pile for 4 days
with no apparent reduction in effective-
ness. Watering was effective for only a
few hours.
This field study showed that the use
of windscreens probably would only be
worthwhile for dust control from an in-
active pile in situations where chemicals
could not completely stabilize the soil in
the pile and drifting of large particles
posed a threat of offsite contamination.
Control of Dust Emissions from Active Site
Cleanup
Water spraying over the area being
worked by the PEL and truck resulted in a
control efficiency of 42 percent for <30-
micrometer particles (TSP) arid 64 percent
for
-------
PRACTICAL METHODS FOR DECONTAMINATING BUILDINGS, STRUCTURES, AND
EQUIPMENT AT SUPERFUND SITES
M. P. Esposito, J. L. McArdle, A. H. Crone, and J. S. Greber
PEI Associates, Inc.
Cincinnati, Ohio
R. Clark, S. Brown, J. B. HalloweTI, A. Langham, and C. D. McCandlish
Battelle Columbus Laboratories
Columbus, Ohio
ABSTRACT
Practical methods for decontaminating buildings, structures, and equipment at Super-
fund sites are described in a new EPA handbook. Techniques presented include: asbestos
abatement, absorption, demolition, dismantling, dusting/ vacuuming/wiping, encapsulation/
enclosure, gritblasting, hydroblasting/waterwashing, painting/coating, K-20, scarifica-
tion, RadKleen, solvent washing, steam cleaning, vapor-phase solvent extraction, acid
etching, bleaching, flaming, drilling and spalling, microbial degradation, and photo-
chemical degradation. Method descriptions include a general discussion of the procedure,
advantages, disadvantages, applicability, effectiveness, engineering considerations (in-
cluding building preparation, process description, and equipment needs), safety require-
ments, waste disposal, and costs. Potential decontamination techniques for a given con-
taminant/structural material combination are identified in a matrix. Other information
presented in the manual includes stepwise guidance for developing a cost-effective decon-
tamination strategy, case studies illustrating the actual application of many decontami-
nation methods, cost analyses for the application of each method to a model building, a
discussion of worker health and safety precautions, and a summary of available sampling
techniques.
INTRODUCTION
Cleanup of abandoned hazardous waste
sites is the Nation's top environmental
priority in the 1980's. In a Superfund
reauthorization bill before the House
(HR 5640), a mandatory cleanup schedule is
being considered that would require EPA to
list 1600 sites on the National Priorities
List by January 1, 1988, and to begin
remedial investigations and feasibility
studies at each site within 6 months of
the date it is listed. This greater level
of activity will increase the need for
decontamination of buildings, structures,
and equipment.
In response to this growing need, the
U.S. EPA funded this study to develop a
general decontamination guide or handbook
for government personnel, cleanup con-
tractors, and other individuals respon-
sible for planning and executing decon-
tamination activities at Superfund sites.
As an initial step in the develop-
ment of the handbook, a survey of build-
ing and equipment decontamination prac-
tices at Superfund sites was conducted.
This survey revealed that the methods
used to remove contaminants from build-
ings, structures, and equipment are few
in number and rarely documented in de-
tail. For example, it is common practice
to steam clean equipment such as back-
hoes, bulldozers, and drilling augers,
but testing to verify that the contami-
nants of concern have been adequately
-------
removed is generally not performed. Con-
taminated buildings and structures are
seldom cleaned and returned to active
use. More often, they are closed and
barricaded to prevent further entry and
exposure, or they are torn down and
buried in secure landfills. Contaminated
underground structures such as tanks,
sumps, and sewers are sometimes filled in
place with concrete to prevent their
reuse. These findings clearly pointed to
the need for basic guidance material
dealing with the identification and
selection of appropriate decontamination
methods, as well as their application to
various contaminants and structural mate-
rials.
DECONTAMINATION METHODS AND THEIR
APPLICATION
The user's manual which was devel-
oped contains information on several
methods that have potential use for
decontaminating buildings, structures,
and equipment. The methods presented
range from classical techniques, which
have been extended in application to the
cleanup of hazardous waste sites, to
developmental concepts, which have shown
strong potential as decontamination
techniques in laboratory and initial
field demonstrations. Method descrip-
tions include a general discussion of the
procedure, advantages, disadvantages,
applicability, effectiveness, engineering
considerations (including building prep-
aration, process description, and equip-
ment needs), safety requirements, waste
disposal, and costs. The following
paragraphs briefly describe each example
method:
Asbestos abatement consists of four
techniques: removal, encapsulation,
enclosure, and special operations. In
removal operations, all friable asbes-
tos-containing building materials are
completely removed to eliminate the re-
lease of asbestos fibers into the air.
The other techniques leave the asbestos
fibers in place but limit potential ex-
posure levels through various treatment,
maintenance, and inspection procedures.
Absorption is widely used in indus-
trial settings to clean up chemical and
other liquid spills. It is also commonly
used by emergency response teams such as
fire departments to absorb accidental
spills on highways and other surfaces.
This method is most applicable immedi-
ately following liquid contaminant
spills. Spills rapidly penetrate most
surfaces, and absorbents act to contain
contaminants and prevent such penetra-
tion. Depending on the surface and time
elapsed since the spill, further decon-
tamination procedures may have to be
employed.
Demolition refers to the total de-
struction of a building, structure, or
piece of equipment. Specific demolition
techniques include complete burndown,
controlled blasting, wrecking with balls
or backhoe-mounted rams, rock splitting,
sawing, drilling, and crushing. Many of
these techniques have been employed for
nuclear facility decontamination and for
the clean up of military arsenals.
Dismantling refers to the physical
removal of selected structures (such as
contaminated pipes, tanks, and other
process equipment) from buildings or other
areas. Dismantling can be the sole activ-
ity of decontamination efforts (e.g.,
removal of contaminated structures from an
otherwise clean building), or it can be
used in the initial stage of a more com-
plex building decontamination effort
(e.g., removal of structures prior to
flaming, hydroblasting, or other cleanup
techniques).
Dusting/vacuuming/wiping is simply
the physical removal of hazardous dust and
particles from building and equipment
surfaces by common cleaning techniques.
Variations include vacuuming with a com-
mercial or industrial-type vacuum; dusting
off surfaces such as ledges, sills, pipes,
etc., with a moist cloth or wipe; and
brushing or sweeping up hazardous debris.
Dusting and vacuuming are applicable to
all types of particulate contaminants,
including dioxin, lead, PCB's, and asbes-
tos fibers, and to all types of surfaces.
Dusting/vacuuming/wiping is the state-of-
the-art method for removing dioxin-con-
taminated dust from the interior of homes
and buildings.
Encapsulation/enclosure physically
separates contaminants or contaminated
structures from building occupants and
the ambient environment by means of a
barrier. An encapsulating or enclosing
physical barrier may take different
89
-------
forms; among them are plaster, epoxy, and
concrete casts and walls. Acting as an
impenetrable shield, a barrier keeps
contaminants inside and away from clean
areas, thereby alleviating the hazard.
As a result, contamination of part of a
structure will not result in the contam-
ination of adjacent areas. Encapsulation
has been used on damaged asbestos insula-
tion, leaky PCB-contaminated electrical
transformers, and open maintenance pits
and sumps contaminated by heavy metals.
Gritblasting is a surface removal
technique in which abrasive materials
(such as sand, alumina, steel pellets, or
glass beads) are used for uniform removal
of contaminated surface layers from a
building or structure. Gritblasting has
been used since 1870 to remove surface
layers from metallic and ceramic sur-
faces, and is currently used extensively
throughout industry. For example, sand-
blasting is commonly used to clean the
surfaces of old brick and stone build-
ings. The gritblasting method is ap-
plicable to all surface contaminants
except some highly sensitive explosive
contaminants such as lead azide and lead
styphnate. This method is applicable to
all surface materials except glass,
transite, and Plexiglas.
Hydroblasting/watenjashing refers to
the use of a high-pressure (500 to 50,000
psi) water jet to remove contaminated
debris from surfaces. The debris and
water are then collected and thermally,
physically, or chemically decontaminated.
Hydroblasting has been used to remove
explosives from projectiles, to decontam-
inate military vehicles, and to decontam-
inate nuclear facilities. Hydroblasting
also has been employed commercially to
clean bridges, buildings, heavy machinery,
highways, ships, metal coatings, railroad
cars, heat exchanger tubes, reactors,
piping, etc. Off-the-shelf equipment is
available from many manufacturers and
distributors.
Painting/coating includes three types
of decontamination techniques: 1) the
removal of old layers of paint containing
high levels of toxic metals such as lead;
2) the use of fixative/stabilizer paint
coatings; and (3) the use of adhesive-
backed strippable coatings. In the first
technique, paint containing lead in excess
of 0.06 percent is removed from building
surfaces by commercially available paint
removers and/or physical means (scraping,
scrubbing, waterwashing). Resurfacing or
further decontamination efforts may be
necessary. The second technique involves
the use of various agents as coatings on
contaminated surfaces to fix or stabilize
the contaminant in place, thereby decreas-
ing or eliminating exposure hazards.
Potentially useful stabilizing agents
include molten and solid waxes, carbowaxes
(polyoxyethylene glycol), saligenin
(a,2-dihydroxy .toluene), organic dyes,
epoxy paint films, and polyester resins
(also see K-20). The stabilized contam-
inants can be left in place or removed
later by a secondary treatment. In some
cases, the stabilizer/fixative coating is
applied in situ to desensitize a contam-
inant (such as an explosive residue) and
prevent reaction or ignition during some
other phase of the decontamination process
(for example, to prevent explosions during
dismantling or demolition). In the third
technique, the contaminated surface is
coated with a polymeric mixture. As the
coating polymerizes, the contaminant
becomes entrained in the lattice or at-
tached to the polymer molecules. As the
polymer layer is peeled off, the residue
is removed with it. It may be possible,
in some cases, to add chemicals to the
mixture to inactivate the contaminants.
K-20 coating is a newly developed
commercial product sold as a building
decontaminant. It acts as a penetrant and
sealer that basically encapsulates and
immobilizes toxic contaminants in place.
Its effectiveness is not fully known.
Although it acts more like a barrier than
a detoxifier, early laboratory evidence
indicates K-20 may facilitate chemical
degradation as well as physical separa-
tion of some contaminants.
Scarification is a method that can
be used to remove up to an inch of sur-
face material from contaminated concrete
or similar materials. The scarifier tool
consists of pneumatically operated piston
heads that strike the surface, causing
concrete to chip off. This technique has
been used in the decommissioning of
nuclear facilities and in the cleanup of
military arsenals.
RadKleen is the trademark for a
commercial process involving the use of
Freon 113 (l,l,2-trichloro-l,2,2-tri-
90
-------
fluoroethane or C2F3C13) for solvent
extraction of soluble contaminants from
various materials. The solvent is pres-
sure-sprayed onto the surface, then
collected, treated, and recycled. The
RadKleen process is currently used for
cleaning radioactive material from vari-
ous surfaces. It has been applied to
chemical agents on small objects, and
thus field capability has been demon-
strated. Studies have been conducted for
agent-contaminated clothing materials,
such as polyester-cotton, Nomex, butyl
rubber gloves, and charcoal-impregnated
cloth. Although this method has not been
demonstrated for removing contaminants
from building surfaces, it looks very
promising.
Solvent washing refers to the ap-
plication of an organic solvent (e.g.,
acetone) to the surface of a building to
solubilize contaminants. This technique
has not yet achieved widespread use in
building decontamination, although it is
beginning to be used in the decommission-
ing of nuclear facilities. The method
needs further development in application,
recovery, collection, and efficiency.
The hot solvent soaking process has been
shown to be effective in decontamination
of PCB-contaminated transformers.
Steam cleaning physically extracts
contaminants from building walls and
floors, and from equipment. The steam is
applied through hand-held wands or auto-
mated systems, and the condensate is
collected in a sump or containment area
for treatment. This method is currently
used by explosives handling and manufac-
turing facilities. It has also been used
to remove dioxin-contaminated soil from
vehicles and drilling equipment in Times
Beach, Missouri.
Vapor-phase solvent extraction is a
method in which an organic solvent with a
relatively low boiling point (such as
methyl chloride or acetone) is heated to
vaporization and allowed to circulate in a
contaminated piece of equipment or en-
closed area. The vapors permeate the
contaminated materials, where they con-
dense, solubilize contaminants, and dif-
fuse outward. The contaminant-laden
liquid solvent is collected in a sump and
treated to allow recycling of the solvent.
This method has not yet been applied to
building decontamination, although it is
believed to have good potential.
Acid etching of a contaminated sur-
face is used to promote corrosion and
removal of the surface layer. Muriatic
acid (hydrochloric acid) is used to
remove dirt and grime from brick building
surfaces in urban areas and to clean
metal parts (e.g., pickle liquors from
metal finishing operations). The result-
ing contaminated debris is then neutral-
ized and disposed of. Thermal or chem-
ical treatment of the removed material
may be required to destroy the contami-
nant before disposal. Although this
technique is not known to have been
applied to chemically contaminated build-
ing surfaces, it is believed to have good
potential.
Bleaching formulations are applied
to a contaminated surface, allowed to
react with contaminants, and removed.
Application usually occurs in conjunction
with other decontamination efforts, such
as the use of absorbents and/or water-
washing. Bleach has been used as a
decontaminant against mustard, G and V
chemical agents, and (experimentally)
organophosphorous pesticides.
Flaming refers to the application of
controlled high temperature flames to
contaminated noncombustible surfaces,
providing complete and rapid destruction
of all residues contacted. The flaming
process has been used by the Army to
destroy explosive and low-level radio-
active contaminants on building surfaces.
Its applicability to other contaminants
is not well known. This surface decon-
tamination technique can be used on
painted and unpainted concrete, cement,
brick, and metals. Subsurface decontam-
ination of building materials may be
possible, but extensive damage to the
material would probably result. This
technique can involve high fuel costs.
Drilling and spalling can remove up
to 2 inches of contaminated surface
material from concrete or similar mate-
rials. This technique consists of drill-
ing holes (1 to 1-1/2 in. diameter)
approximately 3 inches deep. The spall-
ing tool bit is inserted into the hole
and hydraulically spreads to spall off
the contaminated concrete. The technique
can achieve deeper penetration (removal)
91
-------
of surfaces than other surface-removal
techniques and it is good for large-scale
applications. However, the treated
surface is very rough and coarse and may
require resurfacing (i.e., capping with
concrete). The drilling and spa!ling
method has been used in the decommission-
ing of nuclear facilities.
Microbial degradation is a develop-
ing process whereby contaminants are
biologically decomposed by microbes
capable of utilizing the contaminant as a
nutrient source. Conceptually, microbes
are applied to the contaminated area in
an aqueous medium and allowed to digest
the contaminant over time; the microbes
are then destroyed chemically or ther-
mally, and washed away. Microbial deg-
radation as a building decontamination
technique is not well developed.
Photochemical degradation refers to
the process of applying intense ultra-
violet light to a contaminated surface
for some period of time. Photodegrada-
tion of the contaminant follows. In
recent years, attention has been focused
on this method because of its usefulness
in degrading chlorinated dioxins (TCDD in
particular). Three conditions have been
found to be essential for the process to
proceed: 1) the ability of the compound
to absorb light energy; 2) the availabil-
ity of light at appropriate wavelengths
and intensity, and 3) the presence of a
hydrogen donor.
Potential decontamination techniques
for a given contaminant/structural mate-
rial combination may be identified with
the aid of a matrix (see Table 1). Each
cell in the matrix represents a specific
contaminant/surface combination and
contains numbers corresponding to decon-
tamination methods just described that
either have been used in the specific
interactions or have the potential for
such use. Each method may be used alone
or possibly in conjunction with one or
more of the other procedures. The reader
is cautioned to remember that although
these decontamination methods have been or
could be used for the mix of the condi-
tions described in the matrix, the effec-
tiveness of the application in many cases
is not known or has not been adequately
documented or verified. Therefore, it is
recommended that the method descriptions
presented in the manual be used to eval-
uate each potential method on a site-
specific basis for efficiency, wastes
generated, equipment and support facil-
ities needed, time and safety require-
ments, structural effects, and costs.
Also, pilot-scale testing of each method
(or combination of methods) should be
performed prior to full-scale implementa-
tion to determine the effectiveness of
the strategy.
OTHER INFORMATION
In addition to the method descrip-
tions, the manual presents guidelines for
developing a successful decontamination
strategy, as outlined in Figure 1. Case
studies illustrating the actual applica-
tion of many of the decontamination
methods described earlier are also in-
cluded in the manual. Table 2 shows a
summary of these case studies, indicating
the contaminants present and the decon-
tamination methods used in each case.
Finally, the manual includes cost anal-
yses for the application of each method
to a model building, a discussion of
worker health and safety precautions, and
information on available sampling
methods.
CONCLUSIONS
The handbook provides much of the
guidance needed by Superfund site cleanup
personnel for decontaminating buildings,
structures, and equipment. However, many
of the decontamination methods currently
in use have not been verified and docu-
mented with respect to their effective-
ness on various contaminant/material
combinations. For example, the degree to
which steam cleaning removes dioxin-con-
taminated soil particles from drilling
augers has not been established. .Addi-
tional research to demonstrate the ef-
fectiveness of available methods under
various conditions is badly needed.
Also, decontamination methods that have
not previously been applied to specific
contaminant/substrate combinations, but
show a strong potential applicability,
should be validated in pilot investiga-
tions. Additions/deletions to the matrix
should be made accordingly. New decon-
tamination technologies that become
available also should be evaluated and
added to the matrix.
92
-------
TABLE 1. EXAMPLES OF PRACTICAL DECONTAMINATION METHODS FOR
VARIOUS CONTAMINANTS AND STRUCTURAL MATERIALS9
Contaminant
Asbestos
Acids
Alkalis
Dioxins
Explosives
Heavy metals and
cyanide
Low-level
radiation
Organic solvents
Pesticides
PCB's
Material
Brick
1,3,4,5,6
2,3,4,6,7,8,
9,13
2,3,4,6,7,8,
9,13,16
2,3,4,5,6,7,
9,11,12,13,
14,19,20,21
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,5,6,7,
8,9,12,13,14,
15,19
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,6,7,8,
9,11,12,13,
14,20,21
2,3,4,5,6,7,
8,9,11,12,13,
14,16,19,20,
21
2,3,4,5,6,7,
8,9,11,12,13,
14,19,20,21
Concrete
1,3,5,6
2,3,6,7,8,9,
10,13,18
2,3,6,7,8,9,
10,13,15,18
2,3,5,6,7,9,
10,11,12,13,
14,18,19,20,
21
2,3,6,7,8,9,
10,11,12,13,
14,17,18
2,3,5,6,7,8,
9,10,12,13,
14,15,18,19
2,3,6,7,8,9,
10,11,12,13,
14,17,18
2,3,6,7,8,9,
10,11,12,13,
14,18,20,21
2,3,5,6,7,8,
9,10,11,12,
13,14,16,18,
19,20,21
2,3,5,6,7,8,
9,10,11,12,
13,14,18,19,
20,21
Glass
1,3,4,5,6
2,3,4,6,9,13
2,3,4,6,9,13
15
2,3,4,5,6,9,
11,12,13,14,
20,21
2,3,4,6,11,
12,13,14
2,3,4,5,6,12,
13,14,15
2,3,4,5,6,11,
12,13,14
2,3,4,6,9,11,
12,13,14,20,
21
2,3,4,5,6,9,
11,12,13,14,-
20,21
2,3,4,5,6,11,
'12,13,14,20,
21
Metal
1,3,4,5,6
2,3,4,6,7,8,
9,13
2,3,4,6,7,8,
9,13,15
2,3,4,5,6,9,
11,12,13,14,
20,21
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,5,6,7,
8,9,12,13,14,
15
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,6,7,8,
9,11,12,13,
14,20,21
2,3,4,5,6,7,
8,9,11,12,13,
14,16,20,21
2,3,4,5,6,7,
8,9,11,12,13,
14,20,21
Plastic
1,3,4,5,6
2,3,4,6,9,13
2,3,4,6,9,13
2,3,4,5,6,9,
11,12,13,14,
20,21
2,3,4,6,9,11,
12,13,14
2,3,4,5,6,9,
12,13,14
2,3,4,6,9,
11,12,13,14
2,3,4,6,9,11,
12,13,14,20,
21
2,3,4,5,6,9,
11,12,13,14,
20,21
2,3,4,5,6,9,
11,12,13,14,
20,21
Wood
1,3,4,5,6
2,3,4,6,7,9,
13
2,3,4,6,7,9,
13,15
2,3,4,5,6,9,
11,12,13,14,
19,20,21
2,3,4,6,7,9,
11,12,13,14
2,3,4,5,6,7,
9,12,13,14,
15,19
2,3,4,6,7,9,
11,12,13,14
2,3,4,6,7,9,
11,12,13,14,
20,21
2,3,4,5,6,7,
9,11,12,13,
14,16,19,20,
21
2,3,4,5,6,7,
9,11,12,13,
14,19,20,21
Equipment
and
auxi 1 lary
structures
1,3,4,5,6
2,3,4,6,7,8,
9,13
2,3,4,6,7,8,
9,13,15
2,3,4,5,6,7,
8,9,11,12,13,
14,20,21
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,5,6,7,
8,9,12,13,14,
15
2,3,4,6,7,8,
9,11,12,13,
14,17
2,3,4,6,7,8,
9,11,12,13,
14,20,21
2,3,4,5,6,7,
8,9,11,12,13,
14,16,20,21
2,3,4,5,6,7,
8,9,11,12,13,
14,20,21
Key for decontamination methods:
1. Asbestos abatement
2. Absorption
3. Demolition0
4. Dismantling
5. Dusting/vacuuming/wiping
6. Encapsulation
7. Gritblasting6
8. Hydroblasting/waterwashing
9. Painting/coating
p f
10. Scarification '
11. RadKleen
12. Solvent washing
13. Steam cleaning
14. Vapor-phase solvent extraction
15. Acid etching
16. Bleaching
17. Flaming
18. Drilling and spelling '
19. K-20 coating
20. Microbial degradation
21. Photochemical degradation
Refer to individual method descriptions in Section 4 to determine whether an indicated technique has actually been used
to treat a particular contaminant/structural material combination, or whether the method is viewed as potentially
applicable.
Applicable only to liquids.
c Some contaminant residues (e.g., asbestos, explosives, toxic residues) may have to be neutralized, stabilized, or removed
prior to demolition to prevent explosions or emissions.
Applicable only to particulates and solids.
Not recommended for removing highly toxic residues or highly sensitive explosives, unless particulates can be controlled.
Applicable only to concrete.
93
-------
Figure 1. Flow diagram illustrating sequence of steps for developing a decontamination strategy.
-------
TABLE 2. SUMMARY OF CASE STUDIES
Site
Contaminants present
Decontamination methods
Homes and other buildings
Seveso, Italy
State Office Building
Binghamton, New York
Sontag Road area8
St. Louis County, Missouri
One Market Plaza Office Complex
San Francisco, California
Frankford Arsenal
Philadelphia, Pennsylvania
Office building
New England
Luminous Processes, Inc.3
Athens, Georgia
Chemical Metals Industries,
Inc.
Baltimore, Maryland
TCDD
PCB's, TCDD, TCDF
TCDD
PCB's, PCDD, PCDF
Explosives
Asbestos
Radiological residues
Heavy metals
Asbestos
Low-level radiation
Heavy metals, acids,
alkalis, cyanide-
and ammonia-bearing
compounds, salts, and
solids and sludges of
unknown composition
Dusting/vacuuming/wiping
Painting/coating
Dismantling
Demolition
Dusting/vacuuming/wiping
Dismantling
Dusting/vacuuming/wiping
Insulation removal
Scrubbing (equipment only)
Steam cleaning (equipment
only)
Insulation removal
Dusting/vacuuming/wiping
Solvent washing
Scraping
Painting/coating
K-20
Gritblasting
Scarification/jackhammering
Dismantling
Hydroblasting/waterwashing
(equipment only)
Flaming
Demolition
Asbestos removal
Dusting/vacuuming/wiping
Hydroblasting/waterwashing
Scarification
Gritblasting
Dismantling
Painting/coating
Asbestos encapsulation
Paint stripping/sanding
Hydroblasting/waterwashing
Dismantling
Gritblasting
Dismantling
Superfund site.
95
-------
ACKNOWLEDGMENTS
This work was funded by the U.S.
Environmental Protection Agency Hazardous
Waste Engineering Research Laboratory,
Land Pollution Control Division,
Cincinnati, Ohio, under Contract No.
68-03-3190. Ms. Naomi P. Barkley was the
Project Officer.
SELECTED BIBLIOGRAPHY
Battelle Columbus Laboratories. 1979.
Final Report on Evaluation of Encap-
sulants for Sprayed-on Asbestos-Con-
taining Materials in Buildings. Prepared
for U.S. Environmental Protection Agency
under Contract No. 68-03-2552.
Benecke, P., et al. 1983. Development
of Novel Decontamination and Inerting
Techniques for Explosives-Contaminated
Facilities. Phase I - Identification and
Evaluation of Concepts. Vols. 1 and 2.
DRXTH-TE-CR-83all.
Crosby, D. G. 1983. Methods of Photo-
chemical Degradation of Halogenated
Dioxins in View of Environmental Recla-
mation. In: Accidental Exposure to
Dioxins. Human Health Aspects. Academic
Press, Inc., New York.
Dempsey, K. B. 1982. Biotechnology Aids
Disposal. Plants, Sites and Parks,
September/October, pp. 1-8+.
Furhman, D. 1983. Decontamination
Operations at Gateway Army Ammunition
Plant. DRXTH-AS-CR-83250. 1979.
Hawthorne, S. H. 1982. Solvent Decon-
tamination of PCB Electrical Equipment.
In: IEEE Conference Proceedings, Vol.
IA-18.
Jones, W. E. 1982. Engineering and
Development Support of General Decon
Technology for the U.S. Army's Intallation
Restoration Program. Task 5, Facility
Decontamination. Defense Technical In-
formation Center, Alexandria, Virginia.
Pub. No. 49-5002-0005.
Liberti, A., et al. 1978. Solar and UV
Photodecomposition of 2,3,7,8-Tetrachloro-
dibenzo-p-dioxin in the Environment. The
Science of the Total Environment, 10:
97-104.
Marion, W. J., and Thomas, S. 1980.
Decommissioning Handbook. DOE/EV/101281.
Natale, A., and H. Levins. 1984. Asbes-
tos Removal and Control. An Insiders
Guide to the Business. Sourcefinders,
Vorhees, New Jersey.
U.S. Environmental Protection Agency.
1979. Asbestos-Containing Materials in
School Buildings: A Guidance Document.
Part 1.
U.S. Environmental Protection Agency.
1983. Guidance for Controlling Friable
Asbestos-Containing Material in Buildings.
EPA-560/ 5-83-002.
Wolfe, H. R., et al. 1976. Some Problems
Related to Cleanup of Parathion-Contami-
nated Surfaces Following Spillage. In:
Proceedings of the 1976 National Confer-
ence on Control of Hazardous Material
Spills. EPA-600/5-76-002.
Wong, A. S., and D. G. Crosby. 1978.
Decontamination of 2,3,7,8-Tetrachloro-
dibenzo-p-dioxin by Photochemical Action.
In: Dioxin: Toxicological and Chemical
Aspects. Spectrum Publications, Inc., New
York.
96
-------
IN SITU TREATMENT OF CONTAMINATED GROUNDWATER AND SOILS,
KELLY AIR FORCE BASE, TEXAS
Roger S. Wetzel, Susan M. Henry, Philip A. Spooner
JRB Associates
McLean, Virginia 22102
Stephen C. James
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Edward Heyse
U.S. Air Force Engineering and Services Laboratory
Tyndall Air Force Base, Florida
ABSTRACT
In situ biological degradation has been selected for field demonstration at a waste
disposal site at Kelly Air Force Base, Texas. The in situ biodegradation treatment of
this site will involve the introduction of oxygen and nutrients to the subsurface and
the circulation of groundwater by an extraction/infiltration system. Extensive investiga-
tion has been necessary to characterize the site, demonstrate the effectiveness of the
biological treatment in the laboratory, and determine the modifications of site conditions
required to approach the optimum for treatment. The four elements of the project necessary
to provide the information required for the design and implementation of the field demon-
stration are 1) geological investigation and site characterization, 2) determination of
the contaminant profile, 3) microbiological investigation, and 4) laboratory biodegradation
investigation. The results of each element of the project are summarized and recommenda-
tions for design and implementation of the field demonstration are made based on these
results.
INTRODUCTION
In situ biodegradation of contami-
nated soil and groundwater is based on
the concept of stimulating the indigenous
subsurface microbial population to degrade
organic contaminants. To enhance biodegra-
dation it is necessary to "correct" condi-
tions that may limit bacterial growth and
metabolism. In the subsurface environment
oxygen is usually the limiting condition
for aerobic biodegradation, although
inorganic nutrients such as nitrogen,
phosphorus and trace metals, and sometimes
organic carbon, can also be limiting.
Optimum conditions can be provided by
injecting an oxygen source and nutrients
into the subsurface as required, and circu-
lating the groundwater using an injection
system and extraction wells.
The in situ approach has some dis-
tinct advantages over conventional "pump
and treat" technologies which involve
pumping the water to the surface for
treatment. First, because the active
treatment zone is in the subsurface, in
situ biodegradation has the potential to
remove contaminants sorbed to the soil
matrix. Second, in situ biodegradation
requires minimal equipment, i.e., that
associated with pumping, nutrient mixing
and injection, and oxygen addition.
"Pump and treat" technologies are general-
ly more expensive because of the cost of
an aboveground treatment system.
The most common problem encountered
in previous in situ biodegradation work
97
-------
has been the limited amount of oxygen
that can be supplied by aeration to the
subsurface microorganisms (3, 4). Hydro-
gen peroxide (^02) is now being invest-
igated as an alternative oxygen source.
Although hydrogen peroxide is cytotoxic
at higher concentrations, studies indicate
that it can be added to groundwater at
concentrations up to 100 ppm without
adverse effects assuming proper formula-
tion and application control (5). This
amount provides five times more oxygen
than can be introduced by aeration alone.
The Air Force has selected a waste
disposal site at Kelly Air Force Base,
Texas, for the field demonstration of an
in situ technology. The site was origin-
ally used as a disposal pit for chromium
sludges and other electroplating wastes.
Prior to being closed in 1966, the site
had been used as a chemical evaporation
pit for chlorinated solvents, cresols,
chlorobenzenes, and other chemicals.
JRB has chosen in situ biological
degradation as most suitable for demon-
stration at the Kelly site. The site
offers a number of advantages with respect
to conducting an in situ biodegradation
demonstration project, namely small size,
near optimum soil and groundwater tempera-
tures, the presence of biodegradable
organics, presence of a perched aquifer
not used for drinking, and a highly
adaptive and substantial microbial popula-
tion. The in situ treatment will involve
providing nutrients, oxygen in the form
of hydrogen peroxide to the subsurface,
and circulating water by use of an extrac-
tion/infiltration system.
PURPOSE
The purpose of this project is to
evaluate the feasibility and effectiveness
of in situ biological degradation as a
hazardous waste disposal site remediation
technique. Field demonstrations of in
situ biodegradation using hydrogen peroxide
as an oxygen source are being conducted.
However, all in situ biodegradation
projects involving the use of hydrogen
peroxide have been associated with gasoline
leaks or spills of one or a relatively
few chemicals. No information is available
on the feasibility of using hydrogen
peroxide as an oxygen source to enhance
biological degradation in the presence of
the complex variety of contaminants which
can be present at hazardous waste disposal
sites. Indeed, little work has been done
to evaluate the effectiveness of in situ
biodegradation at hazardous waste disposal
sites.
APPROACH
This project has included a feasi-
bility study and a laboratory study to
evaluate the potential in situ treatment
options and optimize their design. This
is being followed by the design, installa-
tion, startup, and monitoring of the field
demonstration of an in situ treatment
system.
The investigative work which has been
required for the design and implementation
of the field demonstration has involved the
following four elements:
Geological investigation and site
characterization
Determination of the extent and
nature of contamination
Microbiological investigation
Laboratory biodegradation research.
The geological investigation required
to characterize the site has included the
determination of site stratigraphy and site
hydrology, which is being completed, and
permeability studies which are currently
being conducted by K..W. Brown Associates.
Priority pollutant analysis according
to EPA recommended protocol for organics
and inorganics has been completed for five
borehole and three groundwater locations.
Further sampling and analysis may be
required based on requirements for design
of the demonstration system.
Microbiological investigation carried
out at Memphis State University involved
enumerating indigenous microorganisms using
a direct-count procedure, acridine orange
epiflourescence, and a viable agar plate-
counting procedure using seven different
media. These enumerations were conducted
to determine if viable bacterial popula-
tions existed at the site, and if there
was any correlation of bacterial number
with soil type, depth, location, and
contaminant concentrations.
The laboratory biodegradation study
is nearing completion. This study is
98
-------
necessary to determine, under given
treatment conditions, what will biodegrade,
the rates of biodegradation, the intermedi-
ate metabolites and metabolic products,
and any problems that may be encountered
with a particular treatment approach.
The study has involved GC/MS analyses of
240 ml closed microcosms consisting of
soil/groundwater slurries. Microcosms
are being sacrificed for analysis on days
0, 25, 50, and 100. Aerobic and anaerobic
biodegradation is being evaluated accord-
ing to the following amendment regime*
aerobic: \\2®2 + nutrients
aerobic control: sterilized
H202 + nutrients
aerobic: oxygen + nutrients
anaerobic + nutrients
anaerobic control: sterilized
anaerobic + nutrients.
RESULTS
Geological Characterization
The characterization of the site has
been complicated by the stratigraphic
heterogeneity of the on-site soils. The
high degree of variability, including
gravel lenses and layers, has made it
very difficult to construct subsurface
profiles. It also compounds the difficulty
of predicting the flow of contaminants in
the subsurface, thus complicating the
approach for treatment of the site.
The site under study is an elevated
peninsula-like landform, surrounded on
three sides by surface drainage (Figure 1).
Leon Creek borders the site on the west,
and an unnamed surface drainage ditch
borders it on the east and south. The site
is approximately 15 to 20 feet above this
drainage, and has steeply sloping sides.
It will be difficult to isolate the site
from surface waters and to prevent contami-
nated groundwater from exiting the sides
of the site. Only the very dry weather
conditions prevent such losses at this
time.
There is an abundance of fine-grained
soils and sludges at this site. Although
*The hydrogen peroxide used was Restore™
105 Microbial Nutrient supplied by FMC
Aquifer Remediation Systems.
no hard data are yet available, the hydrau-
lic conductivity of these materials can
be expected to be low. This means that
the task of delivering treatment solutions
to the contaminated areas is likely to be
difficult and lengthy.
The problems with particle size also
extend to coarser materials that are
present. The gravel layers and lenses with-
in the site are several orders of magnitude
more permeable than the fines. The differ-
ence in the rate at which these materials
will accept infiltrating fluids will
greatly complicate the design of an
infiltration network.
The perched aquifer beneath this site
is relatively thin, ranging from four to
eight feet in thickness, and exhibits
seasonal fluctuations. Cross Section
C-C', shown in Figure 2, illustrates the
perched aquifer, as well as the correla-
table Navarro Formation beneath it, and
the uncorrelatable alluvial deposits
above it. Although the aquifer is small,
water levels beneath the site may have to
be raised for effective treatment. Because
the water-bearing materials are likely to
hydraulically connect with the adjacent
•surface waters, some form of hydraulic or
physical barrier, such as an extraction
well system or a slurry wall, would be
required to contain the groundwater for
treatment.
Contaminant Profile
Priority pollutant analyses of soil
and groundwater samples indicate extensive
contamination with a wide variety of organ-
ics and inorganics. Table 1 summarizes
results of analyses for organics conducted
on soil samples from 5 boreholes and
water samples from 3 wells. This is not
a complete listing of organics detected,
but lists some representative compounds
present at high concentrations, and the
sum of some compounds by class.
The principle priority pollutants
found in soil samples at levels over 10
ppm include 1,2-dichlorobenzene and 1,4-
dichlorobenzene. Priority pollutants found
in groundwater include tetrachloroethylene
(PCE) as high as 15 ppm, trichloroethylene
(TCE) as high as 6.7 ppm, and trans-1,2-
dichloroethylene as high as 7.7 ppm.
99
-------
/ g UNNAMED
DRAINAGE
a MONITORING WELL
OBOREHOLE
Figure 1. Site Plan with Cross Section Location.
100
-------
EL
C
BH-7 B
-•30
-•20
-•10
-•00
• 90
-••o
-•70
EVAT
C'
."B BH-12
"^^^^
^^^
INTER
SOIL
V
r^
LAYERED FINE-Q
8 WITH QRAVEL L
— — _ _ SATURATED ZONE
MULTICOLORED YELLOW-BR
GREEN AND BLACK SILT LO
DARK BLUE-QREY CLA1
ION
RAINED
AYERS
OWN
AM
' DARK OREENI8H-QREY
LOAMY FINE SAND
I.I,
100
Figure 2. Cross Section C-C'
Petroleum-based hydrocarbon concentrations
were particularly high in soil samples
from the shallow sampling depth of bore-
hole 5 and the deep sampling depth of
borehole 1, at concentrations of 879 ppm
and 221 ppm respectively. Groundwater
from well AA contained 3595 ppm hydro-
carbons.
The main inorganics present in all
soil samples at concentrations greater
than 10 ppm are antimony, chromium,
copper, lead, nickel, and zinc. In 3 of
the soil samples analyzed chromium concen-
trations exceeded 340 ppm. Cadmium and
silver were present as high as 8 ppm, and
mercury as high as 0.3 ppm. Although
concentrations in groundwater were gener-
ally low and did not exceed 1 ppm, analyses
indicated contamination levels slightly
exceeding the National Interim Primary
Drinking Water Standards for cadmium,
chromium, and lead.
The distribution of contaminants
throughout the site is extremely hetero-
genous and a defined plume of contamination
has not been identified. However, based
upon the high concentrations of organics,
particularly hydrocarbons, present in the
samples of boreholes 4 and 5 and at depth
in borehole 1, it is surmised that migra-
tion is occuring, carrying contaminants
from the former evaporation pit around
boreholes 4 and 5 into the saturated zone
in the area of borehole 1 (Figure 1).
Microbiological Investigations
Subsurface samples were taken and
shipped to the Memphis State University
laboratory for direct and viable cell
counts. Acridine orange direct cell
counts were performed according to the
method of Ghiorse and Balkwill (1).
Estimates of viable bacteria density were
made using dilution-spread plate procedures
on seven media: two nutient rich, two low-
nutrient, and three dilute groundwater
agar. Direct count estimates ranged from
7.6 x 10^ to 1.7 x 10^ cells per wet gram
sample. Viable cell counts ranged from
less than 100 to 7 x 106 colony forming
units per wet gram sample. The data in-
dicated that pollutants present did not
reduce the number of total bacteria
compared to other sites. Of particular
note was the apparent non-selectivity of
any of the seven media. All media yielded
similar numbers of cells for any one
101
-------
TABLE 1. REPRESENTATIVE CONTAMINANT CONCENTRATIONS, PPM
Tetrachloro- Trichloro- 1,1,-Trichloro- Trans 1,2-Dichloro- Chloro- Dichloro- Alkyl- Hydrocarbons
ethylene ethylene ethane ethylene benzenes benzenes benzenes Toluene Ketones n-Alkanes Branched
Soil
Borehole
11
11
11
#1 0.001
#2
#3 0.022
#4
#5 0.019
0.01 0.10
0.016 0.97
-
0.91
0.082
0.093
0.15
-
0.16
0.012 38.
0.005 0.038 0.037 40
0.005 0.30
1.0
0.020 0.022 2.2 58
78. 0.13 26. 454
158.
.
0.009
0.11
456.
Groundwater
Wei 1 : AA
" BB
11 CC
-
15.
0.13
0.076
0.12 0.12
6.7
7.7 - 0.26
0.13 - 0.005
2.2 - <0.001
50 - <1.0 1145
<1.0
<1.0
2450.
_
-
Note: Values represent highest concentrations found at each borehole. Three soil samples from each borehole were analyze'd.
-------
sample. This lack of selectivity may
indicate the presence of highly adaptive
populations of bacteria. There was
little variation noted as a function of
either site location, depth, or pollutant
concentration.
Laboratory Biodegradation Study
.The previously completed contamina-
tion profile and microbiological investi-
gation were used in planning subsurface
sampling for the laboratory biodegradation
study. Subsurface samples were sent to
the ORB laboratory, where both aerobic
and anaerobic microcosms were assembled.
The results of the lab studies as of the
day 50 analysis indicate that biological
degradation was occurring in the test
microcosms. Significant differences were
evident between the aerobic and anaerobic
microcosms in terms of the nature of the
compounds degraded. The patterns of
degradation correlate well with the
existing literature. Compounds which
degraded in the aerobic microcosms include
the hydrocarbons, aromatics, and the
higher molecular weight halogenated
organics such as chlorobenzene. Compounds
which degraded anaerobically but not aero-
bically were most of the lower molecular
weight halogenated hydrocarbons, including
PCE and TCE. The compounds 1,1-dichloro-
ethane and 1,1-dichloroethene disappeared
by day 50 in both the aerobic and anaero-
bic microcosms, although rates were
generally faster in the aerobic microcosms.
Table 2 presents selected results of
the GC analysis of the base neutral/ acid
extractable (BN/A) fractions of the micro-
cosm soil samples. (Concentrations of
organics in the microcosm water fractions
were repeatedly so low that BN/A analysis
of this fraction was discontinued). The
results are characteristic of hydrocarbon
biological degradation patterns. The
concentrations of total resolved hydro-
carbons, representing preferentially
biodegraded n-alkanes, were reduced rapidly
in the aerobic microcosms, and remained
essentially unchanged in the anaerobic
microcosms. Concentrations of total
unresolved hydrocarbons, representing
less preferentially degraded branched
alkanes, were also degraded aerobically
but remained unchanged in the anaerobic
microcosms. The ratio of unresolved to
TABLE 2. CONCENTRATIONS OF HYDROCARBONS IN
MICROCOSM SOIL FRACTIONS
HYDROCARBONS, PPNI
Day-Treatment
Total Resolved
Hydrocarbons
X* CV**
Total Unresolved
Hydrocarbons
X CV
Ratio of
Unresolved to Resolved
X CV
1 -
24 -
49 -
1 -
24 -
49 -
1 -
25 -
50 -
*
**
***
H202
H202
H202
Oxygen
Oxygen
Oxygen***
Anaerobic
Anaerobic
Anaerobic
20.3
14.0
6.91
19.6
9.78
3.52
12.5
14.4
13.5
All values are the
14
19
23
5
23
67
33
57
25
205
176
102
151
170
71.0
89.5
132
119
22
16
22
15
16
61
27
51
35
means of three replicates
CV: coefficient of variation.
Means of two
repl i
cates.
(Standard
10
12
14
7
17
20
7
9
8
.1
.6
.8
.71
.7
.9
.26
.60
.69
10
5
3
10
15
8
9
12
13
unless otherwise noted.
deviation/mean
x 100).
103
-------
resolved hydrocarbons increased signif-
icantly in the aerobic microcosms, as
the preferred n-alkanes were more rapidly
degraded.
The preliminary available data indi-
cate that in some cases the oxygen-treated
microcosms appeared to exhibit more rapid
rates of degradation than the hydrogen
peroxide treated microcosms. The total
resolved and unresolved hydrocarbons
disappeared more rapidly in the oxygen-
treated microcosms (Table 2). Oxygen-
treated microcosms exhibited decreases in
resolved and unresolved hydrocarbons of
82 percent and 53 percent, respectively.
Peroxide-treated microcosms exhibited
decreases in resolved and unresolved
hydrocarbons of 66 percent and 50 percent,
respectively. A similar trend was repeated
for 1,1-dichloroethane, 1,1-dichloro-
ethylene, and chlorobenzene (Figure 3).
The trends illustrated for 1,1-dichloro-
ethane and chlorobenzene suggest that an
acclimation period of approximately one
month may be required for the microbial
population to adapt to the peroxide
treatment. After an acclimation period,
degradation appears to proceed as rapidly
as in the oxygen-treated microcosms.
Results from the day 100 analyses, which
are not yet complete, are necessary to
confirm these trends.
Anaerobic degradation resulted in the
disappearance of most of the lower mole-
cular weight halogenated hydrocarbons
(Figure 4). 1,1-Dichloroethane, 1,1-
dichloroethylene, and 1,1,1-trichloroethane
disappeared by day 25 in the anaerobic
microcosms. PCE and TCE had disappeared
by day 50. The concentrations of trans-
1,2-dichloroethylene increased, probably
because it is being produced as a degrada-
tion product of PCE and TCE (2).
CONCLUSIONS
Biological degradation is a feasible
treatment alternative and will be effec-
tive at the Kelly site if the metabolism
of the indigenous microorganisms can be
optimally enhanced. The limiting factors
regarding site remediation include the
nature of the biological degradation and
the site hydrogeology.
The nature of the site stratigraphy
and site hydrology will make it difficult
to provide a uniform distribution of the
permeating solution. The clayey soils
have a low permeability. The distribution
of contaminants throughout the site is
fairly heterogenous and there does not
appear to be any defined plume of con-
tamination. The depth of the saturated
zone is variable depending upon climatic
350
• HYDROGEN PEROXIDE
D OXYGEN
150-
100
50
CHLOROBENZENE
10 20 30
TIME (DAYS)
40
50
2 fl-
Figure 3. Change in Concentrations of 1,1-Dichloroethane, Chlorobenzene, and
1,1-Dichloroethylene with Time in Aerobic Microcosms. (Each data
point represents a mean of three replicates.)
104
-------
1600
1400-
1200-
D TETRACHLOROETHYLENE (PCE)
• TRICHLOROETHYLENE (TCE)
O 1,1,1-TRICHLOROETHANE
A 1,1-DICHLOROETHANE
• TRANS-1.2-DICHLOROETHYLENE
A 1,1-DICHLOROETHYLENE
20 30
TIME (DAYS)
40
50
Figure 4. Change in Concentrations of
Lower Molecular Weight Haloge-
nated Hydrocarbons with Time
in Anaerobic Microcosms.
(Each data point represents a
mean of three replicates.)
conditions, and there is a possible
hydraulic connection of the groundwater
with surrounding surface waters.
Aerobic biodegradation can be ex-
pected to reduce the hydrocarbon, aromatic,
and halogenated aromatic concentrations.
The hydrocarbons and unsubstituted aroma-
tics will disappear most rapidly. PCE
and TCE are not expected to degrade
aerobically. For reducing the concentra-
tion of these lower molecular weight
halogenated solvents, aerobic biological
degradation will have to be coordinated
with other treatments. It is unlikely
that the metal concentrations at the site
will be reduced without additional treat-
ment. It is possible that the metals
present will be mobilized, depending upon
the treatment approach. This possibility
will be investigated in further laboratory
studies.
Metals are bound to the soil matrix
in the form of organo-metal chelates and
as sulfides. Under oxidizing conditions,
particularly in the presence of a very
strong oxidant such as peroxide, the
reduced complexes may be oxidized. If
the metals are not re-precipitated as
ferric or manganic complexes, they may be
mobilized into the groundwater. This
would be particularly problematic if
there is a leakage of groundwater into
Leon Creek. (Leakage of peroxide/nutrient
treated groundwater alone would not be
problematic. The peroxide would rapidly
decompose to oxygen and water, and the
nutrients, at worst, could lead to an
algal bloom.)
Aerobic degradation will be used at
the site for two reasons. First, based
on our laboratory studies anaerobic
degradation was not as effective in
degrading as wide a variety of contami-
nants, and second, aerobic degradation is
much more easily accomplished in the
field with a transfer of technologies
from such incidents as gasoline spills.
Oxygen is apparently most easily and
efficiently delivered to the subsurface
in the form of hydrogen peroxide. Matters
which must be resolved before hydrogen
peroxide is used in the field are the
possible mobilization of heavy metals,
and possible precipitation of iron and
manganese oxides and hydroxides and
calcium and iron phosphates that could
clog the injection wells.
Plans for the field demonstration will
be formalized in early 1985. Currently
we plan to design an injection well system
near the center of the area to be treated,
and an extraction well system surrounding
this area. The treatment solution (hydro-
gen peroxide, nutrients and water) will
be prepared in mixing tanks and added to
the injection wells. Groundwater will be
extracted at the site perimeter and
recirculated to the mixing tanks. Moni-
toring will be conducted at the well
head, at the mixing tanks and at other
site locations to maintain an optimum
treatment solution. Other monitoring
points may include points on site to
indicate progress and off site to indicate
any problems with migration of pollutants
or treatment solution. Our treatment
goals at the site are to degrade (within
the time constraints of our project)
105
-------
those contaminants shown to degrade in
the laboratory. These contaminants
include hydrocarbons, aromatics and
halogenated aromatics. Following aerobic
degradation, other treatment techniques,
such as activated carbon filtration and/
or air stripping may be needed to further
address site problems.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the
assistance of William Ellis, Mark Floyd,
Jim Payne, and Kathi Wagner of JRB Associ-
ates; Paul Rogoshewski; C.H. Ward and
Mike Lee of Rice University; Stephen
Klaine and Lawrence M. Mallory of Memphis
State University; and FMC Aquifer Remedia-
tion Systems.
REFERENCES
1. Ghiorse, W.C. and D.L. Balkwill, 1983.
Enumeration and morphological charac-
terization of bacteria indigenous to
subsurface environments. In C.H.
Nash, and L.A. Underkofler, eds.,
Developments in Industrial Micro-
biology. Vol 24. John D. Lucas
Printing Co., Baltimore, Maryland,
pp. 213-224.
2. Parsons, Frances, Gladys Lage, Ramona
Rice, Melissa Astraskis, and Raja
Nassar, 1982. Behavior and Fate of
Hazardous Organic Chemicals in
Contami nated GroundwateTlFor~the
Florida Department of Environmental
Regulation, Tallahassee, Florida.
3. Raymond, R.L., 1983. Personal Communi-
cation. Groundwater/Envi ronmental
Consultants, Inc, Wilmington, Delaware.
4. Raymond, R.L., V.W. Jamison, J.D.
Hudson, R.E. Mitchell, and V.E.
Farmer, 1978. Final Report: Field
Application of Subsurface Biodegra-
dation of Gasoline in a Sand Formation.
For the American Petroleum Institute,
Washington, D.C.
5. Texas Research Institute, Inc. 1982.
Final Report: Enhancing the Microbial
Degradation of Underground Gasoline
by Increasing Available Oxygen. For
the American Petroleum Institute,
Washington, D.C.
106
-------
CASE STUDY INVESTIGATIONS OF REMEDIAL RESPONSE PROGRAMS
AT UNCONTROLLED HAZARDOUS WASTE SITES
Claudia Furman
JRB Associates
McLean, Virginia 22102
S. Robert Cochran
Earth Technology
Alexandria, Virginia 22314
ABSTRACT
Section 105(7) of the Comprehensive Environmental Response, Compensation and
Liability Act of 1980 (CERCLA) requires that USEPA establish means of ensuring that
remedial actions are cost-effective over the period of potential exposure to hazardous
substances or contaminated materials. In partial support of this objective, USEPA has
been directing the preparation of detailed case studies on remedial responses at uncon-
trolled hazardous waste sites. Detailed information has been collected on cost, insti-
tutional, and technical issues related to the implementation and effectiveness of
remedial actions at these sites. Compiled into a succinct case study, this historical
information can then be used in planning, selecting, designing, and cost analyzing
future remedial actions.
The project discussed here is a continuation of earlier case study activities
where 23 case studies were published (2). The focus of this more recent series of 8
case studies, however, was on fund-financed or enforcement actions, as opposed to the
earlier studies which focused on private- and State-lead actions. Within this report,
the 8 case studies are briefly described in table format, followed by a more comprehen-
sive discussion of 4 of the 8 remedial case studies.
INTRODUCTION
As the Superfund program expands and
the number of remedial responses increases,
the nature and type of remedial action
technologies implemented will continue to
grow. It is also apparent that valuable
lessons can be learned from the field
implementation of remedial technologies
at hazardous waste sites throughout the
course of the Superfund program. As a
result, it has become necessary, for
programmatic and scientific reasons, to
provide a systematic way of recording the
results of remedial actions.
PURPOSE
A series of case studies of remedial
actions completed at fund-financed and
enforcement sites has been conducted. At
selected sites, detailed case studies
were performed to document the performance
of applied remediation technologies, and
to determine the limitations and appli-
cability of these technologies in other
remediation activities.
The results of the case study invest-
igation are currently being used as an
information transfer between those who
have experience in using specific remedial
actions, and those planning, assessing,
or implementing site remediation efforts.
The case studies have been prepared for
an audience consisting of members of
industry and commerce, state agencies,
local authorities, and the USEPA.
107
-------
Case study reports have been struc-
tured so that they provide detailed data
on the remediation techniques employed;
the circumstances and conditions under
which they were implemented; their appar-
ent effectiveness in correcting or con-
trolling the problem; and their potential
uses in other remedial action scenarios.
The following is an outline of the typical
case study report:
I. Introduction
• Background
• Synopsis of Site Response
II. Site Description
• Surface Characteristics
• Hydrogeology
III. Waste Disposal History
IV. Description of Contamination
V. Planning the Site Response
• Initiation of Site Response
• Selection of Response
Technology
• Extent of Site Response
VI. Design and Execution of Site
Response
VII. Cost and Funding
• Source of Funding
• Selection of Contractors
• Project Cost
VIII. Performance Evaluation
IX. Bibliography
APPROACH
Initially a survey was conducted of
fund-financed sites and enforcement sites
to identify candidate case study sites.
The survey identified over 30 sites where
remedial responses had been partially or
entirely completed. The project team sel-
ected 8 sites for case study reports.
The criteria used for selection included:
• Current status with regard to
percent completion of the remedial
action
• Availability, accessibility, and
completeness of remedial action,
cost and engineering data
• Type of remedial action technology
implemented, so that a range of
remedial action techniques were
investigated
• Type of waste management practice,
so that a wide range of technolo-
gies, commonly used in hazardous
waste management, were studied
• Types of waste and contaminants
present at the facility to ensure
that a variety of waste streams
and pollutants were included
• Hydrogeologic setting, so that a
variety of settings were repre-
sented
• Enforcement sensitivity.
Once the case study sites had been
identified, an in-depth and site-specific
data acquisition effort was begun. All
available background information relative
to each site's environmental setting,
operational history, remedial actions and
Superfund involvement was collected. Two
readily available sources of information
were used to complete this task. The
first source was site-specific information
collected from public information data
sources, including the following:
• USGS topographic and geologic
maps, geologic and hydro!ogic
reports, soils and climatological
maps and reports
• USEPA and state files, including
site-specific data on operational
history, Superfund involvement,
detailed remedial action, engi-
neering design and cost data,
previously documented cost stud-
ies, and litigative case support
data.
The second source of site-specific
data was direct correspondence with the
USEPA Superfund on-scene coordinators,
the facility operators, and the parties
primarily responsible for designing and
implementing the site remedial action.
During correspondence, additional site-
specific information was collected, and
108
-------
written or verbal approval for site
visits was obtained. Prior to any field
visits, collected information was thorough-
ly reviewed and data gaps identified for
clarification during the site visits.
A schedule for all site visits was
then developed. Follow-up telephone calls
were made to appropriate contacts to coor-
dinate field visit activities and schedules.
Each site visit was two days in duration,
and included the following activities:
• Detailed site inspection, includ-
ing documentation of site layout,
evidence of environmental contam-
ination, and observation of
remedial action technologies
being employed
• Review of all available remedial
action design drawings and exist-
ing cost information
• Detailed review of the remedial
action data previously collected,
including pre-remediation and
monitoring activities, remedial
action and treatment technologies,
implementation and cost details,
and remedial action design data.
All sources of information and site-
specific data were carefully referenced
in field logs for later documentation in
the case study reports prepared for each
site. Photographic slides or prints de-
veloped from the site visits were collected
and developed to track the response
history.
After every site visit, individual
case study reports were prepared and for-
warded to the USEPA Project Officer, Task
Managers, and parties involved in imple-
menting the remedial action for their
review.
RESULTS
The 8 final case study reports de-
scribe an array of remedial technologies
for implementation at uncontrolled hazard-
ous waste sites. In addition to technical
discussions, the case studies present
available cost and funding data for each
site, and describe the more important
institutional issues that affected the
implementation of the response programs.
Table 1 summarizes the site conditions at
the 8 sites prior to implementation of
remedial response actions, the remedial
technologies utilized, and the total
estimated costs associated with the 8
response programs. The remainder of this
section presents a more comprehensive
discussion of the case study investiga-
tions conducted for the 4 sites at which
remedial actions have been completed.
Gulf Coast Lead - Tampa, Florida
Gulf Coast Lead (GCL) is a small,
secondary lead smelting facility whose
operation consists of recovering lead
from discharged lead acid batteries and
lead groupings from battery manufacturers.
Until 1979, rinsewater contaminated with
sulfuric acid and heavy metals, and
battery casings from plant operations,
were deposited in an unlined surface
depression near the center of company
property, causing soil and shallow ground-
water contamination. Surface runoff from
the battery casing area threatened to
spread the contamination to off-site
areas, while the underlying Floridan
Aquifer, Tampa's principal source of drink-
ing water, was also threatened. A federal
suit filed in October 1980 under RCRA
resulted in a consent decree signed later
that same month. Under this agreement
between EPA and GCL, the remedial response
at the site was completed.
The pollution problem at the Gulf
Coast Lead facility was discovered in
December 1978 by the Hillsborough County
Environmental Protection Commission
during an effort to bring all secondary
lead smelting and battery recycling
operations in the county into compliance
with State and local regulations. On
November 1, 1979, after receiving citizen
complaints regarding GCL's noncompliance
with State water quality regulations,
USEPA's Surveillance and Analysis Division
conducted a groundwater investigation.
As a result of this investigation, ground-
water from two on-site monitoring wells
was found to contain high inorganic
contaminant concentrations, metals above
the Maximum Concentration Levels estab-
lished under the Safe Drinking Water Act
(SDWA), and lowered pH. Dead and dis-
colored vegetation was found onsite; and
significant amounts of lead, nickel,
antimony and sulfuric acid were found in
109
-------
TABLE 1. SUMMARY OF CASE STUDY INVESTIGATIONS
SITE SITE CONDITIONS
SI1
b
RESPONSE
STATUS OF
REMEDIAL
PROGRAM
ESTIMATED
COSTS
Bruin Lagoon,
Bruin Borough,
Pennsylvania
Denney Farm,
Aurora,
Missouri
Earthen-diked lagoon constructed
in 1930's and used as a repository
for oil production wastes, waste
motor oil re-refining residues, coal
fines and fly ash; began to draw
public attention following a 3,000-
gallon sludge spill into Bear
Creek which lies adjacent to site.
Estimated 80-90 drums of wastes
containing dioxins discarded
in a trench on property known as
Denney Farm.
Lagoon embankment improve-
ments
Construction of concrete
retaining wall along Bear
Creek
Tank demolition, removal and
offsite disposal
Removal and off-site disposal
of liquid supernatant contained
in lagoon
Sludge stabilization
Installation of venting system
for acidic gas
$1.76 mil lion
• Excavation of waste materials
and contaminated soils
• Placement of soils with
visible contamination in drums
• Placement of other contaminated
soils in microbiological degra-
dation basins
• On-site storage of drummed
wastes (solid waste and
contaminated soils)
• Trench backfilling and capping
• On-site incineration for ultimate
destruction and disposal of wastes
currently under investigation
I
N/A
I - Incomplete
C - Complete
N/A - Not Available
(continued)
-------
TABLE 1. (continued)
SITE
SI
IE
CONDI
TIONS
SITE
RESPONSE
S
TATUS OF
REMEDIAL
PROGRAM
ESI
IMATED
COSTS
Gulf Coast
Lead,
Tampa, Florida
Lipari
Landfill,
Pitman, New
Jersey
Contaminated rinsewater and battery
casings from lead recovery operations
at the GCL smelting facility were
dumped in unlined surface depression
on company property. Soil and
shallow groundwater became con-
taminated with sulfuric acid and
heavy metals, thus presenting a
threat to an underlying aquifer.
Seven-acre area used for disposal
of industrial and domestic wastes
including solvents, formaldehyde,
paints, phenol and anine wastes;
concern focused on contamination
of two nearby streams, a lake and
two aquifers.
• Implementation of groundwater
monitoring program
• Removed battery casings from
surface depression
• Capping of site
t Installation of retention basin
to collect and recycle contami-
nated surface runoff from site
t Installation of underdrain sys-
tem to collect rainwater runoff
from uncontaminated areas for
discharge to Tampa storm drains
t Installed a groundwater treat-
ment system
• Subsurface acid reaction barrier
to intercept and neutralize con-
taminated groundwater that had
migrated off-site was under study
at the time of this writing
• Construction of a soil-
bentonite slurry trench cutoff
wall around a 16-acre area which
included the 7-acre landfill
• Installation of synthetic mem-
brane cap over the 16-acre site
• Installation of passive gas
collection and venting system
• Completion of groundwater treat-
ability study and design of a
collection and treatment system
$700,000
$2.5 million
I - Incomplete
C - Complete
N/A - Not Available
-------
TABLE 1. (continued)
SITE
SITE CONDITIONS
SITE RESPONSE
STATUS OF
REMEDIAL
PROGRAM
ESTIMATED COST
Picillo Farm,
Coventry,
Rhode
Island
Rose Park
Sludge Pit,
Salt Lake
City, Utah
A 7.5-acre area used as a hazardous
waste disposal site during the
1970's; over 10,000 drums containing
industrial solvents and oils, pesti
cides, PCBs, paint wastes and
explosives were buried in 4
trenches on site; leachate
migrated through soil into nearby
swamp, contaminating groundwater,
surface water and soil in vicinity.
Six-acre disposal site located in a
public park; between 1920 and 1957
acid sludges were discarded in unlined
pits and covered with lime and soil;
the site was covered with a soil cap
in 1960; in 1976 construction
activities at the site forced
sludge to the surface; low concen-
tions of contaminants were found
in groundwater immediate area but
had not migrated appreciably.
Drum removal and off-site
disposal
Ground and surface water
monitoring of swamp and
contaminated groundwater
plume
Detonation of unidentified,
explosive materials
Landfarming of contaminated
soils
Completion of a feasibility
study on groundwater recovery,
treatment and discharge
Installation of groundwater
monitoring wells around sludge
pit
Construction of a soil-
bentonite slurry trench
cutoff wall around the site
Construction of a sand layer,
filter fabric and clay cap
system
Installation of fence around
perimeter of site
$4.96 million
$1.2 million
I - Incomplete
C - Complete
N/A - Not available
(continued)
-------
TABLE 1. (continued)
SITE
SI
I'E
CONDI
TIONS
SITE
RESPONSE
STATUS OF
REMEDIAL
PROGRAM
ESTIMATED
COST
Tacoma Well Well 12A, the northern most of 13 •
12A, Tacoma, wells in a well field in South
Washington Tacoma Channel, was taken off line
after chlorinated organic solvents
were detected in a groundwater sample;
the source was identified as being •
north of the well field; the
spreading of contaminants to other
wells after 12A was taken out of
service indicated that further con-
tamination of the well field could
be prevented if Well 12A remained
operational to act as an intercepter
or barrier.
Taylor Road A 42.5-acre sanitary landfill in •
Landfill, which unknown quantities of •
Seffner, hazardous wastes including solvents,
Florida paint thinners, insecticides,
herbicides, fungicides and sludges t
were deposited; groundwater samples
collected at the site were found to
contain VOCs and metals in concen- •
trations above acceptable safe
drinking water standards.
Installation of an air-
stripping system at Well 12A
to remove volatile organics
(considered an interim
remedial measure)
Completion of studies to
identify contaminant source,
and design and implement a
permanent solution
Upgrading of surface cap
Channelization of surface
drainage along eastern and
southern site boundaries
Installation of a methane
gas collection and control
system
Development of long-term
groundwater and methane gas
monitoring program
$948,133
T-Incomplete
C - Complete
N/A - Not Available
$3.56 mil lion
-------
the shallow groundwater. A later investi-
gation confirmed the source of the problem
as acid infiltration from untreated or
partially treated wastewater discharged
into the shallow aquifer, and metals
being leached from the surface into the
groundwater.
In order to immediately reduce the
threat to the local private water supply
caused by the discharge of untreated or
partially treated wastewaters, GCL was
ordered to upgrade its treatment system
to meet the influent requirements of the
Tampa Publicly Owned Treatment Works
(POTW). Hence, discharge to the unlined
surface depression was halted. The
consent decree signed between EPA and GCL
in October 1980, then required that GCL
either eliminate, or isolate and contain,
all hazardous constitutents in the ground-
water attributable to GCL operations.
The selected remedial plan at the
GCL site consisted of capping, combined
with groundwater withdrawal and treatment.
This plan provided the greatest assurance
that further contamination in the area
would be eliminated. This was at a later
date augmented, through negotiations with
the EPA, by the addition of an "acid
reaction barrier" to intercept and neutra-
lize contaminated groundwater that had
already migrated offsite.
The remedial plan implemented at GCL
consisted of the following remedial
technologies:
• An earthen cap over part of the
unpaved area of the site and an
asphalt cap over the remainder
• A surface water control
system
• A groundwater withdrawal and
treatment system
• A long-term groundwater
monitoring program
• A subsurface acid reaction
barrier.
The installation of these technologies
is briefly discussed below.
Capping -
The objective of this remedial
technology was to prevent further infiltra-
tion of rainwater. Therefore, all unpaved
areas at the site were capped with either
asphalt or clay. The clay cap was placed
over a 3.4-acre area of the site overlying
the buried battery casings. This cap
consisted of a 4-inch layer of compacted
clay over the existing grade, followed by
8 feet of topsoil and grass sod.
Surface Water Controls --
Surface water controls were imple-
mented to collect the increased runoff
from the cap surface. The surface water
control system consisted of two main
components: a stormwater detention pond
and a stormwater detention basin. The
stormwater detention pond is lined with
.035-inches of reinforced chlorinated
polyethylene. It receives all rainwater
falling on the paved areas of the GCL
plant. In addition, the pond receives
runoff from sprinkling systems located on
all buildings which are used to reduce
the levels of lead in the air. The water
collected in the pond (called graywater)
is recycled for use in the sprinklers and
in the battery washing system. The storm-
water detention basin is designed to
collect surface runoff from the clay cap
for conveyance to the City stormwater
ditch which parallels the northern boundary
of the GCL site.
Groundwater Withdrawal and Treatment --
The objectives of the groundwater
withdrawal and treatment system are: (1)
to contain (by reversal of the groundwater
gradient) the off-site migration of
contaminants, and (2) to eliminate the
sources of the contamination. Pumping
was, therefore, implemented in two areas.
One area was directly above the battery
casing disposal location and the other
area was along the western and south-
western boundaries of the GCL site. The
battery casing area was selected because
it was the source of groundwater contami-
nation, and pumping would result in the
quickest improvements in groundwater
quality. The western and southwestern
boundaries of the site were selected for
groundwater withdrawal because pumping
these areas would intercept contaminated
groundwater flowing away from the site,
and also would draw contaminated ground-
water offsite and into the treatment
system.
114
-------
Withdrawn groundwater is treated in
GCL's wastewater treatment system, which
consists of primary clarification, pH
adjustment, and final clarification, and
then discharged to the Tampa POTW.
Groundwater Monitoring --
The groundwater monitoring program
at the GCL site was designed to detect
any unsuspected migration of contaminants
within the upper aquifer, and to detect
any downward movement of contaminants
into the Floridan Aquifer. These objec-
tives are being met through quarterly
sampling of 6 upgradient and 21 downgra-
dient wells at the site.
Subsurface Acid Reaction Barrier --
The purpose of the subsurface barrier
is to intercept and neutralize any contam-
inated groundwater that had already
migrated offsite. At the time of this
writing, studies were being conducted to
determine the proper depth and placement
of this barrier which will consist of a
deep trench filled with oyster shells and
.crushed limestone.
Cleanup Objectives and Effectiveness --
The main objective of the response
effort was to successfully isolate and -
contain hazardous contaminants attributable
to company operations that were found in
the soil and shallow groundwater beneath
the site. In this way, the threat of
contaminating the Floridan Aquifer, and
thus, area drinking water supplies, would
be eliminated. Capping the site, which
was one element of the stormwater manage-
ment strategy, served to minimize infil-
tration of rainwater and thus prevent
displacement of the contaminated fluids.
It also served to isolate the source of
contamination and partially mitigate
threats to public health and the environ-
ment. Pumping and treating the contami-
nated groundwater directly also served
this purpose. In total, the remedial
actions are expected to correct the
contamination problem at the site. The
remedial plan will be considered success-
ful when all wells onsite achieve accept-
able water quality levels based on SDWA
Maximum Concentration Levels.
The effectiveness of the remedial
response technologies at the GCL site
cannot be evaluated at this writing.
However, the selection of these response
technologies appears to have considered
all viable alternatives, and the operation
of the response equipment has proceeded
as expected. Unless major problems arise,
such as clogging of the extraction wells,
improvements in the extracted groundwater
should be apparent within a 5-year period.
Rose Park Sludge Pit - Salt Lake City,
Utah
The Rose Park Sludge Pit is a former
waste disposal site located on 5 to 6
acres of a 33-acre public park in an
industrial/residential area of northwest
Salt Lake City, Utah. From the 1920's
until the mid 1950's, an undocumented
amount of acid waste sludge from petroleum
refinery operations at the Utah Oil and
Refining Company was discarded into
unlined pits onsite and then covered with
lime and soil. The waste acid sludges
were generated from the acid treatment of
light petroleum oil products at the
refining company. Remedial work at Rose
Park was funded as a private party re-
sponse.
In 1957 Salt Lake City purchased the
site in response to residents' complaints
about the dumping. An estimated 40 to 100
truckloads of sludge were excavated and
transported to a local dump, after which
the area was covered with soil and left
undisturbed for almost 20 years.
During a 1976 expansion of park
facilities in the area, plans to build a
Little League ball field and tennis courts
were halted when a bulldozer broke through
the overburden cap and started the effu-
sion of sludge from the site. Workers
began to complain of respiratory and eye
irritations and headaches due to sludge
fumes, prompting the Salt Lake City/County
Health Department to analyze samples of
the sludge. In late 1977, the Utah
Geological and Mineral Survey (UGMS)
began a preliminary investigation of
groundwater contamination at the site,
and in July 1979 USEPA Region VIII
visited the site.
Analysis of the sludge by the County
Health Department revealed that the sludge
was acidic and that the mist released was
sulfuric acid. The UGMS study concluded
115
-------
that the site was possibly contributing
oil and grease contamination to the
groundwater downgradient from the site.
During the EPA preliminary assessment,
chemical analyses of the sludge samples
indicated the presence of various pollu-
tants including xylene, o-dichlorobenzene,
pyrene, fluoranthrene and phenanthrene.
The primary immediate hazard at the
Rose Park site was the risk of exposure
through direct contact with the waste
materials by humans. Due to the viscous
nature of the waste material, contaminant
migration potential was very low, and the
direct exposure route was a primary
concern in developing the remedial action.
Short-term measures taken to reduce the
immediate hazard of direct contact were
the installation of a 6-foot high chain-
link fence with a barbed wire top rail,
extension of the property boundary by
approximately fifty feet and new signs
warning of the hazard. A detailed montor-
ing program to improve characterization
of the problem was then implemented.
The remedial response actions agreed
upon in an Intergovernmental/Corporate
Cooperative Agreement between the State
and the responsible private party, provided
for monitoring, capping, slurry wall
construction, and security measures. The
implementation of these actions are
briefly described below.
Groundwater Monitoring --
Sampling by EPA of all 10 monitoring
wells installed at the site will occur
quarterly in the first year of remedial
action operation. After year one and
through year 30, sampling of only six of
the wells becomes the responsibility of
the City and the private party. The
sampling program is intended to ensure
that the remedial action is operating
properly over its effective life.
Slurry Wall --
Final specifications called for
construction of a soil-bentonite slurry
wall which extended to a depth of about
30 feet into the underlying low permea-
bility silty clay, and around the 2,326-
foot perimeter of the sludge pit. Speci-
fications called for the use of Wyoming-
type bentonite which conformed to American
Petroleum Institute specification 13A -
Oil Well Drilling Fluid Material (3).
The main trench was excavated to a
depth of about 30 feet. This allowed the
wall to penetrate at least 10 to 15 feet
into the underlying low permeability clay
soil, 15 feet below the lowest observed
oily residue zone and 10 feet below the
deepest known extent of the sludge pit.
The completed trench was a minimum of 3
feet and a maximum of 5 feet thick.
Capping --
The second phase of the remedial
action involved placement of a cover over
the sludge pit. The components of the
cover construction included the following:
cleaning and site grading; placement of
bank run fill, clean sand fill, filter
fabric, clay cover, and topsoil; and re-
vegetation of the disposal area.
Site grading was necessary to achieve
a 2 percent minimum grade. Following the
site grading activity, a sand layer and
gas vents, designed to provide a pathway
through which gases could be atmospheri-
cally vented, were installed. Filter
fabric was then placed loosely over the
sand fill to provide stability and to
prevent upward migration of contaminants.
A 2-foot thick clay cap of a low permea-
bility silt was then placed over the
site, followed by 8 inches of topsoil.
Revegetation was postponed until weather
conditions were favorable during the
growing season, and has since been com-
pleted.
Site Security --
Prior to vacating the site in July
1984, the chain link fence to the north
of the site, which had been removed at
the start of the trenching operation, was
reinstalled.
A vehicular barrier was erected in
October 1984. The barrier is a corrugated
steel rail mounted on steel posts and
stands about 3 feet high. It runs along
the west, south and east sides butting
into the chain link fences on the north.
Warning signs have been placed at each
entrance prohibiting digging or vehicular
traffic.
116
-------
Cleanup Objectives and Effectiveness --
The primary objective of the remedial
action at the Rose Park site is contain-
ment of the waste material. The purpose
of containment is to remove the threat of
direct physical contact with the sludge
and its health-threatening components.
The sludge reportedly has low potential for
migration and does not pose an offsite
hazard to health and the environment.
However, the selected remedial action
considered the flow of groundwater through
the site and migration of the contaminants,
in addition to removing the threat of
direct physical contact. Monitoring of
the site to insure the integrity of the
containment system was accomplished with
wells provided by a private party.
Available information indicates that
the Rose Park remedial actions were
carried out according to plan and have
been successful. The containment system
greatly reduces the potential for direct
contact, as well as providing the added
benefit of eliminating odors. Further,
the containment system has greatly reduced
the possibility of contaminant migration,
and further decreased the possibility of
groundwater contamination.
Tacoma Well 12A - Tacoma, Washington
The Tacoma Well 12A site is located
in Tacoma, Washington, a historically in-
dustrial area. During the summer of
1981, chlorinated organic solvents were
detected in the water from City Well 12A,
one of 13 in a well field used to meet
increased demand for potable water during
summer months. The source of contamination
at the well field has not yet been con-
clusively identified, but a 1983 study
revealed that it may be waste sludges
from an oil reclamation company that was
located in the area but has since gone
bankrupt and is no longer in operation.
Enforcement actions began with sufficient
information to link a specific party to
the pollution, USEPA and the State entered
into a Superfund Cooperative Agreement
for remedial work in June 1982.
The sequence of events leading to
the discovery of the contamination problem
at Well 12A began with routine sampling
and testing of Tacoma well water during
the National Organic Monitoring Survey in
March 1976. Traces of trichloroethylene
were found, prompting the Washington
Department of Ecology (DOE) and the
Tacoma/Pierce County Health Department to
initiate an investigation of possible
source(s) of the contamination. When
over 600 drums of stored chemicals were
found at the American Surplus Company
(buyers and sellers of chemical solvents
and paint thinner), only two blocks from
Well 12A, EPA and the Washington DOE
decided to sample Well 12A.
Results of this sampling indicated
the presence of such chlorinated organic
solvents as 1,1,2,2-tetrachloroethane,
trichloroethylene, 1,2-transdichloroethy-
lene and tetrachloroethylene. The Wash-
ington Department of Health and Human
Services concluded that the use of Well
12A for drinking purposes would constitute
a definite health threat, and subsequently
removed the well from service. However,
the same pollutants that were found in
Well 12A began to show up in other City
wells and monitoring wells constructed
during the remedial investigation.
The need for potable water in the
area could not be met by simple conserva-
tion methods or other alternatives, so
abandoning the well field was not con-
sidered a viable option. Consequently,
an Initial Remedial Measure was developed
to contain the contaminated plume and to
provide safe drinking water, until a more
permanent solution could be found following
completion of additional source identifi-
cation activities. Pumping and treatment
of the water from Well 12A was the selected
response to the health hazard posed by
the contaminants. Pumping Well 12A at an
appropriate rate prevents contaminant
migration to other wells in the field by
altering the hydraulic gradient. In
addition, water withdrawn from Well 12A
is treated and used for water supply pur-
poses during summer months.
Of the treatment alternatives con-
sidered, air stripping and carbon adsorp-
tion treatment were considered viable
responses. Data from bench scale and
pilot plant studies of these technologies
indicated the air stripping was more
economical and efficient at the scale
required for Well 12A. The air stripping
system installed at the Well 12A site is
briefly discussed below.
117
-------
Air Stripping Treatment System --
The pumping/air stripping system was
designed as a temporary measure to provide
necessary drinking water and control
further pollutant migration. A design
life of five pumping seasons was selected,
although actual operation of the system
will be dependent on the outcome of
further studies and the implementation of
a permanent cleanup plan.
A pumping rate for Well 12A of 3500
gallons per minute (gpm) results in the
desired hydraulic gradient, which acts as
a barrier to further migration of con-
taminated groundwater to other wells in
the field (which, without pumping of Well
12A, would be downgradient of Well 12A).
Five air stripping towers, with design
capacity of 700 gpm each, are used in
parallel to treat the total flow of 3500
gpm. The air flow rate necessary for
stripping is 29,000 standard cubic feet
per minute per tower. An air permit was
obtained from the Puget Sound Air Pollu-
tion Control Agency for the emission of
volatile organic compounds stripped from
the withdrawn groundwater.
Cleanup Objectives and Effectiveness --
The objectives of the treatment system
enacted at Tacoma Well 12A were:
• To block the spread of contami-
nants
• To supply drinking water within
applicable water quality standards
• To partially clean up the aquifer.
To block the spread of contaminants,
the treatment system for Well 12A had to
operate over extended periods of time
either discharging to Commencement Bay or
to the Hood Street Reservoir when the
well field was needed as water supply in
the summer months. In order to operate
over a period of time and to meet appli-
cable water quality standards, discharge
limits were placed on treated effluent as
follows:
t A maximum of 15.5 pounds/year of
1,1,2,2-tetrachloroethane could
be discharged to the Hood Street
Reservoi r.
• The average concentration of
effluent to Commencement Bay of
1,1,2,2tetrachloroethane had to
be 50 ppb or less.
• The maximum concentration allowed
for discharge to Commencement
Bay was 100 ppb.
Therefore, the treatment system had
to operate so that less than the maximum
quantity of 1,1,2,2-tetrachloroethane
would enter Hood Street Reservoir during
the period when the well field was operat-
ing, and so that effluent limits would be
within acceptable ranges when treated
effluent was discharged to the bay. If
the 15.5-lb limit was reached, water
could no longer be discharg ed to the
drinking water system, and if the concen-
tration limits were exceeded, the well
would be forced to shut down.
The efficiency of the Well 12A treat-
ment system was approximately 98 percent
during the period from August 4, 1983 to
November 4, 1983. Calculations made in
October 1983 showed that effluent from
the well could be sent to the Hood Street
Reservoir for 305 days before the 15.5-lb
limit was reached. In addition, the
effluent discharged never rose above the
50-ppb limit set as the average for
discharge to Commencement Bay. This
record of performance indicated that the
system could be operated for periods much
longer than required to protect other wells
and still be within the discharge limits.
Because the effluent concentrations were
low, treated water could be discharged to
the Hood Street Reservoir and into the
City's distribution system for a period
much greater than required to meet drink-
ing water supply needs. Therefore, the
system accomplished two of the three
stated objectives:
• The system permitted the well to
operate for periods while the
well field was in operation,
thus blocking the spread of
contaminants.
• The system treated Well 12A
groundwater such that effluent
could be discharged to the Hood
Street Reservoir if needed.
The system appears to be meeting its
third objective—partially cleaning up
118
-------
the aquifer—but sampling data, at the
time of this writing, was only available
from a single pumping season.
Taylor Road Landfill - Seffner, Florida
Taylor Road is a 42.5-acre landfill,
which had been owned and operated by
Hillsborough County near Seffner, Florida,
approximately 10 miles east of Tampa.
Although originally intended only for the
disposal of municipal refuse, the landfill
had unknown quantities of solid and
hazardous wastes including solvents,
paint thinner, halogens, insecticides,
herbicides, oil, sludges, fungicides and
other wastes deposited there between 1976
and 1979. Improper site management and
disposal operations contributed to the
areal groundwater, soil and air contamina-
tion which was attributed to Taylor Road
and one of two other Countyowned waste
repositories adjacent to the site. A
suit was filed under RCRA and SDWA in
October 1980 and later amended to include
CERCLA. The case was settled in June
1983 when Hillsborough County, the USEPA
and the Florida Department of Environmental
Regulation (DER) signed a consent decree
for a $2 million remedial program.
In early 1979, the Taylor Road Land-
fill site came to the attention of USEPA
Region IV personnel when area residents
expressed their concern about a proposed
expansion of the existing landfill by the
County. The site is situated near several
private residences and within a 1.5-mile
radius of various commercial, industrial
and institutional establishments. In
early August 1979, EPA scheduled a site
inspection in response to the numerous
citizen complaints and as part of the
nationwide, Congressionally mandated
inventory of landfills.
Analysis completed in December 1979
by EPA's Water Supply Branch revealed that
soil and monitoring wells were heavily
contaminated with organic compounds and
heavy metals, and that potable wells also
contained traces of various volatile
organic compounds (VOCs). Similar sampl-
ing by County consultants supported EPA's
findings. A March 1980 preliminary
assessment by EPA, documents continued
contamination of the water supply, notice-
able odors, and evidence of unknown
amounts of solvents and hospital wastes
buried at the site. Tests conducted on
August 20, 1980 by a team comprised of
Hillsborough County and DER personnel,
found methane gas offsite at concentra-
tions exceeding the lower explosive limit
even though a preliminary gas venting
system had previously been installed.
Following the determination by EPA
in 1979 that groundwater in the vicinity
of Taylor Road was contaminated, EPA sent
warnings to residents and businesses in
the area that were using private wells
for their water supply, to discontinue
use of all wells, and contacted State and
County officials to recommend that they
provide alternate supplies of water to
these residents and businesses. Hills-
borough County immediately initiated
delivery of bottled water to affected
well users to abate the hazard posed by
use of the contaminated groundwater, and
has plans for extending County water
service to these users.
A remedial investigation was begun
in November 1981. This study was intended
to evaluate the site, and confirm the
presence and extent of contamination,
evaluate existing pollution controls (i.e.,
cap, drainage ditches, and gas control
system) and recommend remedial alterna-
tives for the mitigation of groundwater
contamination originating from the land-
fill.
The selected remedial alternative was
the installation of the impermeable cap
and surface drainage system. The system
was selected because it was technically
feasible, did not jeopardize the existing
integrity of the Taylor Road or adjacent
landfills, and had essentially no human
health or environmental risk associated
with its implementation.
The response action at the Taylor
Road Landfill site was comprised of two
phases: (1) upgrading of the facility
cap, and (2) improving surface drainage
to prevent cap erosion and water infiltra-
tion. Monitoring at the site was also
required to detect any future problems
and to determine the effectiveness of the
response actions. The implementation of
the selected response actions are briefly
described below.
119
-------
Cap Upgrading --
The consent decree mandated that
about 65 percent of the surface area of
the site be upgraded. Permeability tests
were to be performed on an additional 6
acres, and if the permeability was not
acceptable, then these areas would also
have to be upgraded.
Areas requiring upgrading were to be
improved such that the entire area of the
site would be covered by a minimum of 2
feet of clay soil with a permeability
of 1 x 10~6 cm/sec or less. Upgrading was
accomplished by removing topsoil and sub-
soil on the existing cover, mixing subsoil
with a montmorillonite clay, and replacing
and compacting the soil-clay admix to
meet permeability specifications.
Due to capping with impermeable clay
(which prevents venting of gases through
the landfill surface), a gas collection
and venting system was required to prevent
horizontal migration of methane generated
by degradation of organic wastes. The gas
vent system is composed of 42 collection
wells, header pipes, condensation traps,
and a blower and flare station. The header
system connects all wells to the blower
and flare station. Sample points, conden-
sation traps and control valves are
located throughout the system to allow
for maintenance control and inspection of
the system. The County has been initially
flaring the methane gas but eventually
intends to recover and use the methane.
Surface Runoff Control --
Surface runoff control from the cap
was improved to prevent ponding and ero-
sion. Clay-lined drainage channels were
constructed or upgraded around the cap
perimeter. The original clay linings
were found to erode relatively quickly
due to the heavy rainstorms that typically
occur in the area, and the clay linings
were replaced with concrete by the County
to minimize erosion and future maintenance
costs. Additionally, the County installed
concrete flumes along the side slopes to
channel water from the surface down the
sides of the fill with minimum erosion.
Monitoring Systems —
To insure that the initial remedial
actions were effective, a detailed moni-
toring program for groundwater and methane
gas was implemented. Groundwater samples
are collected quarterly from 4 wells, and
semi annually from 3 additional wells.
Additionally, three new wells may be in-
stalled for additional monitoring.
Flexibility in monitoring allows for the
relocation of monitoring points but not
for increase in the number of points. In
turn, this allows for the accurate monitor-
ing of the groundwater if the flow direc-
tion changes.
A methane gas monitoring system was
established around the Taylor Road Land-
fill and two adjacent County-owned waste
repositories, for the purpose of detecting
the migration of methane gas beyond the
boundaries of the fill area. This system
consists of 32 wells extending from
surface to bedrock.
Cleanup Objectives and Effectiveness --
The overall short term objective of
the remedial action program at Taylor Road
was to control the migration of existing
contamination as well as to prevent new
contamination. The response was designed
to minimize surface water infiltration and
the migration of methane gas. The long
term remedial action goal is to ensure
that the system is effective and that
future contamination is prevented, thereby
eliminating the risk of exposure to area
residents from a polluted drinking water
supply.
There are no data currently available
to evaluate the effectiveness of the reme-
dial actions. However, based on the site
conditions and information that is avail-
able, the remedial actions selected and
implemented should be effective in the
long-term. If properly maintained, the
systems should effectively control leachate
from the Taylor Road Landfill by prevent-
ing infiltration, which prevents the
displacement of leachate currently in the
landfill and prevents formation of addi-
tional leachate.
Summary
Together, the 8 case study reports
describe a variety of remedial technolo-
gies that have been implemented at fund-
financed and enforcement action sites.
The initial environmental conditions,
i.e., hydrogeology, type of contamination,
contaminant levels, population affected,
120
-------
etc, that characterized each site were
unique, as were the circumstances and
events leading to each site's discovery
and remediation (see Table 1). The use-
fulness of the information contained in
these reports lies with this characteristic
that each report presents an entirely
unique set of circumstances under which
the regulatory authorities (Federal,
state, and local), and the public sector
had to act in order to design a resolution
to a problem that posed a threat to the
environment and public health. The case
study reports provide detailed accounts
of remedial program design, initiation
and implementation. They also include
brief evaluations of the effectiveness of
completed remedial response programs.
With the type of information contained in
these reports, they can serve as useful
learning tools for those involved in all
phases of remediation programs implemented
at uncontrolled hazardous waste sites.
ACKNOWLEDGEMENTS
JRB Associates and the Environmental
Law Institute (ELI) prepared the eight
case study reports for USEPA's Office of
Research and Development, Hazardous Waste
Engineering Research Laboratory, Land
Pollution Control Division. These reports
were prepared through the coordinated
efforts and guidance of S. Robert Cocnran,
JRB Project Manager, Claudia Furman, JRB
Task Manager and Dr. Edward Yang, ELI
Project Manager. The following persons
from JRB and ELI all played major roles
in conducting the case study site visits
and report preparation: Dirk Bouma,
Richard Eades, Mark Evans, Linda Schwartz,
Edward Tokarski , Tim Van Epp, and Kathleen
Wagner.
The project team wishes to acknow-
ledge and thank Douglas Ammon, Stephen
James, Bruce Clemens, and John Cross, EPA
Task Managers, for their assistance and
support in the development of the reports.
REFERENCES
1. Hazardous Materials Intelligence
Report, 1981. Periodic interim lists
of Superfund priority waste sites,
1981-1984. World Information Systems,
Cambridge, MA.
2. USEPA, Office of Emergency and Remedial
Response, and Office of Research and
Development, Municipal Environmental
Research Laboratory, 1984a. Case
Studies, Remedial Response at Hazard-
ous waste Sites. EPA-540/2-84^00T;
3. USEPA, Office of Emergency and Remedial
Response, and Office of Research and
Development, Municipal Environmental
Research Laboratory, 1984b. Slurry
Trench Construction for Pollution
Migration Control, EPA-540/2-84-001.
4. USEPA, Office of Emergency and Remedial
Response, and Office of Research and
Development, Hazardous Waste Engineer-
ing Research Laboratory, 1984c.
Draft case study reports on the
following eight sites: Bruin Lagoon,
Penney Farm, Gulf Coast Lead, Lipari
Landfill, Picillo Farm, Rose Park,
Tacoma Well 12A and Taylor Road
Landfill (unpublished).
5. USEPA, Office of Solid Waste and
Emergency Response 1983. Hazardous
Waste Sites. National Priorities
List. HSW-7.1.
121
-------
EFFECT OF FREEZING ON THE LEVEL OF CONTAMINANTS
IN UNCONTROLLED HAZARDOUS WASTE SITES - PART I:
LITERATURE REVIEW AND CONCEPTS
I.K. Iskandar
U.S. Army Cold Regions Research and Engineering Laboratory
Hanover, New Hampshire 03755-1290
J.M. Houthoofd
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
A literature search indicated that natural freezing may have detrimental effects at
uncontrolled hazardous waste sites in the cold-dominated areas because of frost action on
buried materials and ion movement in soils. Natural and artificial freezing, however,
can be used beneficially to concentrate effluents, and to dewater sludges, contaminated
sediment and soils. The process of artificial ground freezing can also be used as an al-
ternative to temporarily immobilize contaminant transport and potentially for decontami-
nation of soils, sediments and sludges. A cost and economic analysis procedure was de-
veloped and used to evaluate ground freezing.
INTRODUCTION
Hazardous waste treatment and disposal
is one of the major environmental concerns
currently facing the nation. The amount of
waste generated annually in the United
States is estimated to be 50 million tons
(9) and is expected to increase rapidly.
In addition to the concern for proper
treatment and disposal of hazardous waste
produced now and in the future, there is
even more concern about past disposal prac-
tices. As is widely known, a large number
of uncontrolled hazardous waste sites have
been discovered (9) and the number is in-
creasing. Although these sites have been
abandoned, the damage to the environment
has either already occurred or can be ex-
pected to occur in the future. A survey of
204 disposal sites (15) concluded that
available techniques for remedial action
are not adequate for the complete cleanup
of ground and surface water. Neely et al.
(15) emphasized the need for innovative
techniques for treatment, and effective
methods for the complete cleanup of uncon-
trolled hazardous waste sites.
OBJECTIVES
This project has been conducted in two
phases. The objectives of Phase I are:
1. To review the literature on po-
tential effects of freezing on
the level of contaminants at ex-
isting hazardous waste sites.
2. To evaluate the feasibility of
using natural and artificial
freezing as an alternative tech-
nique for treatment of hazardous
waste.
3. To report on the technical con-
cept for the artificial freezing
process and availability of hard-
ware.
4. To develop a cost and economic
analysis of artificial freezing
for hazardous waste remedial ac-
tion.
122
-------
TABLE 1. EXPERIMENTAL PLAN, ION TRANSPORT IN FREEZING AND FROZEN SOILS
Treatment
Water content
Heavy metal added
Freeze/thaw cycles
Drainage
Drainage period
Leachate chemical analysis
Soil chemical analysis
Photos
Thin section
160%
yes
1
no
-
yes
frozen
yes
yes
160%
yes
1
yes
2 days
yes
thawed
no
no
160%
yes
3
yes
2 days
yes
thawed
no
no
160%
yes
1
yes
extended
yes
thawed
no
no
160%
yes
0
yes
extended
yes
thawed
no
no
The objective of Phase II of the pro-
ject is to conduct laboratory studies on
the effect of freezing on the fate of se-
lected metals and organics in contaminated
soils.
APPROACH
Phase I of this project was conducted
by searching the literature on the use and
effects of artificial and natural freezing
on contaminants in uncontrolled hazardous
waste sites, sea water, wastewater, sludge
and soils. The search indicated that there
is very little information available on the
effects of freezing on the level of contam-
inants at uncontrolled hazardous waste
sites. However, a great deal of research
has been conducted on the use of freezing
for dewatering sludge, for desalting sea
water, and for constructing mines and other
civil engineering projects. The relevant
findings will be summarized in this paper
and in more detail in a separate project
report submitted to the USEPA.
The approach taken for Phase II was to
investigate the fate of selected heavy met-
als and organics in soils upon freezing.
The selected metals were Cu, Zn, Cd and Ni,
and the organics included toluene, benzene,
chloroform and tetrachloroethylene. The
heavy metals experiment was conducted sepa-
rately from the organics experiment to
avoid possible interactions and the need
for different techniques for sample hand-
ling. The effect of freeze/thaw cycles on
the quantity and quality of leachate was
investigated. Table 1 shows the experi-
mental design for the metals portion of the
study. The setup for the organics was sim-
ilar to that of the metals except that soil
column 4 was subjected to two cycles of
freeze/thaw and soil column 3 was subjected
to five cycles of freeze/thaw. The columns
used were made from plexiglass, 12.5 centi-
meters (cm) in diameter and 80 cm in
height, and were filled with either metal-
spiked or organic-spiked soils. Visual ob-
servations were made during the freeze/thaw
cycles and the volumes of leachate were re-
corded with time. Subsamples of leachate
were analyzed for the desired metals or or-
ganics. Metals were analyzed by atomic ab-
sorption spectrophotometer or by inductive-
ly coupled plasma (ICP). The organics were
determined with a gas chromatography-mass
spectrophotometer (HP 5992 GC-MS) equipped
wiht a purge trap sampler (HP 7675A). Deu-
terobenzene (C6D6) in tetraglyme was added
to each sample as an internal standard just
prior to purging. The rate of frost pene-
tration was recorded with time using ther-
mocouples and a data logger. At the con-
clusion of the experiments, photographs and
thin sections of the frozen column were
made to document the effects of freeze/thaw
on soil structure and the formation of ice
lenses. Subsamples of the soils with depth
were chemically analyzed for metals or or-
gani cs.
PROBLEMS ENCOUNTERED
Some difficulties in the extraction
and the analysis of organics were encount-
ered which made it necessary to improve on
current analytical methods. The improved
techniques will be published elsewhere
(Jenkins, in preparation).
RESULTS
The literature search indicated that
natural and artificial freezing has been
used for over 100 years for mining (4), for
desalting of sea water (11,12,18,20), for
123
-------
TABLE 2. NORTH AMERICAN MINE SHAFT GROUND FREEZING (4)
Mine
Chapin Mining Co.
PCA
PCA
Myles Salt Co.
International Min.
Duval
International Min.
Allan Potash Mi nes
Alwinsal Potash
Noranda Mines Ltd.
Cominco Ltd.
PCA
PCS
PCS
Cargill
Cargill
PCS
Amax Coal Co.
Diamond Crystal
Island Creek Co.
Morton Salt Co.
Peabody Coal Co.
Selco Mining Corp.
White County C. Corp.
Domtar
Les Mine Sal in
Turn's Coal Co.
Asarco Expl oration
— • • • - —
Location
Iron Mt. , Michigan
Carlsbad, NM
Sask., Canada
Louisiana
Canada
Carlsbad, N.M.
Canada
Canada
Canada
Canada
Canada
Sask., Canada
Canada
Canada
Louisiana
Louisiana
Canada
11 linois
Louisiana
Kentucky
Louisiana
Illinois
Quebec, Canada
Illinois
Goderich, Ontario
Mad. Is. , Canada
11 linois
Timmons, Ontario
Year
1888
1952
1955
1957
1963
1963
1964
1964
1965
1965
1967
1967
1967
1969
1973
1974
1974
1975
1975
1977
1978
1979
1979
1980
1980
1981
1981
Shaft
dimen . , m
4x5
4.6
4.9
4.9
5.5
4.26
5.64
4.9
5.5
4.9
5.64
5.5
4.9
4.9
4.9
4.9
4.26
6.09
2.44
6.09
5.5
6.09
3.65
6
6.09
6.70
7.3
3.65
Shaft
depth, m
233
1,051
228
1,030
305
1,021
1,089
1,004
1,052
1,089
1,068
1,021
1,000
378
478
1,004
228
470
122
381
122
61
335
518
122
87
174
Freeze
depth, m
30
107
914
761
363-437
128
468
625
527
591
684
548
461
469
70
76
527
38
70
54
64
53
61
41
61
43
64
43
treating wastewater and sludges (2,3,5,10),
for supporting large buildings during exca-
vation and construction, and for tunnel
construction (4,6,8,13,21). Braun and
Nash's (4) information on the use of arti-
ficial ground freezing for building mine
shafts in North America, presented in Table
2, indicates that artificial ground freez-
ing can be extended to a depth of more than
900 m below the ground surface and that the
hardware is currently available.
Artificial and natural freezing in de-
salting projects works because, when the
temperature of the saline solution is grad-
ually lowered below 0°C, ice nucleation
starts and ions are rejected. As the tem-
perature is further lowered, a large mass
of ice is formed and a concentrated brine
is left behind. The ice is then washed to
remove external salt and then it is removed
to another site and allowed to melt, form-
ing fresh water. Alternatively, the brine
is pumped out and the ice is left to melt.
The same principle can be used for treat-
ment of industrial effluent. However, in
this case we will be interested in the con-
centrated solutions for reuse, recovery or
further treatment. The gradual lowering of
temperature is very important because, if
the temperature drops fast, ions or salts
are entrapped between the ice particles and
the separation process will not be com-
plete.
The main purpose of sludge treatment
by freezing is for dewatering or volume re-
duction. Alum [Al 2 (S04)3] enhances coagu-
lation and improves the physical properties
of the sludge. It also improves the quali-
ty of the separated water, as heavy metals
and nutrients such as phosphate will pre-
cipitate with Al ions. Some systems, such
as those in Nekoosa and Port Edwards, Wis-
consin (16), and Burlington, Canada (17),
operate on a large scale. Both systems are
using natural freezing. The cost of energy
was a drawback in some projects; however,
124
-------
this method should be evaluated in the case
of hazardous waste.
Chamberlain and Blouin (7) investigat-
ed the use of freezing for dewatering pol-
luted dredged material, and they found that
freezing and thawing were effective.
Freezing affected the consolidation of
dredge material by 1) aggregating and re-
structuring soil particles, 2) densifying
aggregates, and 3) segregating ice.
Ground freezing and its effects have
been extensively studied in the laboratory.
Anderson and Morgenstern (1) summarized the
many effects of freezing on soil physical
and chemical properties as they relate to
engineering. The phenomenon of frost heave
is probably the most widely recognized ef-
fect. Frost heave results from the growth
of ice lenses in soils during freezing,
which causes an expansion of the soil mass
and a subsequent rise of the soil surface
above the freezing zone. Ice lens growth
requires water movement to the lenses from
the unfrozen soil beneath.
Frost heave should be a concern in un-
controlled hazardous waste sites where
frost penetration reaches the depth where
waste materials have been buried. For ex-
ample, soil-covered asbestos, or even bur-
ied barrels containing chemicals, could
slowly move upward as a result of frost
heaving. Properly designed soil covers
must be deep enough so that frost penetra-
tion cannot reach the hazardous waste.
Another important finding is that some
amount of unfrozen water is always present
in frozen soils as a thin film around the
particles. The amount of the unfrozen wa-
ter depends on factors related to soil
physical and chemical characteristics and
temperature. This is important in hazar-
dous waste sites because chemicals can move
even in frozen soils as concentrated solu-
tions in the unfrozen water films. Also,
in calculating the freezing point of con-
taminated soils, engineers must remember
that the chemicals could depress the freez-
ing point of the soil solution.
The successful use of artificial
ground freezing in many large-scale engi-
neering projects (Table 2) indicates that
the technology and the necessary hardware
are available. Artificial ground freezing
uses the Poetsch method. Briefly, this
method involves spacing vertical drill
holes with concentric refrigeration lines
along the desired freezing line. A header
or manifold system provides coolant to an
interior pipe, with the return line being
the outer casing (Fig. 1). A self-con-
tained refrigeration system pumps coolant
around the freezing loop. Temperature-
measuring instrumentation is appropriately
placed to monitor the progress of the
freeze front. The parameters controlling
the energy and time needed to freeze a soil
include soil thermal properties, refrigera-
tion line spacing, soil water content, ini-
tial temperature of soil and atmosphere,
and temperature of coolant.
Basically, two techniques for coolant
circulation have been used. The first in-
volves drilling holes in the ground around
the area to be excavated, and circulating
brine solution or other coolants at low
temperature as discussed above (Fig. la).
This technique cools the soil slowly and
allows for soil water or groundwater to
migrate and to form large ice lenses. The
Expansion
Valve
Cooler
Soil
Compressor
a.
Soil
Figure 1. Artificial ground freezing by
brine (a) and liquid nitrogen (b).
125
-------
advantages of this technique are its low
initial cost and maintenance cost, and the
regularly shaped walls it forms. Its dis-
advantages are the length of time it takes
to reach the desired low temperature and
the fact that the lowest temperature
achievable is only -55°C. The second tech-
nique uses liquid nitrogen (LN2) for cool-
ing (Fig. lb). The use of LN2 has the ad-
vantage of reaching much lower temperatures
(-196°C) in a much shorter time compared
with the brine cooling technique. Its dis-
advantages, however, are its high cost and
the irregularly shaped walls it forms.
Both techniques may be used in uncontrolled
hazardous waste sites, depending on the
situation and purpose.
For example, the brine cooling tech-
nique may be used for separation of solids
from liquids, forcing the contaminants to
move in soil solution in a desired direc-
tion, and in situations where time is not
an important factor. The use of LN2 will
be necessary if one wants to entrap contam-
inants in the soil mass, when time is a
critical factor, and when a very low tem-
perature is required. A very low tempera-
ture is necessary where high concentrations
of soluble salts or chemicals are present.
In some situations, a combination of the
two techniques - using LN2 initially to
reach the low temperature and then brine
cooling to maintain the freezing condition
- may be ideal.
In summary, ground freezing may have a
detrimental effect on uncontrolled hazard-
ous waste sites because of frost heaving
and ion migration in the leachate upon
freezing and thawing. In the northern
United States where many freeze/thaw cycles
occur every winter, the impact on contami-
nant movement and soil cover must be evalu-
ated. However, natural and artifical
freezing can be used beneficially in some
situations under proper engineering design.
The beneficial uses of freezing at uncon-
trolled hazardous waste sites are to immo-
bilize contaminants and to change the phys-
ical properties of sludge and sediments,
and currently it is being tested for treat-
ment of contaminated soils. The hardware
for artificial ground freezing is now
available.
COST AND ECONOMIC ANALYSIS
Artificial ground freezing for hazard-
ous waste containment is a proposed alter-
native to the traditional slurry wall or
grout curtain barrier techniques for imme-
diate action to halt contaminant movement
and for treatment of soils and sediments.
The parameters quantified for the cost
analysis of ground freezing include the
thermal properties of soils, refrigeration
line spacing, equipment mobilization and
freezing time constraints. Sullivan et al.
(19) developed a computer program and a
procedure for analyzing the cost of using
artificial ground freezing at uncontrolled
hazardous waste sites. The thermal data
required for the cost analysis are obtained
primarily from soil texture, water content
and temperature measurements. The specific
heat of soils depends on the water content
because the volumetric heat capacity ratio
of water to most dry soils is about 5. The
thermal conductivity of coarse-grained
soils is significantly higher than that of
fine-grained soils. Increasing the water
content of soils decreases the thermal con-
ductivity. A water-saturated soil system
is desirable for formation of an imperme-
able frozen barrier. Lunardini (14) pro-
vides extensive data relating soil physical
properties to soil thermal properties.
Therefore, soil thermal properties can be
predicted from soil physical properties.
The time constraint for forming the
frozen wall plays a critical role in some
situations where immediate action is re-
quired. In these cases, the drilled holes
should be closer to each other and liquid
nitrogen should be used. The cost associ-
ated with this process is high. If time is
not important, such as in cases where
treatment of contaminated soils or sediment
is desired, drilling spacing can be large
and reusable refrigerant used; the cost
would be much less. The cost for equipment
mobilization and site preparation depends
on the site location, proximity to highways
and contractors, and the nature of the
site.
The following hypothetical case gives
an overview of the calculations necessary
to obtain an optimum design for a ground
freezing project and to compare the cost
with other alternatives. Assume a chemical
spill occurred as a result of a truck acci-
dent or a derailed train car (Fig. 2).
Surrounding towns impose a 30-day time con-
straint on the chemical and transportation
companies for containment of the spill to
avoid contamination of a nearby water re-
source. A preliminary 1-week field inves-
126
-------
250
Connecticut River
Freeze Ring Placement
(130ft radius)
Toxic Spill
200ft
Interstate (Route 5)
Figure 2. A hypothetical chemical spill
(after Sullivan et al. (19)).
tigation to define the problem and to ob-
tain data on soil physical and thermal
characteristics is required. Assuming that
the pollutant diffuses horizontally at a
rate of 1 foot (ft) per day, the frozen
wall is planned as a circle with a radius
of 130 ft around the spill. From the in-
vestigation, the pollutant had reached a
15-ft depth. Figure 3 summarizes the cost
analysis data obtained.
The optimum cost design (least cost)
is about $70,000 (Fig. 3). This design
calls for 3.8-ft drill spacing with a 12-
day freezing time. If time is not suffi-
cient to freeze the soil, drill spacing
must be decreased to 3.7 ft and the cost of
drilling will be higher. From Figure 3,
the cost of fuel is not as sensitive as
time, drilling equipment rental and drill-
ing expenses. The time and drilling cost
(dependent on drilling spacing) are the
most sensitive parameters in the cost of
artificially freezing the soils. The more
time that is available, the lower the
drilling cost.
Although this is not an actual case,
it provides the reader with an idea of how
the program works and what parameters
150
o 100
50
" Drill Expenses
Fuel ~~~
50
30
zo
2 3 4 5 6 7 8"
Drill Spacing (ft)
Figure 3. Cost of artificial ground freez-
ing as a function of time and drill spacing
(19).
should be considered in planning to artifi-
cially freeze the ground. For details on
the model and the case discussed, the read-
er should consult Sullivan et al. (19). In
this example, ground freezing is used for
immediate action and only as a temporary
treatment because of the maintenance cost.
The added features of low noise and minimal
environmental disturbance when using ground
freezing may be desirable in urban areas.
The data needed to calculate the required
thermal properties of soils for technical
and economical analysis are routinely ob-
tained during the geotechnical and hydro-
logical site investigations. Consequently,
there are no additional site examination
costs.
ACKNOWLEDGMENT
This work was financially supported by
the U.S. Environmental Protection Agency
(EPA), Containment Branch, Land Pollution
Control Division, Hazardous Waste Engineer-
ing Research Laboratory, Cincinnati, Ohio,
and the U.S. Army Cold Regions Research and
Engineering Laboratory (CRREL), under In-
teragency Agreement DW 930180-01-0. The
authors thank Dr. Walter Grube and Jonathan
Herrmann of EPA, Cincinnati, for supporting
the project and initiating the agreement,
and for reviewing a draft of the resulting
technical report. This paper was techni-
cally reviewed by Francis Sayles of CRREL,
Dr. Dennis Keeney of the University of Wis-
consin, Madison, Dr. Grube and Mr. Herr-
mann. The information contained in this
127
-------
paper represents the authors' opinion and
not necessarily that of EPA or CRREL. Ci-
tation of brand names is just for conven-
ience and is not to be used for promotion
or advertising purposes.
REFERENCES
1. Anderson, D.M. and N.R. Morgenstern,
1973. Physics, chemistry and mechan-
ics of frozen ground: A review. Pro-
ceedings, 2nd International Conference
on Frozen Ground. NAS, Washington,
D.C., pp257-288.
2. Bishop, S.L., 1971. Methods for
treating wastes from water treatment
plants. New England Water Works Asso-
ciation. 85, pplO-24.
3. Bishop, S.L. and G.P. Fulton, 1968.
Lagooning and freezing for disposal of
water plant sludge. Public Works, 99,
pp94-96.
4. Braun, B. and W.R. Nash, 1982. Ground
freezing applications in underground
mining construction. Proceedings, 3rd
International Symposium on Ground
Freezing. U.S. Army Cold Regions Re-
search and Engineering Laboratory,
Hanover, N.H., pp319-326.
5. Burd, R.S., 1968. A study of sludge
handling and disposal. FWPCA, U.S.
Department of Interior, Publication
WP-20-4.
6. Careaga, J.A. and E.R. Mayer, 1972.
Frozen soil: A material to solve
problems in the construction industry.
Proceedings, Third Inter-American Con-
ference on Materials Technology, Mexi-
co, pp57-64.(U.S. Army Cold Regions
Research and Engineering Laboratory,
Draft Translation 480, ADA-012-170,
1975.)
7. Chamberlain, E.J. and S.E. Blouin,
1977. Frost action as a factor in en-
hancement of the drainage and consoli-
dation of fine-grained dredged materi-
al. U.S. Army Waterways Experiment
Station. DMRP, Technical Report
D-77-16.
8. Oorman, Ya.A., 1971. Artificial
freezing of soil in subway construc-
tion. Izd-vo. Transport. USSR (in
RussianT^
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
EPA, 1982. Handbook for remedial ac-
tion at waste disposal sites: EPA Re-
port No. 625/6-82-006.
Fulton, G.P., 1970. New York communi-
ty improves water supply systems. Wa-
ter and Sewage Works, 117, pp!44-147.
Geller, S.Yu., 1962. Desalting of wa-
ter by natural freezing for farm use.
Izvestiya Akademii Nauk, SSSR, Seriya
Geograficheskaya, No. 5, pp 71-77.
Heller, S.J., 1939. New method to ob-
tain fresh water in the desert. Prob-
lems of Physical Geography No. VII
NSSR., Translation by Heinze, Sas-
katchewan Research Council.
Low, G.J., 1960. Soi1 freezing to re-
construct a railway tunnel.Journal
of the Construction Division, ASCE,
86.
Lunardini, V.J., 1981. Heat Transfer
in Cold Climates. Van Nostrand Rein-
hold, New York, 731pp.
Neely, N.S., J.J. Walsh, D.P. Gilles-
pie and F.G. Schauf, 1981. Remedial
action at uncontrolled hazardous waste
sites. Land Disposal, Hazardous
Waste, Proceedings, 7th Annual Re-
search Symposium, 16-18 March, 1981,
Philadelphia, Pa., pp312-319.
Nekoosa Paper Inc., 1981. Nepco Lake
dredging disposal. Report no. 2, sub-
mitted to Wisconsin Department of Na-
tural Resources, Madison.
Rush, R.J. and A.R. Stickney, 1979.
Natural freeze/thaw sewage sludge con-
ditioning and dewatering. Canadian
Environmental Protection Service, En-
vironment Canada, Report 4-WP-79-1.
Stinson, D.L., 1976. Atmospheric
freezing for water desalinization.
In: Gray F. Bennett (Ed.), Water
1976. I: Physical. Chemical Wastewa-
ter Treatment. AICHE Symposium Ser-
ies, 166, vol. 73.
Sullivan, J.M., D.R. Lynch and I.K.
Iskandar, 1984. The economics of
ground freezing for management of un-
controlled hazardous waste sites.
Proceedings. 5th National Conference
on Management of Uncontrolled Hazard-
128
-------
ous Waste Sites, 7-9 Nov., 1984, Wash- 21. Tsytovich et al., 1973. Physics,
ington, D.C. Physical Chemistry and Mechanics of
Permafrost and Ice. Yakutsk Book Pub-
20. Szekely, T., 1964. Water purification lishing Co., Yakutsk, USSR. (U.S.
by freezing in dugouts. Eng. Div., Army Cold Regions Research and Engi-
Saskatchewan Research Council. neering Laboratory, Draft Translation
439.)
129
-------
OPERATION OF THE CENTER HILL PILOT PLANT AND SOILS LABORATORY
FOR SUPERFUND AND RCRA RESEARCH PROJECTS
Joseph Burkart
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
and
Gerard Roberto
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
The Center Hill Pilot Plant and Soils
Laboratory is operated under government
contract with the University of Cincinnati
and provides the U.S. EPA Hazardous Waste
Engineering Research Laboratory (HWERL)
with technical support for various Super-
fund and RCRA research projects.
A poster and slide presentation will
present the capabilities of the facility
and the general conduct of current research
projects.
The facility consists of an engineer-
ing laboratory complex to support geoscien-
tific and geotechnical research with empha-
sis on soil/chemical interactions.
A collection of "one-pager" handouts
on each project will be available. Many
of the U.C. principal investigators and EPA
technical project monitors will be attend-
ing the seminar and be available to discuss
their projects.
Capabilities of the University of Cin-
cinnati at the Center Hill Facility in-
clude:
Geotechnical and geochemical laboratories
to perform all conventional soil mechan-
ics lab tests and those special tests
which deal with hazardous waste.
A special permeability laboratory where
studies on soils and waste material are
performed to determine hydrological and
chemical resistance characteristics.
A geoscientific laboratory where pilot
scale and bench scale studies for the
treatment of hazardous waste and contam-
inated soils dealing with biological and
other special treatment processes are
conducted.
A high bay area where pilot scale testing
equipment for studies on the performance
and treatment of various waste remedial
techniques are conducted.
A wet chemistry laboratory for conven-
tional lab studies for treatment proces-
ses.
The facilities for conducting studies of
gas permeation through flexible membrane
liners.
130
-------
RECLAMATION AND REUSE OF CONTAMINATED LAND
IN THE UNITED STATES, ENGLAND, AND WALES
Tayler H. Bingham and Garrie L. Kingsbury
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
ABSTRACT
This poster will provide information
on selected case studies illustrating the
reclamation and reuse of land contaminated
with hazardous wastes. Examples presented
will be taken from a larger set of case
studies compiled by the Research Triangle
Institute as part of a study to investi-
gate issues surrounding the reclamation
and reuse of contaminated land including
uncontrolled hazardous waste sites.
Examples of redeveloped sites in the
U.S. and the U.K. have been documented
through record searches and site visits.
The following sites are included.
Frankford Arsenal, Philadelphia,
PA—cleanup and redevelopment of
160-year old arsenal for reuse as
downtown commercial/industrial
center.
Hercules Properties, Hercules,
CA--cleanup and redevelopment of
70-year old explosives manufacturing
property into commercial office park
and convention center.
Kapkowski Road Site, Elizabeth,
NJ--cleanup of uncontrolled waste
dump contaminated with PCB-laden oil
and redevelopment as industrial park.
Thamesmead, London, England—
reclamation of properties associated
with 300-year old Woolwick Arsenal
for redevelopment as housing for
60,000 people.
Beaumont Leys, Leicester,
England—cleanup and redevelopment of
former sewage works and 200-acre
sewage farm for redevelopment as a
new town.
Ravenfield Tip, South Yorkshire,
England—cleanup of former uncon-
trolled hazardous waste site and
subsequent landscaping as football
field and public open space.
Where meaningful, comparisons will be
made between the practices of the U.S. and
the U.K. with respect to reclamation and
reuse of contaminated lands.
131
-------
COMPUTER ANALYSIS OF HAZARDOUS W\STE DIKES
Andrew Bodocsi and Richard McCandless
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
The objective of the study was to
develop a "user-friendly", interactive com-
puter program for the geotechnical analysis
of dikes used around hazardous waste sites.
The program was designed to guide a permit
reviewer along the steps of the analysis of
an earth dike, considering slope stability,
bearing capacity, settlement, liquefaction,
piping, hydraulic flow and pressure condi-
tions. The program is to be used on the
IBM-PC/XT microcomputer.
In addition to the input and output sub-
routines, and the controlling main program,
the following subroutines have been devel-
oped:
- A slope stability and bearing capacity
analysis program. For this, the REAME
program by Dr. Huang of the University
of Kentucky was adapted and rendered
user-friendly. This program is based on
the Modified Bishop Method and, in its
adapted form, it will perform an auto-
matic search for the minimum factor of
safety with respect to slope stability
and bearing capacity;
- A settlement subroutine analysis program
to calculate the maximum settlement of a
dike and also the differential settle-
ments between the center!ine, crestline
and toe of a dike. The subroutine han-
dles the cases of both normally-consoli-
dated and overconsolidated sub-layers;
- Hydraulic subroutine programs to analyze
the flow of water or leachate through and
under a dike to calculate hydraulic pres-
sures and to check for potential lique-
faction and piping in the dike system.
A User's Manual was also prepared to facil-
itate the use of the program.
132
-------
EVALUATION OF STABILIZED DIOXIN CONTAMINATED SOILS
Donald E. Sanning
Edward J. Opatken
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
William D. Ellis
William H. Vick
Science Applications International Corporation
La Jolla, California 92038
ABSTRACT
Asphalt and portland cement are
being tested as stabilizers to reduce
the release of dioxin from three
contaminated soils from the St. Louis,
Missouri, area. The soils include a
sandy loan from the Piazza Road site,
a sandy silty loam from the the Minker
site, and a silty loam from the Sontag
Road site. Standard mix design
methods for preparing soil asphalt
using emulsified asphalt and cutback
asphalt, and for soil cement, are
being followed to prepare test
specimens meeting highway industry
criteria for strength and weathering
performance. Cement contents ranging
from 7 to 12 percent have produced
acceptable specimens for the three
soils. The initially designed
emulsified asphalt specimens did not
meet acceptance criteria at any
asphalt concentration, probably due to
the high soil fines content, so a lime
treatment is being tested to
agglomerate the fines before asphalt
stabilization.
Once acceptable soil cement and
soil asphalt mixes are designed, fresh
test specimens will be prepared and
subjected sequentially to repeated
wetting and drying cycles, then to
repeated freezing and thawing cycles,
and then to prolonged leaching with
water according to a modified solid
waste leaching protocol. Dioxin is
known to have very low water
solubility, so movement of dissolved
dioxin into the environment is not
expected to be a significant factor.
But the movement of dioxin into the
environment by erosion and the
suspension in surface runoff of dioxin
contaminated clay and fine silt
particles can potentially be a
significant release mechanism.
Therefore, both the leachate and soil
fines suspended in the leachate will
be analyzed for 2,3,7,8-tetra-
clorodibenzodioxin (TCDD), the most
toxic dioxin isomer, to assess the
effectiveness of cement and asphalt
stabilization methods for preventing
the escape of dioxin into the
environment.
133
-------
TWO AUTOMATED SYSTEMS FOR ON-SCENE COORDINATORS: THE COUNTERMEASURE
SELECTION SYSTEM AND THE CASE HISTORY FILE OF RESPONSES TO
HAZARDOUS SUBSTANCES RELEASES
R. F. Hildreth, S. C. Fullerton
Science Applications International Corporation
8400 Westpark Drive
McLean, Virginia 22102
R. A. Griffiths
Releases Control Branch
U.S. Environmental Protection Agency
ABSTRACT
Two automated data bases have been
developed for the Releases Control Branch,
Land Pollution Control Division, Hazardous
Waste Engineering Research Laboratory,
USEPA by Science Applications International
Corporation (SAIC). They are intended to
be used by On-Scene Coordinators (OSC) and
their staffs in planning and implementing
spill cleanups and remedial response
actions. They can be accessed directly by
computer terminal from the response site if
desired.
The Countermeasure Selection System
provides On-Scene Coordinators (OSC) with a
computerized database of information to
help them select appropriate countermeasures
when responding to hazardous spills. The
system is based on the EPA "Manual of
Countermeasures for Hazardous Substance
Release", a compendium of information on
the hazards and behavior of chemicals when
spilled and potential spill countermeasures.
The system guides the user along a
decisionmaking path by providing information
focussed on the incident at hand. The
information in the database consists of data
arranged in three tables:
• Characteristics of the hazardous
substances spilled
• Applicable countermeasures for the
chemical class of the spilled substance
t Limitations and requirements of
applicable countermeasures
For example, an OSC might want to
determine applicable countermeasures for an
incident involving a railroad tank car
spill of acrylonitrile in a flood plain.
The OSC would look up the chemical in the
first table to determine its characteristics;
that leads to the second table, which
identifies technically feasible counter-
measures. The third table presents the
limitations and special requirements of the
countermeasures and gives references to
paragraphs in the manual having further
information to help the OSC reach a final
decision.
The Case History File provides a
database of information pertaining to past
responses to releases of hazardous
substances. The File is a computerized
system that helps an OSC's staff determine:
• if there have been previous incidents
like this one,
• how they were handled,
• how successful the effort was,
• how much it cost.
Previous incidents can be characterized
and accessed by:
• date
• location
• hazardous material involved
134
-------
• interacting substances
• geographical and geo-hydrological
characteristics
• affected area, population, and
resources
• containment actions
• cleanup actions
• treatment technologies used
• other factors.
The File also contains narratives with
additional information on each entry,
including the OSC's rationale for taking a
particular course of action. Cost
information is included to help gauge the
feasibility of a given response action.
For example, an OSC might want to find
information for a case involving a railroad
spill of acrylonitrile in a flood plain.
Then he might want to read the text to see
what unique problems might be encountered
or what containment methods were used and
how well they worked.
The system also provides an interactive
procedure for entering after-action reports
on new incidents. This helps ensure the
timely entry of new data and improves the
currency of the data base.
The Countermeasure Selection System
and the Case History File are two automated
systems for OSC use that are presently
being developed for the Releases Control
Branch, Land Pollution Control Division,
Hazardous Waste Engineering Research
Laboratory.
135
-------
VOLATILE ORGANIC EMISSIONS REDUCTIONS FROM SURFACE
IMPOUNDMENTS BY THE USE OF WIND FENCES AND WIND BARRIERS
Louis J. Thibodeaux
Hazardous Waste Research Center
Louisiana State University
Baton Rouge, Louisiana 70803, U.S.A.
Charles Springer and Rebecca S. Parker
Department of Chemical Engineering
University of Arkansas
Fayetteville, AR (USA) 72701
ABSTRACT
Quantities of hazardous waste materials in liquid, sludge or solid form can be
disposed on or near the soil surface. A common disposal operation uses surface impound-
ments for aqueous waste. Proper design and operation procedures can result in signifi-
cantly reduced emissions of hazardous volatile species to the surrounding air.
Surface impoundments usually consist of shallow earthen basins. Aqueous and non-
aqueous waste are placed in these open top basins for treatment of selected waste or for
storage. Volatiles may escape to the air boundary layer but significant reduction in
emissions rate is possible by the use of perimeter fences and network fences. Desorption
data was obtained on a pilot scale apparatus using ethyl ether as the model chemical.
Forty percent reduction in emission of ethyl ether was achieved with this liquid-phase
controlled volatile chemical using solid perimeter fences. Seventy eight percent re-
duction was achieved with solid fence networks.
INTRODUCTION
A surface impoundment is typically
a large, open-top vessel of earthen
construction containing aqueous waste.
These vessels can also contain a float-
ing hydrocarbon layer on the surface
and/or a semi-solid sediment on the
bottom. One important characteristic is
that it has a liquid interface in direct
contact with the atmosphere. Hazardous
substances including chlorinated sol-
vents, hydrocarbon solvents, pesticide
waste, inorganic and other classes of
volatile chemical compounds are con-
tained in some of these impoundments.
These are usually surface impoundments
used for managing hazardous waste.
Uses of surface impoundments for
hazardous waste management include:
biochemical treatment of the organic
substances, chemical treatment of organics
and inorganics (i.e., neutralization),
holding (storage) prior to performing
other treatment operations and water
evaporation for liquid volume reduction.
While these uses are occurring, volatile
chemical species in solution of the
aqueous phase can evaporate to the air.
There are problems associated with
the emission of volatile organic chemical
from surface impoundments. The loss of
solvents in large quantities can modify
the general ambient air quality of the
nearby air mass from the point of view
of photochemical oxidants content.
Odors, associated with some selected
organic and inorganic compounds, are a
136
-------
common impact on local environment near
residential areas (Wilson, et. al.,
1980). The emission of specific hazard-
ous chemicals such as benzene
(Thibodeaux, et. al.) can result in
increased levels of exposure (and health
risk), for both workers and nearby
residents.
Control methods for volatiles have
been proposed and there are several
possible techniques that can be used to
reduce emissions. These control methods
fall into three general categories:
covers, fences and adsorbents. Covers
involve the placement of a barrier at
the water-air interface. Oil (slicks)
layers, plastic films, mats of floating
materials and foams have been proposed.
Fences, both perimeter and surface
networks, have been proposed to partial-
ly nullify the wind enhanced
volatilization. On-bottom or suspended
adsorbents (liquid droplets and solid
particles) have been proposed as a means
of retarding the evaporation by reducing
the effective concentration in the water
(Cudahy and Sandifer, 1980). All three
techniques inhibit evaporation allowing
for some treatment to be effective.
This manuscript presents a brief
description of the chemodynamics mecha-
nisms and mosels for volatile chemical
emission from surface impoundments.
Analysis of the rate equation suggests
ways of controlling volatile emissions.
Finally experimental results of emission
reduction schemes using wind barriers in
pilot-scale equipment are presented.
Preliminary design equations are pro-
posed.
CHEMODYNAMIC MECHANISMS OF VOLATILIZATION
The general concepts to explain
emission of volatile chemicals from
aqueous solutions in surface impoundment
type operations were developed by
Thibodeaux and Parker in 1974 (Thibo-
deaux, 1979). A brief review of these
mechanisms and models is needed.
Volatilization will be considered
to be occurring with the chemical in
aqueous solution at a concentration of
C , g/cm3. If the overlying air mass
contains the volatile chemical species
at concentration C,,, g/cm3, and C./H is
less than C , than the emission flux
rate from trie water surface is
N = K (Cw - CA/H) (1)
with N in g/cm2-s. H is the Henry's
constant for the volatile species in cm3
H20/cm3 air and K is the overall transport
coefficient in cm/s. This coefficient is
commonly expressed in terms of two resis-
tances:
1/K = 1/kj + 1/k H'
(2)
where the 1/ki term accounts for the
resistance on the liquid-side of the
interface and 1/k H is the gas-side
resistance. For most chemicals of a
hazardous nature k H » kj and the liquid
phase controls thegvolatilization process.
Each of these terms must account for both
mechanical and nature stimulated process-
es.
Wind constitutes a natural process
that has a large influence on the magni-
tude of ki and k . Generally kx increases
with the square of the wind velocity and
k increases proportional to the wind
velocity. In the absences of wind the
temperature difference between water and
air is the important independent variable.
If the water is warmer than air, both kx
and k are a function of AT to the 1/4
power? Mechanical surface aerators,
installed for increasing the rate of
molecular oxygen transfer to the water,
also have a dramatic effect on kx and to a
lesser extent on k .
IMPLICATIONS FOR VOLATILE EMISSION CONTROLS
Laboratory and field studies indi-
cate that the above mechanisms are
operative in surface impoundments and
that Equations 1 and 2 represent adequate
models for quantifying volatile chemical
emission rates. This being the case,
then an analysis of the model equations
and concepts may lead to methods by which
the emission rate may be reduced.
Volatilization is a well known first
order process as indicated by Eq. 1.
Reducing the concentration of the vola-
tile species in the water will then
reduce the emission rate a proportional
amount. A reduction in the overall
transport coefficient can bring about the
137
-------
same effect. For a given concentration
level in the water, reducing K provides
the most practical approach to reducing
the emission rate. Means of adding
additional resistances in series or by
increasing kL will add to the overall
resistance. This can be seen by in-
specting Eq. 2.
Adding additional resistances can
be performed by placing material at the
air-water interface. This is the
technique of oil films, plastic films,
floating mats, etc. mentioned above.
This process has some disadvantages that
include reduced oxygen transfer, high
maintenance, incomplete coverage (oil
films) and additional volatile source
(oil evaporation).
The use of fences to reduce the
effect of the wind on both kj and k is
a potential control method that in-9
creases the overall resistance. Parker
(1984), employing a theoretical model
developed by Levich for the influence of
protrusions in a fluid stream in reduc-
ing mass transport, along with Chilton-
Colburn analogy theory, developed an
expression that related kj to the
protrusion height:
= constant h
-i
(3)
where h is the vertical fence height in
cm. Advantages of fences are low
installation and maintenance costs.
PILOT-SCALE EXPERIMENTS UITH FENCES
Parker (1984) performed a series of
pilot-scale simulations in a wind tunnel
water tank (WWT) to investigate the
effectiveness of fences in reducing
volatile chemical emissions from water
bodies. The wind tunnel was 6 meters
long and 0.6m x 0.6m in air flow area.
Underneath was a water tank 0.6 m
maximum depth and 0.6 m wide. The
air-water contact surface length was 2.4
m. Ethyl ether, which is a liquid-phase
controlled volatile, was placed in the
water tank at an initial concentration
of approximately 100 mg/1. Batch evapo-
ration experiments, involving monitoring
of the decrease in ethyl ether concen-
tration with time and extraction of the
first order rate constant, yielded
valves of k!. Numerous batch experi-
ments with various fence arrangements
yielded various values of kj. The uncon-
trolled kj was obtained without a fence or
obstruction present in the WWT.
A total of 37 batch desorptions were
performed with various fence arrangement
around the water surface in the WWT. Two
basic fence patterns were investigated.
These were perimeter fences and network
fences. A perimeter fence is one that
encircles the impoundment at the water's
edge. Network fences involve the instal-
lation of fences on (i.e., floating or
piles) the water surface creating a series
of uniformly spaced cells.
The simulation experiments performed
in the WWT were scaled to an existing
surface impoundment studied at one parti-
cular hazardous waste facility in western
New York State (Thibodeaux, et al.). This
surface impoundment was nearly square in
plan area with sides of 186 meters and
4.23 meters in depth. Table 1 shows the
relation and scaling factors between the
field impoundment and the laboratory
model.
The degree of emission reduction is
defined by:
Reduction = NW/O
x 100
N.
(4)
Wo
where N , and N is the emission rate
without fences aPld with fences respect-
ively. To nullify the effect of air
concentration, C. was kept at zero or at
C./H « C in the inlet air at all times.
ATso the effect of C can be nullified by
assuming identical values both with and
without fences. Substituting Eq. 1, into
Eq. 4 with the above conditions, yields:
Reduction =
x 100
(5)
k
l,w/o
So it is possible to quantify the emission
reduction potential of a particular fence
arrangement by measuring kj values.
The results of experiments at wind
speed of 390 cm/s appears in Table 2.
These results represent an average of
three to five batch experiments for each
fence arrangement. The 90° angle indi-
138
-------
cates a vertical fence; the 80° angle
indicates a 10° downwind tilt and 45°
angle a 45° downwind tilt to the upwind
fence. The porous fence as a commercial
fabric sold under the trade name
DUSTAMER®*. The two series of network
fence studies are for fences placed 5h
and 20h distance apart on the water
surface, where h is the fence height.
The simulation test results repre-
sented in Table 2 include the ethyl
ether desorption mass-transfer (m-t)
coefficient, kj in cm/h, and the stan-
dard deviation (sd). The first line
entry in the table is kj without any
barriers. The percent reduction for
each barrier arrangement is calculated
using Eq. 5 and appears in the last
column. For a vertical perimeter fence
with an air speed of 390 cm/s, the
volatile chemical desorption rate is
reduced by 19.7% for a low (1.7 cm.
model height = 1.2 m. field height)
fence and 34.3% for a high (6.4 cm.
model = 4.8 m. field) fence.
Results from a high fence tilted 10
degrees downwind indicates a 39.1%
reduction. However, if the kj. ± 2 sd is
calculated and compared for each angle
there is significant overlap. This
suggest that the difference in the
average percent reductions are insignif-
icant. A slight downwind tilt, as used
in snow drift control, appears to have
no significant effect in volatile
emission control.
The last three line entries in
Table 2 represent a porous perimeter
fence and a network of solid fences.
The kj ± 2 sd for the no-fence case and
the porous fence case do not overlap.
The 78.31 reduction is therefore signif-
icant and represents a high degree of
volatile emission control. 'Simulation
results for a series of fences (i.e.,
networks) spaced 5x heights and lOx
heights apart over the surface are also
significant and the degree of volatiles
control is also high. (F=fence spacing).
It is possible to combine the
results of both the perimeter and network
experiments by using a dimensionless
scaling factor. The study of ki values
in Table 2 suggest that spacing between
fences is an important independent
variable. The use of the dimensionless
ratio, F/h, seems to allow consolidation
of both sets of data into a single
parameter. Figure 1 is a plot of the 90°
solid perimeter and network fence data.
A linear regression of this data yielded
the following equation:
= 0.328. (F/h)
0.4619
(6)
*Mention of trade names or commercial
products does not constitute endorsement
or recommendation for use by U.S. EPA.
The statistical correlation coefficient
was .9513 indicating a high degree of
correlation. It is also interesting to
note that the experimental ki is propor-
tional to h to the -0.4619 power and the
theoretical model was -0.5 (see Eq. 3).
By employing kj , = 3.424 in Eq. 5 it
is possible to (Jbtain a generalized
equation for the design of a fence
volatile control scheme. Using Eq. 6 for
kljW yields
% Reduction=[l - .096 (F/h)'462]xlOO (7)
for 1.6 < F/h < 160. The data and Fig. 1
also suggest that very little reduction
is achieved for fence spacings of 160 or
greater. Therefore, % reduction = 0, F/h
> 160 completes the design equations.
A limited number of experiments were
performed at a lower wind speed of 290
cm/s with fence heights of 1.7 and 3.2
cm. This data appears in Table 3. The
no-fence m-t coefficient for the low wind
speed is much lower than that for the
high wind speed value. Lunney (1983)
observed this same effect and correlated
coefficients to the square of wind speed.
There is a corresponding less amount of
volatile chemical emitted at the low wind
speed. However, it appears that the
average percent reduction is slightly
higher for the low wind. Comparing data
in Tables 2 and 3 indicates that the low
and medium, 90°, fence heights display
40.9% and 20.3% reductions for the low
wind and 19.7% and 22.8% reductions for
high wind respectively. The true signi-
ficance of the average percent reductions
in the case of low wind is questionable
considering the overlap in the ki ± 2 sd
values.
139
-------
Assuming the 0.462 power relation
holds for the F/h variation on emission
reduction for the 290 cm/s wind it is
possible to obtain a correlation similar
to Eq. 7. This correlation is:
% Reduction=[l-.0836 (F/h)'462]xlOO (8)
for F/h < 215. If F/h > 215 the %
reduction is zero. Figure 2 is a
graphical representation of Equations 7
and 8.
It should be noted that the above
design equations (Eq. 7 and 8) are
tentative in that they are laboratory
observations and involve results based
on one chemical and two wind speeds.
Hopefully continued research will
involve other volatile chemicals,
additional environmental parameters
including water temperature and field
testing.
CONCLUSIONS
A properly conceived and formulated
mathematical model of a natural process
can be a powerful tool for studying
means of controlling transport rates of
chemical species. In the case of
surface impoundments, fences appear to
be an effective volatile chemical
control device in the presence of wind.
a perimeter fence of solid construction
(i.e., non-porous) can give up to
approximately 40% volatile emission
reduction. A network grid can give up
to 80% reduction as can a porous perime-
ter fence. Fence angle does not in-
crease the effectiveness of a solid
perimeter fence as a means of emission
reduction.
ACKNOWLEDGMENT
This work was supported in part by
the U.S. Environmental Protection Agency
Municipal Environmental Research Labo-
ratory, Cincinnati, Ohio under a cooper-
ative agreement No. CR810856 with the
University of Arkansas, Fayetteville,
AR. Mr. Stephen James served as the
Project Officer. This document has been
reviewed by the U.S.E.P.A.
140
-------
Table 1
Scaling Factors for Laboratory Simulation
Parameter
Laboratory
Model
Field
Equivalent
wind speed
wind speed
290 cm/s @ 10 cm
390 cm/s @ 10 cm
5.6 m/s @ 10 m
8.0 m/s @ 10 m
fetch*/depth ratio
fetch*
fence height
fence height
fence height
44
244 cm
3.2 cm
1.7 cm
6.4 cm
44
186 m
2.4 m
1.2 m
4.8 m
*fetch (F) = distance wind travels over open water.
between fences, or fence spacing.
In this study it is distance
Table 2
Fence Control Simulation Results (V = 390 cm/s)
Parameter h, height (cm) mass-
transfer
coefficient
k^cm/h)
standard
deviation
(cm/h)
%
reduction
No fence
Solid, perimeter, 90°
Solid, perimeter, 90°
Solid, perimeter, 90°
Solid, perimeter, 80°
Solid, perimeter, 45°
Porous, perimeter,90°
Solid, network 90°
F = 5 h
F = 10 h
1.7
3.2
6.4
6.4
6.4
6.4
1.7
1.7
424
750
642
251
087
211
.7445
.7328
.7540
.0395
.224
.406
.160
.085
.134
.119
.0908
.0454
0.0
19.7
22.8
34.3
39.1
6.2
78.3
78.6
78.0
141
-------
TABLE 3
Fence Control Simulation Results (v = 290 cm/s)
Parameter
no fence
solid
perimeter, 90°
solid,
perimeter, 90°
h, height (cm)
0.0
1.7
3.2
mass-transfer
coefficient
k: (cm/h)
2.375
1.403
1.893
standard
deviation
(cm/h)
0.333
0.460
0.673
average
°/
7o
reduction
0.0
40.9
20.3
LITERATURE CITED
1. Thibodeaux, L.
dynamics",
1979.
(1979), "Chemo-
J. Wiley, N.Y.,
2. Cudahy, J.J. and R.L. Sandifer
(1980), Emission Control
Options for the Synthetic
Organic Chemicals Manufacturing
Industry. Secondary Emissions
Report.EPA"ContractNoT
68-02-2577, Research Triangle
Park, North Carolina, June
1980.
3. Parker, R.S.P. (1984), Fences for
Reduction of Wind Enhanced
Liquid Phase Mass-Transfer
Coefficients Over Water Bodies,
M.S. Thesis, University of
Arkansas, Fayetteville, AR.
4. Wilson, G.E., J.Y.C. Huang and T.W.
Schroepfer, "Atmospheric
Sublayer Transport and Odor
Control," Jo. Environmental
Engineering Division, ASCE,
Vol. 101, No. EE2, April 1980,
p. 389-40.
5. Thibodeaux, L.J., C. Springer, P.
Lunney, S. James and T.T. Shen,
"Air Emission Monitoring of
Hazardous Waste Sites," Nov.
29-Dec. 1, 1982, Washington,
D.C.
6. Lunney, P.D., "Characteristics of
Wind and Depth Upon Liquid
Phase Mass Transfer Coeffi-
cients: Simulation Studies"^
M.S.Thesis,Universityof
Arkansas, January 1983.
142
-------
1.0
0.8
0.6
~ 0.4
5 0.2
-0.2
-0.4
A NETWORK SOLID FENCES
| SOLID FENCES
t I s.d.
V(IOcm)= 393 cm/s
1.0
2.0 3.0
LN(F/h)
4.0
5.0
FIGURE I. LOG PLOT OF LIQUID COEFFICIENT
VERSUS (F/h)
143
-------
100
DATA BASED ON ETHYL ETHER
IN PILOT SCALE SURFACE
IMPOUNDMENT.
WIND = 390cm/s
MODEL
I8.0m/s FIELD)
v—WIND =290 cm/s MODEL
(5.6 m/s FIELD)
20 40 60 80 100 120 140 160 180 200
WIND FETCH -r FENCE HEIGHT, F/h
FIGURE 2. PRELIMINARY DESIGN-VOLATILE
EMISSION REDUCTION BY WIND BARRIERS
144
-------
INVESTIGATION OF FLOATING IMMISCIBLE LIQUIDS TO CONTROL
VOLATILE ORGANIC CHEMICAL (VOC) EMISSIONS FROM SURFACE
IMPOUNDMENTS
K. T. Valsaraj, Charles Springer, Trac Nguyen and Louis J. Thibodeaux
Department of Chemical Engineering
University of Arkansas
Fayetteville, AR 72701
ABSTRACT
Floating immiscible liquids as a control option for volatile organic chemical
emissions from a laboratory scale waste water impoundment were investigated. Three
different immiscible organic liquids were used, viz., mineral oil, octanol and lauryl
alcohol. A three phase resistance-in-series model was proposed. Experiments with
complete surface coverage and thick mineral oil layer showed a considerable decrease in
the emission rates only for those compounds with low oil solubility. On the other hand
the emission rate was less significantly reduced for chemicals with high oil solubility.
INTRODUCTION
Field measurements as evidenced by
previous work [1-4] have demonstrated that
significant amounts of volatile organic
compounds (VOCs) are being released to the
air from surface impoundments containing
hazardous waste. Several control methods
lare being investigated which would help re-
duce these VOC emissions. Two such control
options currently under study are:
(i) the use of floating non volatile
immiscible organic liquids (under no-wind
conditions) and
(ii) the modification of the turbulent
boundary layer by use of wind barriers (for
wind enhanced conditions).
This paper presents ongoing work on
the use of floating immiscible liquids as
options for controlling VOCs from surface
impoundments
PURPOSE
The goal of this research project is
to provide information on the feasibility
of using floating immiscible liquids as a
:control option for VOC emissions. It has
observed that in the absence of
intentional agitation, the VOC emission
rate is a function of (among other factors)
wind velocity [5-7]. However, in the
absence of wind-induced stirring, natural
convection would be a significant mechanism
in the emission process when impoundment
liquid is warmer than the overlyinq air
[8,9].
APPROACH
Two approaches are being pursued in
this research. One is the procurement of
data on VOC emissions from a laboratory
impoundment simulator containing hydro-
carbons in aqueous solution covered by an
immiscible liquid surface layer. The other
is to develop and verify models for such
emissions.
Pilot-scale Simulator: The impoundment
simulator consisted of a circular tank, six
feet in diameter and two feet deep. It was
insulated on the bottom and sides with
about 6 inches of polystyrene foam insula-
tion. The apparatus was located indoors,
in a room which was approximately 55' x 60'
with a ceiling height of 14'. To further
shield the apparatus from extraneous air
currents, it was surrounded by an enclosure
made of polyethylene film. The enclosure
was annroximatelv 1*' x 16' x 8' hi ah, open
145
-------
at the top. Figure 1 shows a view of the
simulator and designated sample locations
within the polyethylene shield. The
temperature profile of the liquid as well
as of the overlying air could be measured
by thermocouples mounted on a standard in
the center of the tank. Some additional
apparatus consisted of a removable steam
coil for initial heating.
An experimental determination began by
filling the simulator with water and warm-
ing it to about 40-42°C using the steam
coil. Four different chemicals, viz.,
acetone, ethyl ether, isopropanol and ben-
zene were then separately added and stirred
in for a uniform concentration. These
chemicals were chosen because they repre-
sented a broad range of solubility in the
mineral oil and a range of mass transfer
properties, i.e., degree to which the
emission might be gas or liquid phase con-
trolled. The floating immiscible liquid
was then poured onto the surface. Approxi-
mately an hour was allowed for the currents
induced by mixing to subside [9]. Samples
were taken at designated locations using
Vacutainer tubes (supplied by VWR Scien-
tific) and analyzed using a Tracer 565
chromotograph with flame ionization de-
tector.
The experimental overall mass transfer
coefficients were deduced from the slopes
of the plots of ln(C/CQ) versus time [8]
using the equation
In 4- = - 5- t (la)
where d is the depth of the impoundment, C
is the concentration of the organic com-
pound in water, CQ is the initial concen-
tration of the organic compound, and K-| is
the overall mass transfer coefficient for
the organics from the aqueous phase to the
atmosphere.
Three Phase Resistance Model:
A three phase resistance-in-series
model was developed to explain the trans-
port of chemicals from water to air [10].
The relevant interfaces and corresponding
mass transfer coefficients are shown in
Figure 2. The equation for the overall
mass transfer coefficient thus derived was
1
1
k7~
1
1
koil Kp
(Ib)
where
K-| is the overall mass transfer coef-
ficient, cm/s.
k] is the water phase mass transfer coef-
ficient, cm/s.
k -I is the oil phase mass transfer coef-
ficient, cm/s
k is the air phase mass transfer coef-
y ficient, cm/s
K is the partition coefficient of the
chemical between oil and water phases
(dimensionless).
H is the Henry's constant (dimension-
1 \ concentration in vapor
'' concentration in solution
The water and air phase coefficients
can be obtained from previous correlations
[8].
1/3
. o.SO
(2)
(3)
where
D
'Al
g =
vi =
v2 =
diffusivity
air (cmvs)
diffusivity of the chemical in
water (cm2/s)
gravitational constant (cm/s^)
kinematic viscosity of air (cm2/s)
kinematic viscosity of water
(cm2/s)
thermal expansion coefficient
of water at temperature T2 (K~^)
146
-------
ENCLOSURE TOP OPEN
WOOD
FRAME
DOOR
SAMPLER
SUPPORT
THERMOCOUPLE
SUPPORT
INSULATION
SAMPLER
TUBES
POLYETHYLENE
FILM SIDES
l\
WATER
TANK
FIG. I. IMPOUNDMENT SIMULATOR- NO WIND
(Cut away view)
-------
voc
EMlSSIOf
[3) WATER
at T2
FIG 2. CONTROL SYSTEM INTERFACES
T2>T,
148
-------
B-, = thermal expansion coefficient of air
at temperature T^ (K"')
T ... = temperature at the oil-air inter-
face (K)
T . = temperature at the oil-water inter-
OW1 face (K)
T-| = Air temperature (K)
T2 = Water temperature (K)
The temperatures were measured using
the thermocouples located at the various
points on the thermocouple support
(Figure 1).
The oil phase coefficient, k .^ was
estimated using the simple film theory as
a ratio of diffusivity of the chemical in
oil to the height of the oil layer (h0)
koil =
uAo
ho
(4)
The oil phase diffusivity of the
chemical was calculated using the Wilke-
Chang correlation [7]
0.5
117.3E-18 (Moil) Toil
TO
(5)
where
M ., = Molecular weight of oil (kg/mol)
T0-ji = temperature of the oil layer (K)
y ., = viscosity of oil (kg/m.s)
V = Solute molal volume at normal
boiling point (m3/kmol)
The experimental value of overall mass
transfer coefficient determined from
experiment and evaluated by equation (la)
was then compared with the value pre-
dicted by the model from equation (Ib).
RESULTS AND DISCUSSION
As of December 1984 only the mineral
oil system has been completely investi-
gated. Work on the other two systems, viz,
1-octanol and lauryl alcohol has been only
partially completed and results are in-
conclusive. The latter two systems were
chosen because the partition coefficients
of the chemicals between water and 1-
octanol or lauryl alcohol are readily
available in the literature while for
mineral oil (which is a mixture of several
hydrocarbons) these measurements are not
as reliable.
For the mineral oil system four chemi-
cals (acetone, ethyl ether, benzene and
isopropanol) and 5 different oil thick-
nesses (vis., 0.20, 0.27, 0.30, 0.38 and
0.46 cm) were investigated. Table 1
summarizes the effect of oil depth on the
VOC reduction of the 4 chemicals for a
water depth of 49.5 cm. Table 2 gives the
experimental and predicted (by the model)
mass transfer coefficients for the 4
chemicals investigated and their dependence
on oil depth.
As can be observed from Table 1, the
percent reduction increases slowly with the
oil thickness for n-propanol, ether and
benzene; whereas, for acetone the percent
reduction increases rapidly as the oil
layer thickens. For a 0.5 cm of oil thick-
ness, the evaporation rates of acetone and
n-propanol were reduced by 80-90%; however,
with the same thickness of the oil layer
it only shows about 30-40% reduction for
ether and benzene.
Based on the properties of the chemi-
cals, acetone and n-propanol are miscible
in water and have very low oil solubility
as is evident from their mineral oil/water
distribution coefficients determined
experimentally (Table 1). Because of their
low concentration gradients across the oil
film, acetone and n-propanol would be ex-
pected to move through the oil film very
slowly after reaching the water-oil inter-
face from the bulk water. In this case,
the desorption rates of acetone and n-
propanol are well controlled by the oil
phase.
Table 2 shows that the experimentally
determined overall coefficient, K-i, that
was predicted by the model, does not agree
at all for chemicals with high oil phase
resistance (viz. acetone and n-propanol)
while the agreement is satisfactory for
those with low oil phase resistance,
especially benzene. This suggests that
the simple film theory prediction of k .-|
may be in considerable error and this
aspect of the model is presently being in-
vestigated. Also efforts are underway to
improve estimates of oil phase diffusi-
vities.
It has been found recently that
149
-------
Ul
o
Table 1: Percent VOC Reduction of Four Chemicals vs Oil Thickness Under No-Wind Conditions
at a Water Depth., h-49.53 cm.
Oil Thickness, cm
0.20
0.27
0.30
0.38
0.46
Mineral oil /water
distribution
@
Percent VOC reduction
Percent Reduction =
Temperature Ranges, °K
Air
303-309
300-306
296-297
304-308
304-307
here is
Rate of
Water
320-318
321-318
317-314
320-318
319-317
Standard Deviation
defined as follows
desorption without
Percent
Acetone Ether
45.9 15.3
51.2
75.3
64.6 20.0
84.0 26.5
7.66 1.4
0.05±0.02 3.85+0.
(See reference 10)
mineral oil -Rate of
Reduction"
1-Propanol Benzene
83.9
88.5
87.7
89.1
92.4
0.9
20 0.25+0.
desorption
23.4
28.2
27.0
37.5
40.6
1.6
03 2454+300
with mineral oil
Rate of desorption without mineral oil
-------
Table 2: Comparison of Experimental and Model Mass Transfer
Coefficients at Different Oil Thicknesses
Oil Thickness-
cm
0.20cm
0.27cm
0.30cm
0.38cm
9.46cm
0.20cm
0.38cm
0.46cm
0.20cm
0.27cm
0.30cm
0.38cm
0.46cm
0.20cm
0.27cm
0.30cm
0.38cm
0.46cm
Chemical
Acetone
Acetone
Acetone
Acetone
Acetone
Standard Dev.
Ether
Ether
Ether
Standard Dev.
n-Propanol
n-Propanol
n-Propanol
n-Propanol
n-Propanol
Standard Dev.
Benzene
Benzene
Benzene
Benzene
Benzene
Standard Dev.
K-, Experiment
cm/hr
0.97
0.88
0.44
0.64
0.29
0.14
2.09
1.99
1.81
0.04
0.35
0.25
0.27
0.24
0.17
0.02
1.48
1.39
1.41
1.21
1.15
n nfi
K-, Model
cm/hr
8.97E-3
6.66E-3
6.66E-3
4.75E-3
4.06E-3
0.12
0.06
0.05
8.97E-3
6.66E-3
6.66E-3
4.75E-3
4.06E-3
1.07
1.00
0.94
0.86
0.73
partial coverage on the surface and forma-
tion of lenses (a thick layer of organic
liquid whose shape is constrained by the
forces of gravity and interfacial tensions)
sometimes leads to enhanced VOC emissions.
Though work on both octanol and lauryl
alcohol are not far enough along to report
results, there is evidence that some of
those observations are due to increased
circulation of the lenses on the water
.surface resulting from what are called
Vthermocapillary" effects (surface tension
induced motion of lenses on the surface
arising from thermal gradients in the
solution, ref. 11). This explanation is
only tentative and remains to be verified.
Experiments thus far indicate that the
behavior of the oil vis-a-vis the volatile
is of considerable significance in select-
ing a floating immiscible liquid as a
control barrier in VOC emissions work.
REFERENCES
1. Thibodeaux, L. J. and C. Springer,
"Emission of Hazardous Chemicals
From Surface Impoundments to Air",
Draft Final Report to USEPA (MERL),
Cincinnati, OH. (1983).
2. Cox, R. D., "Evaluation of VOC Emission
From Waste Water Systems (Secondary)
Emissions) Field Sampling Test Plan",
EPA Contract No. 68-033038, Radian
Corporation, Austin, TX
.(March 1, 1982).
151
-------
3. Thibodeaux, L. J., C. Springer, P.
Lunney, S. James and T. Shen, "Air
Emission Monitoring of Hazardous Waste
Sites", Proceedings of 3rd National
Conference of Management of Uncontroll-
ed Hazardous Waste Sites", November 29-
December 1, 1982, Washington, D. C.
4. Springer, C., L. J. Thibodeaux, P. D.
Lunney, R. S. Parker and S. C, James,
"Secondary Emissions from Hazardous
Waste Disposal Lagoons: Field Measure-
ments", Proceedings of the Ninth Annual
Research Symposium on Solid and
Hazardous Wastes, EPA-600/9-83-018,
pp. 58-69 (1983.
5. MacKay, D., "Environmental and Labora-
tory Rates of Volatilization of Toxic
Chemicals from Water", Hazard Assess-
ment of Chemicals, Current Develop-
ments, Vol. 1 (19811
6. MacKay, D. and R. S. Matsugo, "Evapor-
ation Rates of Liquid Hydrocarbon
Spills on Land and Water", Canadian
Journal of Chemical Engineering, 51,
434 (1973).
7. Lunney, P. D., "Characterization of
Wind and Depth Effects Upon Liquid
Phase Mass Transfer Coefficients:
Simulation Studies", Master's Thesis,
University of Arkansas (1983)
8. Springer, C. L. J. Thibodeaux and T.
Hedden, "Thermal Effects on the
Emission of Volatile Organic Chemicals
from Surface Waste Water Impoundments
in the Absence of Wind: Laboratory
Measurements", Presented at the 77th
APCA Annual Meeting at San Francisco,
June 24-29 (1984).
9. Hedden, T. E., "Volatile Organic
Chemical Emissions from Waste Water
Impoundments Under No-Wind Conditions",
Master's Thesis, University of
Arkansas, August (1982)
10. Nguyen, T., "Investigation of Floating
Immiscible Liquid as a Control Method
for Volatile Organic Chemical Evapor-
ation from Surface Impoundments",
Master's Thesis, University of Arkansas
(1984).
11. Levich, V. G., "Physiocochemical
Hydrodynamics", Prentice-Hall, Inc.
New Jersey (l°f?), Danes ?°/!-394.
ACKNOWLEDGEMENT
The authors wish to express their
appreciation to the U. S. Environmental
Protection Agency and to the Municipal
Environmental Research Laboratory for the
financial support provided to this project
under Cooperative Agreement
No. CR810856-01-0.
We also wish to thank our project
officer, Mr. Stephen C. James, for his
continuing interest as well as for his
stimulating ideas and encouragement.
152
-------
LEACHING POTENTIAL OF 2,3,7,8-TCDD IN CONTAMINATED SOILS
Danny R. Jackson, *Mike H. Roulier, Henry M. Grotta
Steve W. Rust, J. Scott Warner
Mickey F. Arthur, and Fred L. DeRoos
Battelle Columbus Laboratories *USEPA Hazardous Waste Engineering
Columbus, OH 43201 Research Laboratory
Cincinnati, OH 45268
ABSTRACT
Objectives
Ten soils from sites contaminated with 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD)
were investigated to determine whether TCDD could move in water percolating through
soil and to determine factors controlling potential migration. These objectives were
accomplished by using batch extractions of disturbed soil samples with water and by
passing water through intact core samples. Soils and extracts were analyzed for TCDD.
TCDD in Soils and Leachates
TCDD content in soil (U.S. EPA Region VII protocol) ranged from 8 ppb to
26,300 ppb (ng/g); only one of the ten soils contained more than 2,440 ppb. For the
batch extractions, TCDD concentrations in solution ranged from 0.1 ppt to 56 ppt
(ng/L); 79% of the values were 2.0 ppt or less. For the intact cores, TCDD concentra-
tions in leachate ranged from 0.1 to 14 ppt; 81% of the values were 2.0 ppt or less.
Partition Coefficients
Values for soil partition coefficients (Kp and Koc) were large, indicating that
TCDD will be only very slightly soluble and that only small amounts will be released
into water in contact with the contaminated soils used in this study. Kp ranged from
0.3 x 105 ml/g to 130 x 105 ml/g with a mean of 18 x 105 ml/g. Mean log Koc for the
two batch methods and the intact cores were 7.39, 7.58, and 7.55; these values were
within one standard deviation of a value of 6.95 predicted from solubility data.
Effect of Co-Contaminants
TCDD concentrations in soil and in water extracts were regressed against soil
physical and chemical properties to determine which, if any, were related to TCDD
solubility and could be used to guide further research. Strong correlation was not
obtained between clay and total organic matter content and TCDD solubility. Instead,
both the amount of TCDD in the soil and the amount of TCDD released in a water extract
were strongly related to the total solvent-extractable organic content in soil and the
halogenated semi-volatiles in this solvent-extractable fraction. The solvent-
extractable content of background (uncontaminated) soils was much lower than the
contaminated soils. These results suggest that co-contaminants were in the wastes
along with the TCDD and that these materials have a major role in controlling the
solubility and movement of TCDD in contaminated soils.
153
-------
Predictive Modeling
A solute transport model was used to predict TCDD movement rates in water perco-
lating through soil. Using worst-case conditions (continuous saturated water flow in
the soil and hydraulic conductivity of 10-4 cm/sec) and mean soil physical properties
from the intact cores, predicted movement rates ranged from 4 yr/cm to 5367 yr/cm
depending on the value of Kp and the relative TCDD concentration (C/C0) that was used.
These predicted worst-case rates confirm that TCDD is essentially immobile in water
percolating through contaminated soils used in this study. The relationship between
the levels of co-contaminants and TCDD solubility in soils suggests that TCDD may be
more soluble and mobile in soils containing higher levels of organic co-contaminants
than those in this study.
INTRODUCTION
The environmental persistence and
toxicity of 2,3,7,8-Tetrachlorodibenzo-p-
dioxin (TCDD) has created concern on a
national level for the cleanup and safe
disposal of contaminated soils. Several
investigations using clean soils spiked
with TCDD have documented the extreme
immobility of TCDD in soils (Kearney
et al., 1973; Bartleson, et al., 1975;
Bolton, 1978). However, there have been
no investigations of the solubility or
mobility of TCDD in field soils con-
taminated with other wastes in addition
to TCDD. This study was undertaken to
assess the potential mobility of TCDD in
such soils.
APPROACH
The dissolution of TCDD from soil
and potential mobility of TCDD in water
percolating through soil was determined
either by mixing water with contaminated
soils in batch studies or by passing
water through intact core samples of con-
taminated soils. The concentration of
TCDD in the soil and in the aqueous
extract (soil leachate) was measured;
these values were regressed against soil
physical and chemical properties to
identify mechanisms controlling TCDD
solubility in the soils.
Bulk soil samples were collected by
excavating soil to a depth of approxi-
mately 15 cm. These samples were air-
dried and sieved to pass a 2 mm opening.
Intact soil cores were collected from the
same depths and locations by driving a
10 x 10 cm steel cylinder into the soil.
The soil-filled cylinders were excavated,
placed in plastic bags, and sealed in a
2-quart steel can.
Samples from soils contaminated by
oil-spraying and industrial waste leakage
were obtained from two states. Eight soil
samples were collected from two horse
arenas, a residential site, and three
different roadways in Missouri. One soil
sample was obtained from each of two
locations in New Jersey, a herbicide
manufacturing plant and a scrap metal
yard. This group of soil samples
exhibited a wide range of physical and
chemical characteristics as shown in Table
1. This wide range of characteristics
facilitated interpretation of the TCDD
solubility data from the contaminated
soils and identification of mechanisms
controlling this solubility.
Analytical Methods
Soil pH was determined in a paste
using a soil to water ratio of 1:1
(McLean, 1980). Electrical conductivity
(EC) was measured using a suspension of
soil and water mixed in a ratio of 1:2
(Watson, 1978). Total organic carbon in
soil was determined by oxidation with
KzCrzQ-? and colorimetry (Watson, 1978).
Total organic matter in soil was esti-
mated by multiplying total organic carbon
by 1.742 (Allison, 1965). Cation
exchange capacity (CEC) was determined by
saturating the soil with NH4+ followed by
displacement of the NH4+ with K+. Analy-
sis of displaced NH4+ was conducted using
an ammonium ion-specific electrode (Allen
et al., 1974). The specific surface area
of each soil was determined by the ethyl-
ene glycol monoethyl ether (EGME) method
(Cihacek and Bremmer, 1979). Soil parti-
cle size was determined by the hydrometer
method (ASTM, 1972).
Total Solvent Extractable Content
Total solvent extractable content
(TSEC) of soils was determined by
154
-------
TABLE 1. PHYSICAL AND CHEMICAL PROPERTIES OF SOILS
Ln
Ln
No.
1
2
3
4
5
6
7
8
9
10
Location
New Jersey
Missouri
Missouri
New Jersey
Missouri
Missouri
Missouri
Missouri
Missouri
Missouri
Site
Industrial-2
Shenandoah Stables
(Slough Area)
Piazza Road-1
Industrial-1
Minker-2
Minker-1
Bubbling Springs Stable
Piazza Road-2
Sontag Road
Times Beach
Total*
2,3,7,8-TCDO
Concentration
ng/g
26,316
2,438
1.095
1,083
762
608
275
234
63
8
Organic
Matter (mq/q)
TSECb
21.5
7.1
5.2
19.4
5.3
7.0
3.4
0.7
5.2
0.3
Peak
Area0
274.0
14.8
1.5
20.9
3.5
3.7
2.2
0.6
0.7
0.4
Total
80
60
27
37
25
22
15
76
60
17
Surface
Area
m2/g
22.3
35.7
27.6
38.9
25.3
25.6
15.0
21.5
34.6
19.7
CECd
meq/100 g
12.7
39.0
35.2
17.8
24.4
19.5
12.8
53.8
36.9
23.2
PH
7.5
7.3
6.7
6.7
7.7
8.0
7.6
7.5
8.4
7.7
EC
pmhos/cm
65.0
195.2
95.0
98.8
139.3
119.2
101.5
153.1
350.0
107.1
Bulk
Density
g/cm3
1.27
1.30
1.34
—
1.29
—
1.48
1.01
1.14
1.22
Sand
81.1
42.3
57.4
72.1
33.9
55.9
61.0
60.5
38.5
52.4
Silt
«£
17.9
51.6
30.1
22.8
58.8
37.3
36.7
30.5
58.4
43.2
Clay
1.0
6.1
12.5
5.0
7.3
6.7
2.3
9.0
3.1
4.4
Texture
Loamy Sand
Silty Loam
Sandy Loam
Sandy Loam
Silty Loam
Sandy Loam
Sandy Loam
Sandy Loam
Silty Loam
Sandy Loam
a Determined using modified EPA Region VII protocol (EPA, 1983).
b Total Solvent Extractable Content.
c Total peak area of solvent extractable components detected by GC-ECD analysis (unitless).
d Cation Exchange Capacity.
-------
extracting a suspension of 30 g of soil
in 30 ml of an aqueous solution contain-
ing 30% NaCl and 2% KH2P04 with 30 mL of
methyl tert-butyl ether (MTBE) for 18
hours in a tumbling apparatus (Warner,
1978). The MTBE phase was separated from
the aqueous phase, dried with 2 g of
anhydrous MgS04, and filtered through a
0.5 uM PTFE membrane filter. A 100 yL
aliquot of the filtered extract was
transferred to a tared aluminum weighing
dish, evaporated to dryness under a heat
lamp, and reweighed using a microbalance.
This weight is reported as TSEC in
Table 1. The method is suitable for
extraction of most phenols, anilines, and
neutral semivolatile organic compounds.
Electron Capture Response of Solvent
Extract
The MTBE extract of soil was
analyzed by gas chromatography (GC) using
a fused silica capillary column and a
63Ni electron capture detector (ECD).
The combined GC-ECD peak area of all
chromatographic peaks eluting after
5 minutes, listed as Peak Area in
Table 1, was obtained using a computer-
ized integration and data processing sys-
tem. Since the electron capture detector
detects primarily halogenated compounds,
the combined peak area is an indicator of
total amount of halogenated organic
compounds present.
Background (uncontaminated) soil
samples were collected near the locations
of each of the contaminated soils. TSEC
was determined for each of these
background samples.
Soil Extraction Methods
The analytical method for TCOD in a
soil leachate required a sample volume of
1 liter. Two batch methods and a method
for leaching intact cores were used to
generate soil leachate (aqueous extract)
to accommodate this requirement.
The Solid Waste Extraction
Procedure-Release (SWLP-R) was used to
determine the effect of sequential
extraction on the solubility of TCDD from
soil (Garrett, et al., 1984). This
method is similar to the Standard Leach-
ing Test (SLT) Procedure R developed by
Ham et al. (1979), wherein the same soil
sample is re-extracted with fresh water.
For this procedure, a 100 g aliquot of
soil was placed in a 2.7-L glass bottle
with 1 L of distilled water. The bottle,
with the soil-water suspension was con-
tinuously tumbled for 18 hours using an
electric-powered rotary extractor set at
30 rpm. The suspensions were pressure-
filtered through hydrophilic cellulose
acetate and nitrate membranes with a pore
size of 0.45 yM. Each soil sample was
extracted a second time by repeating the
above procedure using the same soil and
fresh water. This procedure results in
filtered soil leachate (aqueous extract)
representing two sequential extractions
from the same soil sample. Two samples of
soil from each location were extracted by
this procedure (SWLP-R). The soil
leachates were analyzed for TCDD using a
modified version of the U.S. EPA Region
VII protocol (U.S. EPA, 1983).
A second version of the SWLP
procedure was used, which was similar to
the SLT-Procedure C (Ham, et al., 1979),
wherein three fresh soil samples are
extracted sequentially with the same
water. The soil leachate is filtered
after extracting the soil sample and then
used to extract the next soil sample.
This procedure is designated SWLP-C where
C denotes concentration; this is done to
determine the effect of a lower liquid to
solid ratio on TCDD solubility in soil.
Two liters of distilled water were tumbled
for 18 hours with 200 g of contaminated
soil. The soil leachate was pressure-
filtered through a 0.45 uM membrane, and
again tumbled for 18 hours with 200 g of
fresh soil, followed by filtration. This
procedure was repeated with a third fresh
200-g portion of soil. The final volume
of soil leachate from the third filtra-
tion, because of retention by the three
soil samples, ranged from 1430 mL to
1640 mL depending on the texture of the
soil. The overall liquid to solid ratio
for this method was approximately 2.5:1.
Two sets of three soil samples from each
location were extracted by this procedure
(SWLP-C). The final soil leachate was
analyzed for TCDO by the modified U.S.
EPA Region VII protocol.
Intact soil cores were leached with
distilled water for comparison with
156
-------
results from the two SWLP procedures and
to study the effect of soil structure and
secondary porosity on soil-liquid contact
and TCDD leaching. One liter of dis-
tilled water was allowed to percolate
through the soil cores. The soil leach-
ate was filtered through a 0.45 uM mem-
brane. A second 1-L aliquot of distilled
water was percolated through each soil
core by repeating the above procedure.
Each 1-L aliquot, representing a liquid
to solid ratio of approximately 1:1, was
analyzed for TCDD.
Partition Coefficients
Partition coefficients (Kp's)
measure the relative distribution of a
substance between the soil solid phase
and the associated aquequs phase. These
coefficients are a key parameter in the
predictive modeling of TCDD movement in
water percolating through soil. Large
partition coefficients for compounds
indicate high partitioning onto the solid
phase and consequently very slow rates of
movement in soil.
Two partition coefficients (Kp and
KOC)» defined below, were calculated for
the contaminated soils from results of
each of the batch methods and the intact
core studies.
= (nq TCDD per g of soil x 103)
ng TCDD per L of extract)
NDC
r
organic carbon per g of soil)
(1)
(2)
When calculated in this manner, Kp
corresponds to the constant in the
Freundlich equation with the exponent
equal to 1.0 (linear isotherm).
In previous studies a considerable
part of the variability in Kp between
different soils was strongly correlated
with soil organic carbon content (Hassett
et al., 1983). As a result of this
strong correlation, partition coeffi-
cients are often normalized on the basis
of organic carbon content (equation 2).
Normalized partition coefficients (Koc)
were shown to have a strong linear rela-
tionship with the octanol-water partition
coefficients (Kow) using several non-
polar organic compounds (Hassett et al.,
1983; Karickhoff et al., 1979). This
relationship supports the theory that
hydrophobic sorption on soil organic
matter is a major process controlling
retention and movement of compounds such
as TCDD.
Solute Transport Model
Sufficient data on rainfall, evapo-
transpiration, and soil hydrau'lic proper-
ties were not available to make site-
specific predictions of TCDD movement
rates in water percolating through soil.
However, using average values of hydraulic
conductivity, density, and porosity from
the soil cores and making assumptions to
supply missing data, predictions of TCDD
movement rates were made to illustrate the
general magnitude of these rates and the
factors that most influence them.
An analytical solution of a solute
transport equation was used to predict
TCDD movement rates. The governing
equation for one-dimensional convective-
dispersive solute transport with no
production or decay is:
R(aC/3t) = D(32C/3X2) - v(3C/3X) <3)
where: R = the retardation factor
(l+(pb/n)Kp)
DO is the bulk density
(g/cm3)
n is the porosity
(cm3/cm3)
Kp is the partition
coefficient (mL/g)
X = the distance from the soil
surface (cm)
C = concentration in solution
(ng/L)
D = the dispersion coefficient
(cm2/day)
v = the pore water velocity
(cm/day).
t = time (day)
The initial condition was zero solute
(TCDD) concentration in the soil, i.e.,
C(X,0) = 0. Boundary conditions were
constant flux at the soil surface and no
change in concentration at infinite depth
(semi-infinite system).
157
-------
[-D(3C/3X) + vC] X=0 = vC
0
(4)
Constant flux was selected as the bound-
ary condition at the soil surface instead
of the more common constant concentration
boundary condition because van Genuchten
and Alves (1982) and van Genuchten and
Parker (1984) have shown that use of the
constant concentration boundary condition
leads to mass balance errors. Values for
hydraulic conductivity (10*3 and 10-4
cm/s), bulk density (1.25 g/cm^), and
porosity (0.53 cm^/cm^) from the leaching
behavior and physical measurements on the
intact cores were used. A hydraulic
gradient of 1.0 was used to represent
saturated soil with no appreciable
ponding of water on the surface. The
dispersion coefficient was calculated
from the relationship presented by Biggar
and Nielsen (1976).
The retardation equation was used as
a second means of predicting TCDD move-
ment under saturated flow conditions
(Freeze and Cherry, 1979). This equation
is:
where:
Vc = v/R
Vc = contaminant velocity
v = pore water velocity
R = retardation factor.
(5)
The retardation equation is a convenient
tool for estimating solute movement
rates. This equation is limited to pre-
dicting the velocity of the maximum sol-
ute concentration for a slug input or the
velocity of the 0.5 relative concentra-
tion (C/C0) for a continuous input.
However, relative concentrations less
than 0.5 are often of interest. For
example, a C/C0 of 0.01 would correspond
to TCDD solution concentrations ranging
from 5.5 x 10~1 ng/L to 1 x 10~3 ng/L in
leachates from soils in this study.
These concentrations are substantially
above suggested water quality criteria
for aquatic life and human health (U.S.
EPA, 1984). The analytical solution was
used to allow prediction of velocities
for other values of C/C0.
Calculations for the analytical
solution were performed using a Tektronix
4052A computer and a BASIC language pro-
gram adapted from the FORTRAN program of
van Genuchten and Alves (1982). Minimum,
maximum, and mean Kp values from the
average of the intact core leaching and
SWLP-C methods were used in calculations
for both equations.
RESULTS AND DISCUSSION
TCDD in Soil Lea'chate
TCDD concentrations in soil leachates
from the three leaching methods are
summarized in Table 2. For the batch
extractions, TCDD concentrations in
solution ranged from 0.1 ppt to 56 ppt
(ng/L); 79% of the values were 2.0 ppt or
less. For the intact cores, TCDD concen-
trations in solution ranged from 0.1 ppt
to 14 ppt; 81% of the values were 2.0 ppt
or less.
The mean values of 1.0 and 1.3 ng/L
for the SWLP-C and intact core procedures,
respectively, (Table 2) are within one
standard deviation of the mean value for
the first extract of SWLP-R. The former
two methods use a liquid to solid ratio of
less than 2.5:1. Both the intact core
procedure and SWLP-C are relatively
complex and time-consuming. Procedure
SWLP-R with one extraction often gives
equivalent results with far less effort.
Partition Coefficients
Kp and log (Kp) values computed for
each soil and extraction procedure are
listed in Table 3. Mean Kp and mean log
(Kp) values for the first extract of the
SWLP-R, the SWLP-C, and the Intact Core
leaching procedure are very similar. The
partition coefficients for the second
extract of the SWLP-R are lower than
values obtained by the other procedures
due to higher TCDD concentrations in soil
leachate. These results indicate that the
first extract of the SWLP-R, the SWLP-C,
and the Intact Core procedures are
essentially equivalent in their rep-
resentation of the partitioning of TCDD in
these soils.
158
-------
TABLE 2. 2,3,7,8-TCDD CONCENTRATIONS IN SOIL LEACHATES FROM SWLP-R, SWLP-C, AND THE INTACT CORE PROCEDURES
Number
SWLP-R
SWLP-C
Extract
la 2a
Mean
Mean lb
Intact Core
Extract
2b
Mean
nnll
1
2
3
4
5
6
7
8
9
10
Mean
5.8
9.0
0.4r
0.9
0.7
0.2
0.5C
0.1
0.2
0.2C
1.8
(0.75)
(0.07)
(0.63)
(0.28)
(0.46)
(0.09)
(0.01)
(0.21)
(0.02)
(1.08)
55.3
15.1
0.3
1.2
1.5
3.7
2.1
O.ic
O.ic
0.2
8.0
(0.03)
(0.10)
(0.29)
(0.24)
(0.18)
(0.16)
(0.12)
(0.06)
(0.18)
(0.75)
30.5
12.1
0.3
1.0
1.1
2.0
1.3
0.1
0.2
0.2
4.9
3.1 (0.61)
1.8
0.5
0.2
0.9
0.9
0.1
0.9
0.2
1.0
(0.15)
(0.61)
(0.02)
(0.09)
NO
(0.13)
(0.25)
(0.07)
(0.41)
2.0 (0
5.8 (1
0.8C (1
NDd
ND
1.8 (0
0.2C (0
O.lc (0
0.6 (0
O.lc (0
1.4
.23)
.18)
.25)
.56)
.25)
.71)
.61)
.48)
2.3 (0.27)
4.4 (0.98)
0.4C (0.39)
ND
ND
1.8C (0.32)
0.4C (0.97)
O.ic (0.52)
0.4C (0.93)
O.lc • (0.44)
1.2
2.1
5.1
0.7
ND
NO
1.8
0.3
0.1
0.5
0.1
1.3
9 Values reported are the means of 2 measurements. Values in parentheses are the Relative Standard deviations of the
measurements.
b Values reported are the means of 3 measurements. Values 1n parentheses are the Relative Standard deviations of the 3
measurements.
c Mean includes one or more values not meeting the detection limit criteria.
d Not determined.
-------
TABLE 3. SOIL PARTITION COEFFICIENTS (Kp) FOR SOIL LEACHATES FROM THE SWLP-R, SWLP-C, AND THE INTACT CORE PROCEDURES
SWLP-R
Extract
Site
Number
1
2
3
4
5
6
7
8
9
10
Mean
Std. Dev.
1
Kp
10-5mL/g
45.8
2.7
30.9
12.2
11.0
27.0
5.9
23.2
2.6
0.4
16.2
Log (Kp)
6.66
5.43
6.49
6.09
6.04
6.43
5.77
6.37
5.41
4.40
5.91
0.68
2
Kp
ID"5 mL/g
4.8
1.6
43.3
9.1
5.1
1.6
1.3
20.9
7.3
0.4
9.5
Log (Kp)
5.68
5.20
6.64
5.96
5.71
5.20
5.11
6.32
5.86
4.60
5.63
0.61
Mean
Kp
10-5 mL/g
8.6
2.0
36.1
10.4
6.9
3.1
2.2
22.0
3.8
0.4
9.6
Meanb
Log (Kp)
5.94
5.31
6.56
6.02
5.84
5.49
5.34
6.34
5.58
4.59
5.70&
0.57
SWLP
Kp
10-5 mL/g
83.9
13.3
20.6
47.1
8.2
--
3.0
25.7
0.7
0.3
22.5
Log (Kp)
6.92
6.12
6.31
6.67
5.92
--
5.47
6.41
4.85
4.53
5.91
0.81
Intact
Core3
Kp
lO-5 mL/g
123.9
4.8
16.9
--
--
53.8
9.4
20.3
1.3
1.1
22.7
Log (Kp)
7.09
5.68
6.23
--
--
5.58
5.97
6.31
5.11
5.05
5.88
0.68
a Mean of two consecutive leachate fractions.
b Means are calculated from the average leachate concentrations and not the arithmetic means of the Kp values.
-------
Values of log (Koc) were computed
for each soil and compared to a predicted
value provided by Dragun (1984). This
predicted value was calculated using
relationships between aqueous solubility,
Koc, and an octanol-water partition coef-
ficient (Kow). The mean log (Koc) values
for the SWLP-R, SWLP-C, and the Intact
Core procedure were 7.39, 7.58, and 7.55;
these were within one standard deviation
of the value of 6.95 calculated by
Dragun.
The partition coefficients in
Table 3 are based on data from
studies where clean water is brought into
contact with contaminated soil and some
of the TCDO dissolves/partitions into the
water. Partition coefficients in clean
soils may be different. In the field,
when water moves out e-f contaminated soil
and into clean soil, the partitioning
process reverses and some of the TCDD in
the water will adsorb/partition onto the
clean soil while some remains in solu-
tion. As discussed later in this manu-
script, we believe the data indicate that
other organic compounds present in con-
taminated soils have a strong influence
on the partitioning of TCDD between soil
and water. Thus, as far as partitioning
is concerned, the clean soil may be a
different system than the contaminated
soil. Because partition coefficients
calculated in this study with desorption
data from contaminated soils were within
the range of values predicted from
physical/chemical properties of TCDD, it
is reasonable, in the absence of better
data, to assume that partition
coefficients calculated with data from
adsorption of TCDD from extracts of these
contaminated soils onto clean soil will
not be significantly different than the
values reported here. At other
locations, where contaminated soils
contain greater amounts of organic co-
contaminants, there is a greater chance
that the partition coefficients for clean
soils may be different than for the
contaminated soils.
Statistical Analyses
A correlation matrix presented in
Table 4 was constructed to investigate
the relationships between selected soil
properties, partition coefficients, and
TCDD concentrations in soil and leach-
ates. The correlation matrix is based
only on data from the Missouri soils. The
two New Jersey soils were excluded from
this analysis because the TSEC and Peak
Area were significantly greater than for
any of the Missouri soils (see Table 1);
inclusion in the analysis would have
biased the results by providing a corre-
lation between two sets of data rather
than within a single data set. Addi-
tionally, the sources of TCDD, organic co-
contaminants, and method of contamination
were different in New Jersey and in
Missouri. Data from the New Jersey soils
generally fit the relationships shown in
Figure 1. When additional data are
available from soils having TSEC and Peak
Area values between those for the Missouri
and New Jersey soils, the entire data set
can then be reanalyzed.
Observed significance levels (OSLs)
which indicate the significance of the
reported correlation coefficients are
provided in parentheses in Table 4. For
example, an OSL value of 0.05 indicates
that the corresponding correlation coef-
ficient is significantly different from
zero at a confidence level of 95% and an
OSL value of 0.00 indicates a confidence
level of 99.9+%.
The log soil concentration of TCDD is
positively correlated with the leachate
concentrations for all three extraction
procedures. However, the leachate
concentrations were more highly correlated
with the log (Peak Area) measurement. The
relationships between log (Peak Area) and
the soil and leachate TCDD concentrations
are illustrated in Figure 1. The least
squares regression line has been
superimposed on each plot. While log
(Peak Area) is not well defined as a soil
property, it nonetheless appears to be a
most important determinant for TCDD
solubility in the Missouri soils.
The relationship between soil organic
matter content and the Kp values for the
three leaching procedures was investigated
to determine the significance of computing
Koc values for the soils. Correlation
coefficients for organic carbon vs. Kp
fromthe three leaching procedures were
insignificant (SWLP-R = 0.36; SWLP-C =
0.35; intact
161
-------
TABLE 4. CORRELATION MATRIX FOR SOIL PROPERTY VARIABLES VERSUS BULK SOIL AND LEACHATE VARIABLES WITH
OBSERVED SIGNIFICANCE LEVELS IN PARENTHESES
Log [2,3,7,8-TCDD]: Percent Organic Percent
Dependent Variable Bulk Soil Log (Peak Area) Log (TSEC) Carbon Clay
Log [2,3,7,8-TCDD]: 0.84 (0.01) 0.76 (0.03) 0.14 (0.74) 0.44 (0.30)
Bulk Soil
Log [2,3,7,8-TCDD]: 0.66 (0.08) 0.95 (0.00) 0.61 (0.11) -0.16 (0.71) -0.08 (0.80)
SWLP-R
Log [2,3,7,8-TCDD]: 0.80 (0.03) 0.92 (0.00) 0.87 (0.01 0.11 (0.81) 0.15 (0.78)
Intact Core
Log [2,3,7,8-TCDD]: 0.49 (0.27) 0.77 (0.04) 0.84 (0.02) -0.21 (0.65) -0.32 (0.45)
SWLP-C
-------
core = 0.17). Thus, computing Koc values
does not account for a significant amount
of the variability associated with corre-
sponding Kp values. However, log Koc
values were calculated for comparison
with the value predicted from solubility
data and the relation between solubility
and Koc.
Based on published information about
soil partitioning of compounds similar to
TCDD in clean soils, it was expected that
clay content and total organic carbon
content of soils would correlate strongly
with the fraction of total TCDD solubil-
ized by the water extract and that TCDD
would be less soluble in soils containing
greater amounts of clay and total organic
carbon. However, there was not a clear
relationship between either of these
parameters and TCDD solubility from the
contaminated soils used in this study.
Instead, both the amount of TCDD in a
soil and amount of TCDD released in a
water extract were strongly related to
the total solvent-extractable organic
content (TSEC) in soil and to the
halogenated semi-volatiles (peak area) in
this solvent-extractable material.
' In soils with higher TSEC and peak
area, the TCDD concentrations in water
extracts were greater. This occurred in
spite of the fact that TSEC represented
only a portion of the total soil organic
matter. In the contaminated soil, TSEC
ranged from 0.3 mg/g to 21.5 mg/g; total
organic matter in the same soils ranged
from 15 mg/g to 80 mg/g. Because TSEC
for background (uncontaminated) soils was
so low (0.51 to 1.54 mg/g), it is con-
cluded that both TSEC and peak area mea-
sure materials that were in the wastes
along with TCDD. This suggests that
other organic contaminants in wastes have
a major role in controlling the solubil-
ity and movement of TCDD in contaminated
soils and that TCDD behaves differently
in contaminated soils than would be
predicted from studies of clean soils
spiked only with TCDD.
Solute Transport Modeling
The retardation equation and the
analytical solution predict the movement
of TCDD from a layer of contaminated soil
into an underlying layer of uncontami-
nated soil. The analytical solution
assumes that TCDD does not reach the lower
boundary of the uncontaminated soil during
the time period being modeled (semi-
infinite system). All depths are measured
downward from the boundary between
contaminated and uncontaminated soil.
The two models were used first to
predict, for a single depth and a single
hydraulic conductivity (10-4 cm/sec), the
time for relative concentrations (C/C0) to
reach specified values. C0 is the
concentration of TCDD in water leaving the
contaminated soil and C is the con-
centration at 30 cm depth in the uncon-
taminated soil at a particular time. For
example, C/C0 = 0.5 would mean 1.55 ng/L
if the SWLP-C data for site No. 1 was
being used (3.1 ng/L, Table 2). Longer
times are required for C/C0 to reach the
higher values (Table 5) because dispersion
at the contaminant front advances the
lower relative concentrations more
rapidly. Differences in movement rates
between various C/C0 are relatively minor
compared to the direct linear effect of
changes in Kp. For example, when Kp
increases from 0.3 x 105 to 133.6 x 105 (a
relative increase of 445) the predicted
times increase by factors ranging from 444
to 447.
Next, the models were used to predict
changes in concentration with depth for
fixed times and a single C0. Figure 2
shows some of these results using
hydraulic conductivities of 10~3 and 10"^
cm/sec, 1.2 ng/L of TCDD in the water (C0)
coming from the contaminated soil, and the
mean Kp for all soils. Even for the
higher hydraulic conductivity, thousands
of years are required for TCDD to move to
appreciable depths.
These predictions represent worst-
case conditions - continuous saturated
water flow instead of the intermittent
flows (both in time and direction)
observed in field soils. Even under these
conditions the movement rates are so slow
that TCDD is essentially immobile in water
percolating through soil. Other transport
mechanisms such as wind and water erosion
of contaminated soils are likely to be
significantly more important for these
soils.
163
-------
3.5
3.0
1
2.S
T 2.0
£t 1.5
•1.0
0.5
A. Bulk Soil
I I I
I
I
•0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2
Log(Ptak Area)
1.5
1.0
1
2 -0.5
f
•1.0
•1.5
C SWLP-C
•0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2
Log(Peak Area)
1.5
1.0
1
0.5
£1 -0.5
f
•1.0
-1.5
I. SWLP-R
I I I
J I
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2
Log(Ptak Ana)
1.5
1.0
1
to,
0.5
2. -05
I
•1.0
-1.5
D. Intact Core
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2
U>g(Pcak Area)
FIGURE 1. LOG-LOG PLOTS OF 2,3,7,8-TCDD CONCENTRATIONS IN (A) BULK SOIL,
(B) SWLP-R, (C) SWLP-C, AND (D) INTACT CORE LEACHATE VERSUS
PEAK AREA
164
-------
0.1
0.2
Hydraulic Conductivity
1x10
1 x
-3 -
0.5
Soil
Depth
(cm)
10
20
50
100
_— — -"""~"^ 3000
I
I
I
Yrs
I
I
O.OT 0.02 0.05 0.1 0.2 0.5 1.0
[2,3,7,8 - TCDD], ng/L
FIGURE 2. PREDICTED TRANSPORT OF 2,3,7,8-TCDD IN SOIL WHERE
Kp = 22.6 x 105 mL/g AND CQ = 1.2 ng/L.
165
-------
TABLE 5. TIME (YRS) FOR 2,3,7,8-TCDD RELATIVE
CONCENTRATION (C/C0) TO REACH THE
INDICATED VALUE AT 30 CM SOIL DEPTH.
Equation
C/CC
Kp x 10-5.
133
22.6
0.3
Retardation 0.5
1.59 x 105 2.69 x 10* 3.57 x 10*
Analytical
0.
0.
0.
5
1
01
1.
8.
5.
61 x
66 x
39 x
105
104
10*
2
1
9
.72
.46
.05
x
x
X
104
104
103
3.61
1.95
1.21
x 102
x 102
x 102
As discussed above, Kp for TCDD in
contaminated water moving in clean soil
may be different than the Kp used in
these predictions. However, it is
unlikely that they will be significantly
less or that movement rates will be
significantly greater unless there are
large amounts of other organic
contaminants present in the soil with the
TCDD and the TSEC and Peak Area are much
greater than the values measured for
these soils.
CONCLUSIONS
Soil properties such as clay content
and organic matter content were not sig-
nificantly correlated with TCDD soil
partition coefficients. Previous inves-
tigations using clean soils spiked with
TCDD and similar nonpolar compounds indi-
cated that a strong correlation should be
expected.
A strong association was found
between TCDD concentration in soil
leachates (aqueous extracts) and solvent-
extractable halogenated semivolatile
organic matter content of soils as mea-
sured by TSEC and Peak Area. Because the
TSEC and Peak Area of background soils
were much lower than for the contaminated
soils, it is concluded that these param-
eters represent organic materials that
were applied to soil in wastes along with
the TCDD. The strong association between
soluble TCDD and solvent-extractable
material suggests that TCDD dissolution
from contaminated soils is regulated by
these halogenated semivolatiles.
The three extraction methods used for
determining TCDD partitioning coefficients
provided essentially identical results
when only the first extract from the SWLP-
R procedure was compared with results from
the SWLP-C and the Intact Core procedure.
The second extract in the SWLP-R procedure
resulted in greater TCDD solution
concentrations in most cases. These
increased TCDD concentrations were
ascribed to the effect of soil elec-
trolytes on the activity coefficients and
solubility. Based on these results, a
single extract using the SWLP-R procedure
is the most cost-effective means for
determining TCDD partitioning
coefficients.
When the TCDD partition coefficients
are used with two solute transport models,
the predicted worst-case TCDD movement
rates in water percolating through soil
are so slow that other transport
mechanisms such as wind and water erosion
are likely to be more significant.
Partition coefficients and TCDD movement
rates may be different in soils that have
greater amounts of halogenated semi-
volatile organic compounds than the soils
used in this study.
REFERENCES
1. Allison, L.E, 1965. Organic Carbon.
In: Methods of soil analysis: chem-
ical and microbiological properties.
Agronomy Monograph No. 9, Part 2,
Chapter 90. American Society of
Agronomy, Madison, WI 53711.
166
-------
2. Bartleson, F.D., Jr., D. D.
Harrison, and J. D. Morgan, 1975.
Field studies of wildlife exposed to
TCDD contaminated soils. U.S. Air
Force Armament Laboratory, Eglin Air
Force Base, FL, AFATL-TR-75-49.
58 p.
3. Biggar, J. W., and D. R. Nielsen,
1976. Spatial variability of the
leaching characteristics of a field
soil. Water Resour. Res. 12(1):78-
84.
4. Bolton, L, 1978. Seveso dioxin: No
solution in sight. Chem. Engineer.
85(22):78.
5. Cihacek, L. J., and J. M. Bremner,
1979. A simplified ethylene glycol
monoethyl ether procedure for
assessment of soil surface area.
Soil Sci. Soc. Amer. J. 43:821-822.
6. Dragun, J., 1984. Personal communi-
cation. E. C. Jordan Co.,
Southfield, MI.
7. Freeze, R. A., and J. A. Cherry,
1979. Groundwater. Prentice-Hall,
Inc., Englewood Cliffs, NJ 07632.
pp. 403-405.
8. Garrett, B. C., D. R. Jackson, W. E.
Schwartz, and J. S. Warner, 1984.
Solid waste leaching procedure.
Technical resource document for
public comment. U.S. EPA, Office of
Solid Waste and Emergency Response,
SW-924.
9. Ham, R., M. A. Anderson,
R. Stegmann, R. Stanforth, 1979.
Background study on the development
of a standard leaching test. EPA-
600/2-79-109. Industrial Environ-
mental Research Laboratory. U.S.
Environmental Protection A'gency,
Cincinnati, OH 45258.
10. Hansch, C., and A. Leo, 1979.
Substituent constants for correla-
tion analysis in chemistry and
biology. Wiley & Sons, NY.
11. Hassett, J. J., J. C. Means, W. L.
Banwart, S. G. Wood, S. Ali, and A.
Khan, 1980. Sorption of dibenzo-
thiophene by soils and sediments.
J. Environ. Qual. 9(2):184-186.
12. Hassett, J. J., W. L. Banwart, and
R. A. Griffin, 1983. Correlation of
compound properties with sorption
characteristics of nonpolar compounds
by soils and sediments: Concepts and
limitations. Chapter 15. In: The
Environment and Solid Waste
Characterization, Treatment and Dis-
posal - Proceedings of the 4th Oak
Ridge National Laboratory Life
Science Symposium, October 4-8, 1981,
Gatlinburg, Tennessee. Edited by C.
W. Francis and S. I. Auerbach. Ann
Arbor Science Pub.
13. Karickhoff, S. W., D. S. Brown, T. A.
Scott, and A. Trudy, 1979. Sorption
of hydrophobic pollutants on natural
sediments. Water Res. 13(3):241-
248.
14. Kearney, P. C., E. A. Woolson, A. R.
Isensee, and C. S. Helling, 1973.
Tetrachlorodibenzodioxin in the
environment: Sources, fate, and
decontamination. Environ. Health
Pers. 5:273-277.
15. McLean, E. 0., 1980. Recommended pH
and Lime Requirement Tests. Pp. 5-8.
In: Recommended Chemical Soil Test
Procedures for the North Central
Region. North Dakota Agricultural
Experiment Station, NCR Publ. No. 221
(rev.), Fargo NO.
16. U.S. Environmental Protection Agency,
1983. Determination of 2,3,7,8-TCDD
in soil and sediment. September
revision. Region VII Laboratory, 25
Funston Road, Kansas City, KS, 249 p
17. U.S. Environmental Protection Agency,
1984. Ambient water quality criteria
for 2,3,7,8-tetrachlorodibenzo-p-
dioxin. EPA-440/5-84-007, Office of
Water Regulations and Standards,
Washington, D.C. 370 p.
18. van Genuchten, M. Th., and W. J.
Alves, 1982. Analytical solutions of
the one-dimensional convective-
dispersive solute transport equation.
Technical Bulletin Number 1661. USDA
Agricultural Research Service,
Washington, D.C. 20250.
167
-------
19. van Genuchten, M. Th., and J. C.
Parker, 1984. Boundary conditions
for displacement experiments through
short laboratory columns. Soil Sci.
Soc. Amer. J. 48(4):703-708.
20. Warner, J. S., 1978. Chemical char-
acterization of marine samples.
American Petroleum Institute Publ.
4307, 2 p
21. Watson, M. E., 1978. Soil testing
procedures. Ohio Agricultural
Research and Development Center,
Research Extension Analytical
Laboratory, Wooster, OH. 20 p
22. Young, A. L., 1983. Long-term
studies on the persistence and move-
ment of TCDD in a natural ecosystem.
Pp. 173-190. In: Environ. Sci.
Res., 26 (Hum. Environ. Risks
Chlorinated Dioxins Relat. Compd.).
168
-------
DEVELOPMENT OF STANDARDIZED BATCH ADSORPTION
PROCEDURES: EXPERIMENTAL CONSIDERATIONS
W. R. Roy, R. A. Griffin, S. F. J. Chou, C. C. Ainsworth, and I. G. Krapac
Illinois State Geological Survey
Champaign, Illinois 61820
ABSTRACT
A Technical Resource Document (TRD) is being developed by the authors to provide
information on the use of batch procedures for determining the adsorption affinity of
soils for solutes in waste leachates. The TRD will be part of the technical basis for
guidance by the Office ef Solid Waste on design and evaluation of compacted soil liners
for waste storage and disposal facilities. In developing the batch procedures for the
TRD it was found that the method of mixing the solute-soil mixtures influenced their
adsorption by soil; more vigorous shaking resulted in greater adsorption. The National
Bureau of Standards rotating tumbler was adopted as the method of choice. A mixing
interval of 24 hours appears to be a valid operational definition of equilibrium for
these types of batch procedures. Changes in solution concentration of arsenic and cad-
mium exposed to different soil materials were negligible after 10 hours. The selection
of appropriate soil :solution ratios for inorganic solutes must be found empirically, but
suitable ratios for hydrophobic organic solutes can be estimated from the compounds'
solubility in water and the organic carbon content of the soil.
INTRODUCTION
This paper discusses the effect of
method of mixing, equilibration time,
and soil:solution ratio on results from
batch studies of the adsorption capacity
of soils for solutes in waste leach-
ates. This work is part of the develop-
ment of a Technical Resource Document
(TRD) on use of batch procedures in de-
signing compacted soil liners for re-
moval of solutes from waste leachates.
The TRD will be part of the technical
basis for guidance by the Office of
Solid Waste on design and evaluation of
compacted soil liners for waste storage
and disposal facilities.
The capacity of geological materi-
als to attenuate potential pollutants
has been studied by many researchers,
especially during the last 30 years.
The batch adsorption or batch equili-
brium technique has often been used in
laboratory studies to assess the capac-
ity of soils and soil components to
attenuate chemical constituents in
solution. However, the batch procedures
used in these studies vary considerably
in terms of experimental conditions, and
in some cases may yield different re-
<;nlt<; pvpn when the same soils, solutes,
and concentrations are studied.
In principle, the batch adsorption
technique is relatively simple, account-
ing, in part, for its popularity. This
technique consists of mixing a solution
containing aqueous solutes of known com-
position and concentration with a given
mass of adsorbent for a period of
time. The solution is then separated
from the adsorbent and chemically
analyzed to determine changes in solute
concentration. However, a number of
experimental parameters may affect the
adsorption of a given constituent. For
inorganic solutes, these parameters
include time, temperature, method of
mixing, soi1:solution ratio, moisture
content, and the composition and
concentration of other dissolved
constituents in the solution (White,
1966; Barrow and Shaw, 1975, 1979;
169
-------
Helyar et al ., 1976; Hope and Syers,
1976; Griffin and Au, 1977; Barrow,
1978; Ainsworth et al., 1984, and Roy et
al., 1984). For organic solutes,
similar parameters may also affect the
adsorption (Bailey and White, 1970;
Grover and Hance, 1970; Dao and Lavy,
1978; Koskinen and Cheng, 1983; and
Horzempa and DiToro, 1983). In addi-
tion, solution pH, adsorbate volatility,
photodegradation, biodegradation, and
stability can also affect the adsorption
of organic solutes (Harris and Warren,
1964; Chou and Griffin, 1983; and Scott
et al., 1982).
Soil :solution ratios used in batch
procedures have varied from 1:10,000 for
the adsorption of phosphates by gibbsite
(Muljadi et al., 1966) to 1:1 for the
adsorption of nickel by a soil studied
by Bowman et al. (1981). Equilibration
time, another basic experimental para-
meter in batch techniques, has varied
from 30 minutes, for the adsorption of
zinc by soil investigated by Shukla and
Mittal (1979), to 6 days to equilibrate
a phosphate-soil system investigated by
Jones et al., (1979). These experimen-
tal conditions were probably appropriate
for the particular system under study,
but the differences in conditions make
comparisons of data among studies
difficult.
It is apparent from the literature
that there is currently no standard
batch method for the collection of ad-
sorption data. The goal of this project
is to fill this gap by developing rela-
tively simple batch adsorption proce-
dures and guidelines for their use.
These products are intended for use by
the EPA's Office of Solid Waste in de-
veloping guidance on design and evalua-
tion of compacted soil liners for waste
storage and disposal facilities. The
results of this study will be included
in a Technical Resource Document de-
scribing the use of batch procedures,
scientific basis and rationale for the
procedures, and interpretating adsorp-
tion data generated by each procedure.
CURRENT STATUS
Cooperative work with ASTM Commit-
tee D34 (Waste Disposal) is being con-
ducted at a number of laboratories to
develop and test a batch procedure for
determining the affinity of soils for
solutes. Round-robin trials were con-
ducted under the auspices of ASTM
subcommittee D34.02.03 (Predictive
Methods). On the basis of these trials
a batch procedure has evolved for meas-
uring a 24-hour distribution ratio (Rj)
for soils and sediments at one initial
solute concentration and provides one
point on an adsorption isotherm. This
proposed ASTM procedure is being used as
the basis for an expanded procedure that
will generate data at several concentra-
tions so that an adsorption isotherm can
be constructed. Throughout the rest of
this paper, the term "24-hour Rd" will
be used to identify the single point
(ASTM) batch adsorption procedure pro-
posed for use in the TRD.
The 24-hour R^ procedure consists
of agitating a soiksol ution mixture of
1:20 by using a National Bureau of
Standards (MBS) rotating tumbler for 24
hours. The air-dried sample is initial-
ly passed through a 2-mm sieve. During
the mixing interval, neither the pH nor
the ionic strength of the mixture is
controlled. The tumbler is operated at
room temperature (22° + 5°C). After 24
hours, the mixture is either filtered or
centrifuged and the resulting solution
analyzed chemically.
The project is nearing completion;
previous papers on the project reported
findings from a comprehensive literature
review and the initial laboratory stud-
ies (Ainsworth et al., 1984). A paper
by Roy et al. (1984) discussed how com-
petitive interactions between solutes
can influence adsorption data. The in-
formation from these publications will
be integrated into the TRD.
The present paper discusses the
effects of mixing method (shaker vs NBS
tumbler), equlibration time, and
soil:solution ratio on results from
batch studies of the adsorption capacity
of soils.
Mixing Method
A first generation procedure was
formulated around the ASTM-A, Water
Shake extraction procedure (ASTM,
1979). A sensitivity analysis of this
170
-------
early procedure made by a number of
laboratories found that the method of
mixing influenced the amount of cadmium
adsorbed by the Catlin silt loam; when
the shaking was more vigorous greater
amounts of cadmium were adsorbed
(Ainsworth et al., 1984). To improve
the consistency of interlaboratory
results, a National Bureau of Standards
(NBS) rotating tumbler was tested as the
mixing system, and a second sensitivity
analysis was carried out using NBS tum-
blers (Table 1) and the Catlin soil.
The interlaboratory results were in good
TABLE 1. Cadmium adsorption by the Catlin silt loam where a NBS tumbler was
used by each laboratory.
Lab
A.
Trial
B.
Trial
C.
Trial
D.
Trial
Initial 24-hr
cone. cone.
mg/L
1
2
3
1
2
3
1
2
3
1
2
3
200
200
200
mean
std . dev.
200
200
200
mean
std. dev.
200
200
200
mean
std. dev.
190
190
190
mean
std. dev.
overall mean
std. dev.
% CV
35.7
36.2
34.6
35.5
±0.8
31.8
35.8
36.8
34.8
±2.6
35.6
35.6
35.0
35.4
±0.4
31.0
30.0
31.0
30.6
±0.6
34.1
±2.3
6.9
Rd
mL/g
92.0
90.5
95.6
92.7
±2.6
105.8
91.7
88.7
95.4
±9.1
92.4
92.4
94.3
93.0
±1.1
102.5
106.6
102.5
103.9
±2.4
96.3
±5.2
5.4
Lab
A.
1
2
3
B.
1
2
3
C.
1
2
3
D.
1
2
3
Initial 24-hr
cone. cone
mg/L
10.1
10.1
10.1
mean
std. dev.
10.0
10.0
10.0
mean
std. dev.
10.0
10.0
10.0
mean
std . dev .
9.8
9.8
9.8
mean
std . dev.
overall mean
std. dev.
% CV
0.114
0.126
0.125
0.122
±0.01
0.110
0.13b
0.165
0.136
±0.026
0.127
0.127
0.132
0.128
±0.003
0.130
0.110
0.120
0.120
±0.01
0.126
±0.01
7.94
Rd
mL/g
1734
1567
1580
1627
±92.8
1798
1461
1214
1491
±293
1554
1554
1495
1534
±34.0
1487
1761
1613
1620
±137
1568
±66.5
4.2
171
-------
agreement. The coefficient of variation
(% CV) between mean values for each lab-
oratory reflects, in part, the precision
in the mixing method. The coefficient
of variation of Rd values, based on ini-
tial cadmium concentrations of 10 mg/L
and 200 mg/L, were less than 10%. The
inter!aboratory coefficient of variation
for similar concentrations, using the
shaking method, was more than 125% in
interlaboratory comparisons (Table 2).
Because all other parts of the procedure
TABLE 2. Cadmium adsorption by the Catlin silt loam where shakers were used
by each laboratory.
Lab
A.
Trial 1
2
3
Trial 1
2
3
C.
Trial 1
2
3
D.
Trial 1
2
3
Initial 24-hr
cone. cone.
-mg/L
Shaker rate
200
200
200
mean
std. dev.
200
200
200
mean
std. dev.
200
200
200
mean
std. dev.
200
200
200
mean
std. dev.
16.9
13.2
11.4
13.8
±2.8
83.8
88.2
86.7
86.2
±2.2
1.88
1.77
1.70
1.78
±0.09
26.5
21.5
10.0
19.3
±8.46
strokes/min throw (cm)
59
7.62
70
3.81
100
3.18
70
3.18
overall mean
std. dev.
% CV
30.3
38.0
125.5
172
-------
were the same in both cases, the mixing
method was the primary contributor to
variation in the interlaboratory
means. The NBS tumbler was adopted as
the method of choice because of the much
lower coefficient of variation between
laboratory means.
Equilibration Time
The proposed 24-hour Rd procedure
provides a measurement of adsorption
after 24 hours. This may or may not be
long enough for the development of true
chemical equilibrium or steady state.
To evaluate the partitioning of solutes
and adsorbates in systems at equilibri-
um, the time interval necessary for the
development of equilibrium must be ex-
perimentally measured. As indicated
earlier, some investigators have used
equilibration times of days or even
weeks. However, adsorption per se is
generally regarded as a fast reaction,
and subsequent removal of a solute from
solution may be attributed to other pro-
cesses. The ambiguity in the definition
and measurement of equilibration times
has been acknowledged as a major problem
in adsorption studies (Anderson et al.,
1981). The EPA (U.S. EPA, 1982)
suggested that the equilibrium time
should be the minimum amount of time
needed to establish a rate of change of
the solution concentration equal to or
less than 5% per 24-hour interval. The
efficacy of this operational definition
for equilibrium was evaluated, using
seven soil materials. Each of the ad-
sorbents was exposed to arsenic and
cadmium solutions, initially containing
200 mg/L, for periods of up to 72
hours. The solutions were analyzed, and
the rate of removal of the solute was
determined. All of the soil-solute
systems in this study were found to be
in equilibrium as defined by this oper-
ational definition. Figure 1 presents
representative data for cadmium adsorp-
tion.
Soil:Solution Ratio
As previously mentioned, the
soil:solution ratio of the proposed 24-
hour Rd procedure is 1:20. However, a
single soil:solution ratio cannot be
used satisfactorily in all cases.
Cadmium adsorbed per gram of soil (mg/g)
¥ t « °?
f— • . <
I ^^L^
i
-=
s
• ^*~ — . ^
^— • *. —
• ^+——^* j
f*^^ 1
f
^Kaolinite
• •
Sangamon paleosol
Vandalia till (ablation) ^
Vandalia till (altered)
^ Vandalia till (unaltered) ^
f
L
i 1 1 1 1 |- - — T ~
0 10 20 30 40 50 60 70
Time (hours)
Figure 1. Equilibration time determinations for a cadmium solution of 200 mg/L
contacted with various soil materials.
173
-------
During the development of procedures
togenerate adsorption isotherms, it
became necessary to determine
soil:solution ratios that would permit
enough solute to be adsorbed to result
in measurable, statistically significant
differences in solution concentration.
Unfortunately, with inorganic compounds,
this cannot be determined a priori .
An empirical, systematic procedure
was developed to determine a suitable
ratio for a given soil-water system and
concentration range. A value of 10% to
30% adsorption for the highest concen-
tration used is a useful criterion.
This will give a discernible decrease in
solute concentration that is statisti-
cally acceptable with respect to the
initial concentration. An example of
this type of approach is given in Table
3. Using a 1:4 soil :solution ratio,
more than 90% of the cadmium initially
added (200 mg/L) was adsorbed by both
the Sangamon paleosol and the Vandalia
till. If the 1:4 soi1:solution ratio
were used to generate data at lower con-
centrations than the 200 mg/L used in
this example, those points may be at
concentrations below analytical detec-
tion limits. In contrast, when a 1:500
ratio was used at the lower concentra-
tion (10 mg/L), about 60% of the cadmium
initially added was adsorbed by the
Sangamon sample. However, when the 200
mg/L Cd solution was used, only 9.8% was
adsorbed at the same soil :solution
ratio. Essentially, the object is to
select a soi1 :solution ratio that will
serve as a compromise. In this case a
ratio of 1:100 was chosen to generate an
adsorption isotherm because the amount
of cadmium adsorbed from the high con-
centration range of the isotherm (200
mg/L solution) was approximately between
10 and 30%, and at the same time the
amount of cadmium remaining from a low
concentration (10 mg/L) solution was
TABLE 3. Soil:solution ratio determination for the Sangamon soil and Vandalia
Ablation till using Cd as the adsorbate.
Initial concentration - 200 mg/L
SANGAMON
Soil -.solution
Ratio
1:4
1:10
1:20
1:40
1:60
1:100
1:200
1:500
Cd adsorbed
ug/g %
722
1631
2792
4246
5165
6250
7500
9250
95.2
86.1
73.7
56.0
45.4
33.0
19.8
9.8
VANDALIA
(ABLATION)
Cd adsorbed
ug/g %
635
1359
2143
3012
3441
3880
4560
4900
94.1
76.2
44.3
25.4
19.1
13.0
8.0
4.6
Initial concentration - 10 mg/L
1:100
1:200
1:500
1:1000
957
1736
3178
4325
91.1
82.7
60.5
41.2
840
1474
2215
—
80.0
70.2
42.2
__
174
-------
also within analytical detection
limits. It is recommended that
different soil :solution ratios not be
used to generate data points for the
same adsorption isotherm. Incongruent
results due to the effects of
competitive interactions may be observed
(see Roy et al ., 1984).
Although finding a suitable
soil:solution ratio for inorganic
solutes must be done empirically, a
soil:solution ratio for hydrophobic or-
ganic solutes may be calculated if the
organic carbon content of the adsorbent
(soil), and the water solubility of the
solute are known (McCall, 1981). A de-
rivation of this estimation technique
begins with:
let
ug adsorbate/g soil
a
yg solute/g solution
uga/g
v9s/9
[1]
Also, if we let R = g soil/g aqueous
solution [2]
then eq. [1] becomes
„ «a 1
[3]
Since yga + vigs must equal the total
mass of solute initially in solution
(yg°), then
i, a 1
or
sd (ygu - ugj R
d
R =
[4]
[5]
Thus, it is possible to select an ap-
propriate soi1 :solution ratio (R) based
on an estimate of the K^ (partition
coefficient) of the organic solute from
preliminary studies or literature values.
Partition coefficients for volatile
organic solutes may be estimated from
the empirical equation of Hassett et al.
(1983), relating the solubility (S) of
the solute in water to its organic
carbon-solution partition coefficient
(Koc), viz.
log Kor = 3.95 - 0.62 log(S in mg/L) [6]
The Kj can then be calculated if the
organic carbon content (TOC) of the
adsorbent is known by
Kd = Koc (% TOC)/100
[7]
To evaluate the efficacy of this
approach (eqs. [1-7]) soil:solution
ratios for dichloroethane, tetrachloro-
ethylene, and o-xylene were calculated,
then these results were compared to
experimental results obtained by using
the Catlin silt loam (4.04% total
organic carbon) as the adsorbent.
The solubility of dichloroethane
and tetrachloroethylene is 8450 mg/L and
200 mg/L, respectively, at 25°C (Chiou
et al., 1979), and the solubility of
o-xylene is approximately 175 mg/L at
25°C (McAuliffe, 1966). From equations
[6] and [7], the calculated KJ values of
dichloroethane, tetrachloroethylene, and
o-xylene were approximately 1.3, 13.4,
and 14.7, respectively. In order to
obtain 20% adsorption (yga/yg° = 0.2),
the estimated soil:solution ratios would
be
dichloroethane:
R -
K so
( 1 ]
O]
1
5.3
tetrachloroethylene:
o-xylene:
1
R _ 20
K 80 (14.7)
68.7
From these results, a soil:solution
ratio of 1:50 was used for a mixture
containing all three compounds to gener-
ate adsorption isotherms for each organ-
ic compound (Figs. 2 and 3). It was
felt that a tri-solute system would be
more representative of leachate samples
taken in the field. Over the concentra-
tion range studied, the adsorption of
dichloroethane, tetrachloroethylene, and
o-xylene by the Catlin silt loam could
be described by the linear relationship
= KdC
[8]
175
-------
Amount adsorbed per gram of soil, x/m (Mg/g)
CO O
->• 3-
rt O
-J
— ' O
O
OI — '
rt
2
no ro
CO
o cr
O
QI
rt
in
c
-j
2:2
O 00
3- o
o -••
ro o
rt 3
3-
3 00
ro o
rt
O) 3"
3 ro
CL -j
rt 00
ro
-S -h
01
I
£
5. o
§ S'
8
§ §'
3
p
b>
I
p
bo
c:
-j
ro
ro
—i X 3>
O •< CL
01 —'in
3 ro o
3 -5
o> ro -o
rt rt
cr _..
ro<< O
oo 3
o rt
O 3- -J.
• ro oo
O
O rt
Q> 37
rt ro
O
-t)
1 O
Amount adsorbed per gram of soil, x/m en oo o NJ *
-t) cr o 01 rt oo -j rt3rt^CL-D-tirtrt
-jrtro cr ro oj ro — '3
-j-j — ' rt oirt — ' ro CL rt o c
mo oro3coo3'Oi CL Ortrtrt
oj o ro ~o "I fo oo rococo cr 3- ro 3-
3 ro jo ro < rt -•• — ' c-j-oi <-t- i -••
o c — '• 01 — ' o ic oo o :^ ro ojaroo
-) — 'rt Q- oo -". ro cu rt 3- c: -"-co
oo ro -•• — '• o oo <-+ — ' C3~ -". oi — 'rt
Ooocro-ti 3~ ro -<• — ' 3- 70 n -> -> x *
ro oo 3 3 01 c o oo r ^ ro 01 ~o rt-
coo CL3-CrOCTrtCLrt CL rt rt C rt
-•• -h ro O oo rt -••-•• Oi £ c -"• -•• -5 3-
3 — ' o — h c ro 3- — 'OO3— '-"soororo
OOOO 3 CO » CL rt Ol — '00 3
rt o (/>-"• VCOOOT rt CL 3" ro
— ' rt rt Cu rt — '3TCXcrtrDOiOirt
CTCOl— ''"a IF" 3~ — '•COC0003'nrtOO3-
rDrt3O-U3Ol3~3— i- rDOI3*Ql O
-"•CL3rD 3 -SO-} rt3 CTCL
— h O oi Oiro rt* ro ro O
c CL — ti co c 3- ro cr ~t^ o* ro ~J- c 3 3 — fi
3 ~i —i' — *' ro ro o rt — • ro ro
CLCUN rt — '3 n -i ->• — '00-JC03
rtro o — •• oi 3 •013--^-
-"• CL CT CCO CTrtOX
o C7-J oo roro£-'.
oo ro oi ro CL 3 3
1 OO
£-}
o
z
o
r~
c.
oo
( !
o
•z.
00
00 rt
o ro
-•• o
— ' 3"
• • 3
CO ->•
O o
— ' c:
c ro
rt
— J. Qj
o -a
3 -0
ro
-) 01
Ol ~!
rt co
_J.
O c-t-
t^ O
v
(^
— J.
ro
CL
CO
c
— J.
rt
Ol
CT
_j
ro
3 0 — ' 3
-h o oi ro
3 rt 3
-I--0 ro rt
3- O CL CO
o c
3 -h £
V) CL -5 Ol
3 CO O oo
-j. 3
-J 01 <
« 3 rt (T>
Q. 3- T
ro o<
H rt
3- 3- Oi O
= ro jo — •
n cz o
» o ro oo
T o ro
-l-td C
2T OJ 00 rt
-l. 3 O
/) — <• CO
n o rt
T> — ' 3-
/> o c ro
-f oi cr
-"• -j -"• ~i
3 cr — • 01
ai o -•• rt
rt 3 rt -•.
-•• V: o
0 0
3 O O O
3 -h oi
rt — '
ro rt n
3 3" C
rt ro i
00
o
— j.
_^
. .
00
o
_ *
c
o (n
"-J •
en
f — » . .
i—1
--j
i— i »
r\j
CO
O t-"
-5 O
ro o o
en ro i
. rt Cfi
on 3- .
^£> *<; on
— ' on
ro
3 ^~
^-~ ro 4=
• • •
O
i—" or
^—^
ro
rt
en 3-
. OJ
on 3
on ro
. .
f—^
t— »
^_»-
•— ' n
II M Oil
en
o
4^
1-" 00
i — >
O — 'CO
O Oi O
3 rt -••
~D ro — '
o o. ••
C CO CO
3 -li O OJ
Q- ~3 — 'CO
o c: ro
Oi 3 rt CL
CL -i.
oo rt O o
O3-33
-i ro
cr -5 rt
ro -o oi 37
a. ro rt ro
-5 -"•
c o o ro
oo ro x
— J. --3 t^^."^}
13 r+ ;o fp
tO QJ ^— ^ — ^
oo -••
ro ro n 3
jo oj ro
• O 3 3
-h rt
i — i cr oj
J> ro ro — '
l l OJ
.. O -j CL
3- ro oi
0 rt
OJ Ql
' w
o
C OJ
i
' — • — h "5 rt OJ oo rt ' — *' — • oo £
"n-"-r03-Q-ro-s3-c 03-
->• r>a_rDoo3oir~in~>rD
oo-J-c: Ortrt"^-^~-O"-j
oo root—'-cioo— '• cnmroro
. 3 ro ~^ rt o CL
rt003— i. OJ3»» X
rsi o t3 ~^-
^ — * rt rt 3 oo ^ — - OJ 7^ fO 3
oiT^oro -O-C3O.-J
3CL -jroroooCL _ i.
o. - — oiSjoo^ — '• c oo
c -•• 3 n oo 3
oo-hoo-i-OJOir— ---rt
O — '• oo rt — ' — ' — •• oi rt 3"
.-53 -•• .003 ro
TD ro o o •£.
Ol — t r~> 3 o> rt OJ ro Oi
— 'roc oo m 3" Q- — 1-3
— ' 01 -* — • ro Jo ro oo oo o
CL — ' X C O 3- C
rt-J- ^OoirD-jrt-3
3"oort3-t>rtJD-a rt
-j rt O -J. c rt O
ro -5 II rt o -j. -i. -h o
ro-'-o 3-3 — 'o -h
cr 3 7^ ro -J- 3 OJ
ocro-h i — icr D-CO
Ort» O"^ICO-JOCOO
3 -i. —o I — l -•• O O — '
-o o oi ^^n> c 3 "3 c
O 3 3 3d -5 3 oo cr rt
c CL — " 3 ro rt ro ro
3 O CLTD O Oi 3
Q-O7^£ — 'TO 3rtOi
co ro-h 3" -"• ro 3 rt CL
-»i ro o i n l
i -i 3- ro
ro 3
i
-------
empirically, but suitable ratios for
hydrophobia organic solutes may be
estimated from the compounds water
solubility and the organic carbon
content of the soil .
ACKNOWLEDGMENTS
The authors acknowledge the support
of the project by the U.S. Environmental
Protection Agency, Cincinnati, Ohio;
Cooperative Agreement No. R-810245-01;
Dr. Mike H. Roulier is the project
officer. Also, we thank ASTM and the
ASTM Committee D34.02.03 participating
laboratories for contributing their time
and effort so generously.
REFERENCES
Ainsworth, C. C., R. A. Griffin, I. G.
Krapac, and W. R. Roy. 1984. Use
of batch adsorption procedures for
designing earthen liners for land-
fill s, JJT_ Proceedings of the Tenth
Annual Research Symposium of the
Solid and Hazardous Waste Research
Division, Ft. Mitchell, KY, April 3-
5, 1984, U.S. EPA-600/9-84-007, p.
154-161.
American Society for Testing and Mater-
ials. 1979. Proposed methods for
leaching of waste materials: Annual
Book of ASTM-A Standards, Part 31,
Water, Philadelphia, PA, p. 1258-
1261.
Anderson, M. A., C. Bauer, D. Hansmann,
N. Loux, and R. Stanforth. 1981.
Expectations and limitations for
aqueous adsorption chemistry in
Anderson, M. A., and A. J. RuTrfn
(eds.), Adsorption of inorganics at
solid-liquid interfaces, Ann Arbor
Science Publishers, Inc., Chap. 9,
p. 327-347.
Bailey, G. W. and J. L. White. 197U.
Factors influencing the adsorption,
desorption, and movement of pesti-
cides in soil: Residue Reviews, v.
32, p. 29.
Barrow, N. J. 1978. The description of
phosphate adsorption curves: Jour-
nal of Soil Science, v. 29, p. 447-
462.
Barrow, N. J. and T. C. Shaw. 1975.
The slow reactions between soil and
anions. 2. Effect of time and tem-
perature on the decrease in phos-
phate concentration in the soil
solution: Soil Science, v. 119, p.
167-177.
Barrow, N. J. and T. C. Shaw. 1979.
Effects of soil:solution ratio and
vigour of shaking on the rate of
phosphate adsorption by soil:
Journal of Soil Science, v. 30, p.
67-76.
Bowman, R. S., M. E. Essington, and G.
A. O'Connor. 1981. Soil sorption
of nickel influence of solution
composition: Soil Science Society
of America Journal, v. 45, p. 860-
865.
Chiou, C. T., Peters, L. J., and V. H.
Freed. 1979. A physical concept of
soil-water equilibria for nonionic
organic compounds: Science, v. 206,
p. 831-832.
Chou, S. F. J. and R. A. Griffin.
1983. Soil, clay, and caustic soda
effects on solubility, sorption, and
mobility of hexachlorocylopenta-
diene. Environmental Geology Notes
104, Illinois State Geological
Survey, Champaign, IL 61820.
Dao, T. H. anH T. I. lavy. 1 "7° .
Atrazine adsorption on soil as
influenced by temperature, moisture
content and electrolyte concentra-
tion: Weed Science, v. 26, p. 303-
308.
Griffin, R. A and A. K. Au. 1977. Lead
adsorption by montmorillonite using
a competitive Langmuir equation:
Soil Science Society of America
Journal, v. 41, p. 880-882.
Grover, R. and R. J. Hance. 1970.
Effect of ratio of soil to water on
adsorption of linuron and atra-
zine: Soil Science, v. 109, p. 136-
138.
177
-------
Hassett, J. J., W. L. Banwart, and R. A.
Griffin. 1983. Correlation of com-
pound properties with sorption char-
acteristics of nonpolar compounds by
soils and sediments: Concepts and
1 imitations, ji_n_ Francis, C. W. and
S. I. Auerback (eds), Characteriza-
tion, Treatment, and Disposal, En-
vironment and Solid Wastes, Butter-
worth Publishers, Chap. 15, p. 161-
178.
Helyar, K. R., D. N. Munns, and R. G.
Burau. 1976. Adsorption of phos-
phate by gibbsite. I. Effects of
neutral chloride salts of calcium,
magnesium, sodium, and potassium:
Journal of Soil Science, v. 27, p.
307-314.
Hope, G. D. and J. K. Syers. 1976.
Effects of solution:soil ratio on
phosphate sorption by soils: Jour-
nal of Soil Science, v. 27, p. 301-
306.
Jones, J. P., B. B. Singh, M. A.
Fosberg, and A. L. Falen. 1979.
Physical, chemical, and mineralogi-
cal characteristics of soils from
volcanic ash in Northern Idaho: II.
Phosphate sorption: Soil Science
Society of America Journal, v. 43,
p. 657-552.
Koskinen, W. C. and H. H. Cheng.
1983. Effect of experimental
variables on 2,4,5-T adsorption-
desorption in soil: Journal of
Environment Quality, v. 12, p. 325-
330.
Harris, C. I. and G. F. Warren. 1964.
Adsorption and desorption of herbi-
cides by soil: Weeds, v. 12, p.
120.
Horzempa, L. M. and D. M. DiToro.
1983. PCB partitioning in sediment-
water system: The effect of sediment
concentration: Journal of Environ-
ment Quality, v. 12, p. 373-380.
McAuliffe, C. 1966. Solubility in
water of paraffin, cycloparaffin,
olefin, acetylene cycloolefin, and
aromatic hydrocarbons: Journal of
Physical Chemistry, v. 70, p. 1267-
1275.
McCall , P. J. 1981. Standard practice
for determination of sorption con-
stants in soil and sediments. Draft
no. 8 submitted to ASTM committee
E35.21, Environmental Chemistry
Fate-Modeling (Sorption Task Force),
32 p.
Muljadi, D., A. M. Posner, and J. P.
Quirk. 1966. The mechanism of
phosphate adsorption by kaolinite,
gibbsite,and pseudoboehmite:
Journal of Soil Science, v. 17, p.
212-229.
Roy, W. R., C. C. Ainsworth, R. A.
Griffin, and I. G. Krapac. 1984.
Development and application of batch
adsorption procedures for designing
earthen landfill liners in Seventh
Annual Madison Waste Conference,
University of Wisconsin, Madison,
Sept. 11-12, 1984, p. 390-398.
Scott, H. D., D. C. Wolf, and T. L.
Lavy. 1982. Apparent adsorption
and microbial degradation of phenol
by soil: Journal of Environmental
Quality, v. 11, p. 107-111.
Shukla, U. C. and S. B. Mittal. 1979.
Characterization of zinc adsorption
in some soils of India: Soil
Science Society of America Journal,
v. 43, p. 905-908.
U.S. Environmental Protection Agency.
1982. Chemical fate test guide-
lines. EPA 560/6-82-003.
White, R. E. 1966. Studies of the
phosphate potentials of soils. IV.
The mechanisms of the "soil/solution
ratio effect": Australian Journal
of Soil Research, v. 4, p. 77-85.
178
-------
CLAY LINER SYSTEMS: CURRENT DESIGN AND CONSTRUCTION PRACTICES
Leonard J. Goldman
Robert S. Truesdale
Coleen M. Northeim
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
Objective
Information on clay liner design and construction for hazardous waste facilities
was collected as part of the development of the Technical Resource Document: Design.
Construction. Maintenance, and Evaluation of Clay Liners for Hazardous Waste
Facilities. Critical aspects of construction that affect liner performance (e.g.
seepage rates) were identified, in part, from the literature and, more extensively,
from interviews with individuals currently involved in design and construction of
natural clay liners and admixed bentonite clay liners for waste disposal facilities.
Literature Review
There is much information available on construction practices developed to
achieve maximum bearing strength in cohesive (clay) soils for applications such
as roadways and foundations. This information, however, has little application
to design and construction of clay liners where the main objective is minimum
permeability. Except for construction of dam cores, there have been few field
examinations of clay construction related to minimum permeability. Several papers
on the theory of compaction have information applicable to design and construction
of low permeability liners. Among the critical elements in construction of low
permeability clay liners are molding water content (wet of optimum for minimum
permeability), compactive effort (interaction of duration and intensity), soil
preparation (clod size, moisture control), method of compaction (sheeps foot
roller, flat roller), post-construction effects (dessication cracking, damage by
equipment), and uniformity of spreading and mixing for admixed liners (in-place,
central plant).
Interviews
The results of the literature review were used to guide selection of topics
for interviews. These topics were discussed in interviews with more than 30
geotechnical engineers and engineering geologists representing 15 engineering and
waste management firms and several regulatory agencies. Many procedures identified
in the literature review as optimum for achieving minimum permeability were not
completely put into practice in the field because of operational difficulties
(e.g. homogenizing large volumes of soil) or because data from field studies were
not available to establish a quantitative relationship between individual aspects
of construction practice and performance of a completed liner. Some potential
construction problems and preventive measures are identified in Table 1.
179
-------
Conclusions
o The relationships between typical/optimal construction practices and
long-term performance of the completed liner are generally acknowledged
but have not been verified through field studies.
o Leakage of several clay-lined facilities has been documented; the causes
for the failures have not been established.
o Many existing facilities are functioning satisfactorily. The reason
for performance differences between functioning and leaking facilities
is not known due to a lack of post-failure analysis and detailed
construction documentation.
180
-------
TABLE 1. POTENTIAL CLAY LINER DESIGN AND INSTALLATION PROBLEMS AND PREVENTIVE MEASURES
PREVENTIVE MEASURES
PROBLEMS
Sidewall slump and collapse caused by
improper sidewall design due to poor
characterization of soil strength
profile and/or high inward hydraulic
pressure on sidewalls (sites below
water table).
Bottom heave or rupture from high inward
pressure on bottom (sites below water
table).
Accumulation of water in landfill
during construction due to rainfall
or seepage into site (sites below
water table).
Drying and cracking of clay liner
greatly increasing permeability
(dessication cracking).
Freeze/thaw cycling resulting in loss
of liner density (increased
permeability).
Reduction in clay workability due to
low temperature.
Wave erosion of upper liner after
construction (liquid impoundments).
Visible partings between improperly
joined liner lifts and increased
permeability parallel to lifts.
Pockets of high-permeability material
(e.g. sand and gravel) in liner
material.
Leaks due to improper sealing around
penetrating objects.
o Proper characterization of subsurface conditions.
o Design a more gentle sidewall slope (depending on
shear strength of foundation soil).
o Reduce hydraulic head by:
- Installing slurry wall around site perimeter
to cut off groundwater.
- Trenching and pumping around site to cut off
groundwater.
- Pumping from wells to lower local groundwater
table.
o Control depth of excavation to lower head potential.
o Reduce hydraulic head by slurry wall, pumping, or
other technique.
o Fill landfill before heaving occurs.
o Cover site with inflatable dome (reduces leachate
o treatment requirements for continuous operation
facilities).
o Seal-roll liner at end of construction day to
assure proper runoff of rainfall into sump.
o Reduce hydraulic head in surrounding soil.
o For all infiltration:
- Use leachate collection system to remove water.
- Design leachate collection system or detection
system (between liners) to handle extra
water input.
o DQ not construct during extremely hot, dry periods.
o Wet liner during dry periods.
o Cover liner with plastic sheet or soil layer if
liner construction is interrupted.
o Do not leave liner exposed prior to waste emplace-
ment or leachate collection system installation.
o Do not construct during winter in cold climates.
o Cover liner with soil blanket or other insulation
material when cold weather is anticipated.
o Increase compact!ve effort
o Cease construction till spring.
o Place rip-rap on lagoon sidewalls extending from
below liquid level into the freeboard area.
o Scarify or disc lower lift prior to installing
next lift.
o Insure moisture content of last lift and lift being
installed are the same.
o Closely inspect liner material at borrow site or
as it is being installed and reject coarse-
grained material.
o Avoid liner penetrations in design.
o Seal properly around penetrating objects.
181
-------
EFFECTS OF INORGANIC LEACHATES
ON CLAY SOIL PERMEABILITY
J. Jeffrey Peirce, Thomas A. Peel
Department of Civil & Environmental Engineering
Duke University
Durham, North Carolina 27706
ABSTRACT
Hydraulic conductivity laboratory tests results for three clay soils exposed to two
inorganic chemicals are documented, the potential usefulness of five additional labora-
tory tests for predicting clay/chemical compatibility is discussed, and an extensive
literature review is summarized. Many researchers have studied the effects that organic
solvents have on the hydraulic conductivity of compacted clays, and results indicate many
of these solvents can cause increases in the hydraulic conductivities. Little informa-
tion, however, is available in the literature concerning the effects inorganic chemicals
have on the hydraulic conductivities of clays. In this research project, the hydraulic
conductivities of three field clays; White Store, Hoytville, and Faceville permeated with
ferric chloride (500 mg/1) and nickel nitrate (50 mg/1 & 300,000 mg/1) were determined
using both fixed and flexible-wall permeameters. Results indicate that at the concentra-
tions tested, neither chemical significantly altered the hydraulic conductivities of the
three clay soils. The predictive test results are seen to be inconclusive because, as
indicated, no changes in hydraulic conductivities were observed.
INTRODUCTION
Hydraulic conductivity is an impor-
tant parameter in the evaluation of clay
soils as candidate materials for barriers
to contain hazardous wastes deposited in
landfills, pits, ponds, and lagoons. In
research and commercial laboratories
across the nation, this parameter is
studied using fixed-wall and flexible-wall
test cells. Other equipment designs and
operating procedures used historically for
water-clay interaction tests have been
modified in several ways to study the
effects of artificially contaminated
aqueous fluids on clay soils.
The literature review conducted for
this research project summarized reports
that organic solvents can cause signifi-
cant increases in the hydraulic conducti-
vities of clays, while very little
documentation is reported concerning the
effects of inorganic liquids (1).
The literature review conducted for
this research project focused on five
areas: 1) flow of liquid through porous
media, 2) laboratory equipment, 3) factors
impacting hydraulic conductivity, 4) fail-
ure mechanisms potentially impacting clay
liners, and 5) hydraulic conductivity lab-
oratory test results. Numerous research-
ers have studied the impact of organic
chemicals on clay hydraulic conductivity
in the laboratory while far fewer have
examined the effects of inorganic
chemicals.
In the study of organic chemical/clay
combinations, Brown and Anderson have
investigated the effects of various organ-
ic liquids on the hydraulic conductivities
of several clay soils (1). Large increa-
ses in hydraulic conductivity were ob-
served for all soils tested when acetone,
aniline, methanol, and xylene were used
182
-------
as permeant solutions. Daniel and co-
workers are currently involved in testing
three clays to determine the effects that
hydraulic gradient, permeant liquid, and
type of permeameter have on the hydraulic
conductivity of compated clay (5). The
three clays, a mixed cation illite, a non-
calcareous smectitic clay, and a commer-
cially processed kaolinite are being
tested at hydraulic gradients of 10, 50,
100, and 300. The hydraulic conductivi-
ties are being determined using fixed-wall
permeameters, flexible-wall permeameters,
and consolidometers with water (0.01 N
calcium sulfate), heptane, and methanol
being used as permeants. Preliminary
results show that at high hydraulic grad-
ients (150 or larger), kaolinite has a
higher hydraulic conductivity to methanol
than to water regardless of permeameter
type.
Most of the research on the effect
inorganic chemicals have on the hydraulic
conductivity of soils has been conducted
by commercial laboratories or by individ-
uals focusing only on sodium (11).
Simons et al. have undertaken an ambitious
testing program, but have to date only
conducted tests in an apparatus with semi-
fixed walls in which the sample is sur-
rounded by paraffin and no data are pre-
sently available (12). Another group
addressing this issue is active at The
Danish Water Quality Institute; to date
no reliable clay hydraulic conductivity
data have been reported (6).
At Duke University, three previous
studies have been conducted to test the
effects inorganic chemicals have on the
hydraulic conductivities of clays. Coia
tested the effects of electroplating
wastes on the hydraulic conductivities of
four clays: pure montmorinonite, pure
vermiculite, pure kaolinite, and White
Store clay (2). An increase in hydraulic
conductivity was observed for'the pure
montmorillonite while all the other clays
exhibited a decrease. Care should be
taken in interpreting Coia's data since
his equipment and procedures were not well
refined. Monserratte tested the effects
of zinc chloride and chromic acid on the
hydraulic conductivities of processed
bentonite and White Store clays (8). No
significant changes in hydraulic conduc-
tivity were observed. Tulis tested four
clays; Wyoming Bentonite, Groleg Kaoline,
vermiculite, and White Store clay, with
metal hydroxides (13). Test results show
an increase in hydraulic conductivity for
the bentonite exposed to alkaline perme-
ants while the kaolin's hydraulic conduc-
tivity decreased upon permeation with the
basic solutions.
Dunn and Mitchell tested two silty
clay soils which had been selected as
potential liner construction materials for
waste containment facilities (4). Tap
water and a synthetic acidic lead-zinc
tailings leachate were used as permeants.
In all tests a slow decrease in hydraulic
conductivity was observed. The hydraulic
conductivity was also found to decrease
somewhat with increasing gradient and the
decrease was found to be irreversible with
decreasing gradient.
Clays rich in kaolinite and illite
were tested by Laguros and Robertson (7).
Leachates consisting of 15 N sulfuric acid
and 1.5 N sodium hydroxide created in-
creases close to three orders and two
orders of magnitude respectively.
Commercial laboratories reportedly
have conducted a great number of tests on
different clays and leachates. Most of
the test results are not generally avail-
able because of confidentiality concerns,
although some data are available. The
permeants typically are complex mixtures
of chemicals of widely varying character-
istics. For example, D'Appolonia Inc.
has compiled a lengthy list of test re-
sults (3). In some cases a slight change
in hydraulic conductivity was observed,
however no drastic changes occurred.
While the literature indicates that
organic solvents can cause increases in
the hydraulic conductivities of compacted
clays, very little has been reported con-
cerning the effects of inorganic liquids.
Most of the information available on the
effects of inorganics comes from irriga-
tion studies focusing on the effects of
sodium. While this information is help-
ful, care must be used in extrapolating
this data to compacted clay liners for
hazardous waste management facilities.
PURPOSE
The objectives of this reasearch
project were to:
183
-------
1. Document literature published on
the impact of inorganic solutions
on clay soil hydraulic conduct-
tivity.
2. Generate laboratory test data on
the ability of selected clay soils
to prevent the migration of iden-
tified inorganic solutions.
3. Investigate the feasibility of
method(s) to predict deleterious
effects of inorganic solutions on
clay soil hydraulic conductivities.
APPROACH
Three field clays were selected for
this project; White Store clay from Durham
County, North Carolina, Hoytville clay
from Woods County, Ohio, and Faceville
clay from Claredon County, South Carolina.
These soils are dominated by montmorillo-
nite, illite, and kaolinite clay minerals,
respectively.
The selection criteria for inorganic
chemicals used in the laboratory tests
were specified by the U. S. EPA Project
Officer. Chemicals tested:
. represent constituents of inorganic
fluids found under field conditions,
. represent concentrations of
inorganic fluids found under field
conditions.
Based on these criteria ferric chloride
(500 mg/1) and nickel nitrate (50 mg/1)
were selected. A nickel nitrate solution
of approximately 300,000 mg/1 was also
used so that the cation exchange capaci-
ties of the soils would be exceeded. A
solution of 0.01 N calcium sulfate, widely
used by other researchers, was used as a
standard permeant liquid to obtain base-
line values.
No widely accepted standards exist
for measuring the hydraulic conductivity
of compacted clays thus the equipment and
procedures developed for this project
incorporate the most successful aspects
of past and present experimental techni-
ques referenced in the literature and/or
devised at Duke University (1). The
equipment used are modifications to
standard fixed-wall permeameters and
triaxial cells. Sketches of the equip-
ment used are shown in Figure 1.
STONES
Figure 1. Fixed (A) and Flexible-wall (B)
Permeameters used for Hydrau-
lic Conductivity Testing.
These units have been designed and manu-
factured to meet special requirements to
resist inorganic chemical attack. The
hydraulic conductivity testing protocol
used for this project is shown in Table 1.
184
-------
TABLE 1. TEST PROTOCOL USED FOR FIXED AND FLEXIBLE-WALL HYDRAULIC
CONDUCTIVITY TESTS, NUMBER OF REPLICATES
Chemical (Concentration
Fixed-Wall Flexible-Wall
SHORT TERM LONG TERM SHORT TERM LONG TERM
(4-12 Weeks) (12-36 Weeks) (4--I2 Weeks) (12-36 Weeks)
CaS04 (.
FeCl3 (500 mg/1)
Ni(No3)2 (50 mg/1)
Ni(N03)2 (300,000 mg/1)
WS F H WS F H WS F H WS F H
16 14 12 242 223 100
644 220 111 010
644 220 111 000
222 222 001 000
WS = Whitestore Clay
F = Faceville Clay
H = Hoytville Clay
-------
All fixed-wall permeatneter tests were
conducted using gradients ranging from 65
to 190 while all the flexible-wall perme-
ameter tests were conducted at a gradient
of 200. The turbidity and pH of all the
leachates were monitored throughout this
project. Atomic adsorption analysis was
also conducted on selected clay-chemical
combinations.
A major problem in conducting hydrau-
lic conductivity tests is knowing when
enough data have been collected to deter-
mine if steady state has been reached for
the test cells. For this project, read-
ings of permeant liquid inflow level are
taken at certain time intervals throughout
the test, ranging from 4 to 24 hours. The
hydraulic conductivity is computed for
each time interval. From this set of data
the first ten points are taken and a lin-
ear regression analysis is performed to
determine the slope of the hydraulic con-
ductivity vs pore volume curve. The first
point is then dropped and another value is
added on the other end of the pore volume
series. The slope is calculated again,
and in sequence the method is repeated.
In the beginning of the test this slope
could be large but as the test progresses
it could decrease and approach zero when
steady state is obtained.
The criteria for when steady-state
is obtained is defined for this research
by the following two criterion:
. when the slope of the curve cannot
be shown to be significantly dif-
ferent from zero at the 95 percent
confidence level and
. at least one pore-volume of the li-
quid has passed through the sample.
These criterion have proven useful in re-
cent studies conducted at Duke University
(2). When the slope of the regression line
cannot be shown to be significantly dif-
ferent from zero, the y-intercept value is
the equilibrium hydraulic conductivity.
For this project all hydraulic conductivi-
ty results are reported as y-intercept
values or mean y-intercept values when
more than one replicate was run.
The determination of the hydraulic
conductivities of clays to various perme-
ants can be both time consuming and costly.
Because of this, alternative tests are
needed which can estimate the effect that
inorganic chemicals have on clay particles.
In this project the effectiveness of five
laboratory tests for predicting the effect
inorganic chemicals have on clay particles
were evaluated. The five tests include:
Specific Resistance to Filtration (SRF),
Capillary Suction Time (CST), Settling (S),
Plasticity Limits, and Shear Strength.
PROBLEMS
Due to the low water holding capacity
of the clays tested, poor results were
obtained from the SRF and CST tests.
Using standard procedures, these tests are
unsuitable for predicting the effects
inorganic chemicals have on clay particles.
Approximately fifty fixed and flexi-
ble-wall permeameter tests were conducted
for this project. Of all these tests
only two flexible-wall tests had to be
terminated early; one because of a leak
in the membrane and the other due to
chemical failure of the quick connects in
the permeant supply system. In both cases
concentrated nickel nitrate (300,000 mg/1)
was being used as the permeant.
RESULTS
The hydraulic conductivity results
obtained during this research effort can
be seen in Table 2. Ninety percent
confidence intervals were determined for
all the tests in which more than one
replicate was performed. From the Table
it is seen that no significant changes in
hydraulic conductivity occurred in any
of the clays upon permeation of the sam-
ples with the chemical solutions. Also,
the hydraulic conductivity test results
obtained from both the fixed and flexible-
wall permeameters confirm the repeatabil-
ity of the data.
The leachate passing through the
clay soil samples was also analyzed. The
turbidities of all the leachates were
very low indicating that no piping occur-
red in the clay samples. In addition,
sample leachates from fixed and flexible-
wall tests with Faceville clay were
analyzed for metal content using atomic
absorption techniques. The analyses
indicated that no chemical breakthroughs
186
-------
TABLE 2. HYDRAULIC CONDUCTIVITY TEST RESULTS
( * TO"8 cm/sec)
FIXED-WALL
SHORT-TERM LONG TERM
WHITE STORE CLAY
0.01 N CaS04
500 mg/1 FeC13
50 mg/1 Ni(N03)2
300,000 mg/1
Ni(N03)2
HOYTVILLE CLAY
0.01 N CaS04
500 mg/1 FeCl3
50 mg/1 Ni(N03)2
300,000 mg/1
Ni(N03)2
FACEVILLE CLAY
0.01 N CaS04
500 mg/1 FeCl3
50 mg/1 (Ni(N03)2
300,000 mg/1
(Ni (N03)2
2.3+0.3
3.7+1.0
2.8+0.6
2.2+0.5
1.0+0.3
1.0+0.1
0.8+0.1
0.8+1.4
3.8+0.8
6.4+0.5
4.3+2.6
5.1+1.9
1.6+0.1
6.2+2.0
3.0+0.5
ND
1.6+2.8
ND
ND
ND
5.2+1.3
8.0+6.9
11.0+6.0
ND
FLEXIBLE-WALL
SHORT-TERM LONG TERM
2.8+1.6
2.1
2.0
ND
0.8+0.3
1.3
0.9
ND
6.3+1.4
6.2
5.6
ND
1.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
KEY: ND indicates Not Determined
NOTE: Confidence intervals are based on the number of replicates indicated in
Table 1.
187
-------
occurred for the 500 mg/1 ferric chloride
or the 50 mg/1 nickel nitrate solutions.
However, chemical breakthrough did occur
with the 300,000 mg/1 solution of nickel
nitrate. No chemical breakthroughs were
observed at the lower concentrations since
the cation exchange capacities of the
clays were not exceeded: several hundred
pore volumes of the low concentration
liquids would have been necessary to sat-
isfy the CEC's of the clays.
Results from the laboratory tests
conducted to estimate the effect of inor-
ganic chemicals on clay particles show
that two of the tests, Specific Resistance
to Filtration and Capillary Suction Time,
do not provide useful results due to the
low water holding capacities of the clays.
However, while other procedures did pro-
duce laboratory results, including
settling tests, liquid and plastic limits,
the predictive capabilities of these other
tests cannot be adequately evaluated since
none of the chemicals tested altered the
hydraulic conductivities of the clays.
From the results discussed above, the
following conclusions were drawn:
1) Ferric chloride and nickel nitrate did
not significantly alter the hydraulic
conductivities of the clays at the con-
centrations tested. 2) Screening tests
are potentially useful for predicting
changes in hydraulic conductivity caused
by inorganic chemicals; however, because
no significant changes in hydraulic
conductivity were observed in this re-
search project the results of the
screening tests are not conclusive.
ACKNOWLEDGEMENTS
This research was supported under
Contract Number 68-03-3149, 24-1 from the
United States Environmental Protection
Agency. The authors wish to thank
Dr. Walter Grube for his assistance.
REFERENCES
1. Brown, K. W. and D. C. Anderson,
1983. Effects of Organic Solvents on
the Permeability of Clay Soils. EPA-
600/2-83-016, U. S. Environmental
Protection Agency.
2. Coia, M. F., 1981. The Effect of
Electroplating Wastes Upon Clay As
An Impermeable Boundary to Leaching.
Unpublished Master's Thesis. Duke
University.
3. D'Appolonia., Final Report, 1983.
Data Compilation and Presentation for
Column Testing Associated with
Hazardous Waste Isolation. Project
No. 83-1068, Research Triangle
Institute, Research Triangle Park,
N. C.
4. Dunn, R. J., 1983. Hydraulic
Conductivity of Soils in Relation to
Subsurface Movement of Hazardous
Wastes. Ph.D. Dissertation, Universi-
ty of California, Berkeley, California.
5. Foreman, D. E. and D. E. Daniel, 1984.
Effects of Hydraulic Gradient and
Method of Testing on the Hydraulic
Conductivity of Compacted Clay to
Water, Methanol, and Heptane. Un-
pulished Research.
6. Hansen, A., 1984. Water Quality
Institute, Hoersholm, Denmark,
Personal Communication.
7. Laguros, J. G. and J. M. Robertson,
1978. Problems of Interaction
Between Industrial Residues and Clays.
Third Annual Conference on Treatment
and Disposal of Industrial Wastes,
Washwaters, and Residues, Houston,
Texas,
8. Monserrate, M. L., 1982. Evaluation
of the Hydraulic Conductivity of Two
Clays Exposed to Selected Electro-
plating Wastes. Unpublished Master's
Thesis, Duke University.
9. Peirce, J. J., 1984. Effects_of
Inorganic Leachates on~Clay Liner
Permeability. Draft Project Report,
EPA Contract Mo. 68-03-3149-2/1-1.
U. S, EPA.
10. Peirce, J. J., K. Witter, and G.
Sallfors, 1984. Termination Criteria
for Clay Permeability Testing,
Journal of the Geotechnical Engineer-
ing Division of the American Society
of Civil Engineers. Submitted for
publication.
188
-------
11. Roulier, M. and A. Otte, 1983, Per-
formance of Clay Caps and Liners for
Disposal Facilities: Final Report.
U. S. EPA Contract No. 68-03-3149,
Work Assignment 1-1.
12. Simons, H., W. Hansel, and E. Reuter,
1984. Braunschweig University,
Braunschweig, West Germany, Personal
Communication.
13. Tulis, D. S., 1983, Response of Clay
to Simulated Inorganic Textile Wastes.
Unpublished Master's Thesis, Duke
University.
189
-------
EFFECTIVE POROSITY OF GEOLOGIC MATERIALS
James P. Gibb, Michael J. Barcelona, Joseph D. Ritchey, and Mary H. LaFaivre
Illinois State Water Survey
P.O. Box 5050, Station A
Champaign, IL 61820
ABSTRACT
The understanding of effective porosity and the development of a method for measur-
ing the effective porosity of geologic materials is essential to predict the rates of
ground-water solute movement through those materials. This project is designed to
explore the phenomena of effective porosity as a dynamic soil/solute parameter. Modern
chromatographic techniques are being adapted as a method for measuring the effective or
dynamic porosity of unweathered, unoxidized clay soils at three sites. Preliminary work
has isolated some problems associated with collecting "undisturbed" soil samples and
conducting tracer experiments on low hydraulic conductivity soils in laboratory columns.
INTRODUCTION
Soil scientists, soils engineers,
geologists, and ground-water hydrologists
historically have been concerned with the
bulk rates of water movement through soils
and geologic materials. The application of
Darcy's equation has been the traditional
method used for making these calculations.
With the emergence of ground-water pollu-
tion problems, emphasis has shifted towards
the prediction of linear rates of solute
movement through near surface soils and
deeper lying geologic materials. To make
these types of predictions accurately, the
dynamic porosity of the soils and solutes
in question must be known.
For the purposes of this project,
dynamic porosity is defined as that portion
of the total void space of a soil matrix
that is available for fluid flow. Uncon-
nected or dead end pores and some very
small but interconnected pores in a soil
matrix are not available for fluid trans-
port (1). Similarly, that small portion of
fluid (and the void volume it occupies)
held on the individual soil grains by
interfacial forces is not available for
fluid transport (11). Effective porosity
can then be considered to be the same as
the dynamic porosity of the soil matrix.
The volume of voids not available for
fluid flow is a function not only of the
soil matrix, but a combination of the soil
matrix and the chemical nature of the
solution of interest. Fluids containing
large molecules in solution may be excluded
from pores within the soil matrix that
other fluids containing smaller solute
molecules would be capable of penetrating
(3,15). Similarly, the amount of immobile
fluid phase held on the individual soil
grains is a function of both the soil
particle surface characteristics and those
of the fluid.
Effective porosity should not be con-
fused with total porosity, drainable
porosity, specific yield, or gravity drain-
age. Use of any of these parameters as an
estimate or substitute for effective
porosity can affect the resulting estimates
of travel times for solutes through
geologic materials by several orders of
magnitude. Based on current concepts of
effective porosity the use of specific
yield or drainable porosity for relative
coarse grained soils may be acceptable. The
number of "dead end" pores is relatively
small and the amount of volume occupied by
fluid held in place by surface tension
forces is small. Conversely, for fine
grained soils the small size of all pores
190
-------
and the large number of "dead end" pores
increases the potential impact of both
phenomena. Figure 1 illustrates the impor-
tance of an accurate determination of
effective porosity for fine grained soils.
The difference between the total porosity
values and specific yield are somewhat com-
parable to the differences between total
porosity and effective porosity. The dif-
ferences become greater with decreasing
soil grain size.
0.001 0.01 01 1 r 10
Clay Silt Sand Gravel
MEDIAN GRAIN SIZE (mm)
100
Cobbles
from Davis &
DeWiest (19661
Fiqure 1. Total porosity relationship to
other soil/solute parameters.
Based on the general description
above, the measurement of dynamic porosity
is most important when attempting to
predict the travel times of solutes in
clays and other fined grained soils. The
estimates of travel times through artifi-
cial and natural clay liners is particu-
larly dependent on the ability to determine
the dynamic porosity for the anticipated
soil/solute combinations. Travel times
through thick till and clay sections previ-
ously believed to provide "years of protec-
tion" to underlying aquifer systems also
need to be reevaluated. The variability of
dynamic porosity for various classes of
chemical compounds and soil combinations is
currently unknown.
PURPOSE
The purposes of this project are 1) to
develop an understanding of the dynamic
porosity phenomena, 2) to develop a tech-
nique for measuring the dynamic porosity of
fine grained soils, and 3) to attempt to
relate dynamic porosity values to other
physical soil/solute characteristics. An
extensive literature search has been con-
ducted on the subjects of hydrodynamic and
physical properties of soil; flow in satu-
rated porous media; tracer movement in
saturated porous media; measurement of
effective porosity of saturated porous
media; and chemical reactions in porous
media.
APPROACH
The review of associated technical
literature was designed to develop a con-
ceptual understanding of the interaction of
soil matrices and various types of solutes.
The conceptual understanding of the dynamic
porosity phenomena was then used to apply
chromatography techniques to address the
problems associated with performing effec-
tive porosity measurements.
To verify the ability of the selected
testing procedure, preliminary tests were
conducted on columns packed with glass
beads and silica sand. This was designed
to minimize the problems associated with
soil/solute reactions, grain size and
shape, and packing arrangements. Perform-
ing tests on column media of known charac-
teristics also permitted the evaluation of
the sensitivity of the testing apparatus.
To minimize the effects of secondary
porosity, sites for collecting soil samples
were chosen where 1) clay soils are present
within the upper 50 feet of the unconsoli-
dated materials, 2) the soils are un-
weathered and unleached, and 3) oxidation
has been minimal. These selection criteria
provide native samples that most closely
approximate the glass bead columns and the
medium contained in standard chromatography
columns. They also should provide samples
that should most closely exhibit only
primary or intergranular porosity.
Samples are being collected in a spe-
cially constructed split-spoon sampler.
The sampler contains a 24-inch length of
polycarbonate plastic tubing and is fitted
with a cutting shoe sized to match the
inside diameter of the tubing, therefore
minimizing disturbance to the sample. The
polycarbonate tubing is then cut into
appropriate lengths, fitted with porous end
frits and caps, and used as the columns in
the laboratory studies.
Successful tracer studies demand that
extremely low flow rates be maintained.
191
-------
This approach deviates considerably from
the techniques routinely applied to labora-
tory hydraulic conductivity testing. The
relative rates of chemical diffusion of the
selected tracers and the rate of solute
migration through the soil column must be
controlled near field conditions to effec-
tively conduct the desired measurements.
The resulting long periods required to con-
duct experiments under hydraulic gradients
representative of expected field conditions
require the use of special apparatus (see
figure 2).
A computerized flow control, sample
collection, and data collection system was
designed and constructed to operate four
columns simultaneously. This permits four
simultaneous replicate measurements of soil
samples from each site. Integral parts of
the system include: a fixed head source,
flow measurement accumulators, inlet
pressure sensors, tracer injection ports,
soil columns, outlet pressure sensors, and
solute collection devices. The fixed
pressure source is maintained by a regu-
lated supply of a mixture of nitrogen and
carbon dioxide. The gas contacts the water
in the supply reservoir and in each accumu-
lator. The supply reservoir is capable of
supplying water to all our accumulators
simultaneously.
The flow measurement accumulators are
devices used when very low flow rates are
encountered (6). The accumulators in this
system are constructed of 2 ml glass pi pets
enclosed in plexiglass cylinders. Flow
rates are determined by measuring the time
it takes the water/gas interface in the
accumulator to move from an upper to a
lower sensing device. Two electric eyes
serve as the triggering mechanisms which
control the actuation of the three-way
valves to refill each accumulator without
disturbing the flow through the columns.
accumulator.
2-way
solenoid valve
J J J w
X" ^
toggle
valve ~*^&
water
1
0-50 psi
pressure
transducer
column 1
septum
injector •
0-100 psi
pressure
transducer
fraction collector
Figure 2. Laboratory apparatus for hydraulic conductivity
and dynamic porosity measurements.
192
-------
The inlet and outlet column pressure
sensors are strain-gage type transducers.
The transducers are connected to the flow
system with stainless steel fittings
designed to minimize dead space. A tracer
injection port at the head of each column
allows the introduction of small quantities
of solute or tracer into the flow system
with minimal disturbance of the system.
The soil columns, constructed of poly-
carbonate plastic, have an inside diameter
of 1.25 inches and an outside diameter of
1.50 inches. By maintaining at least a
30-40:1 ratio between the column and par-
ticle diameters, local velocity effects,
such as channeling or radial velocity
gradients, can be effectively minimized
(10). The top and bottom of each column
(end caps) are constructed of Delrin
plastic. The end caps are milled to accept
a 5 micrometer pore size stainless steel
frit and fittings to connect to the inflow
and outflow tubing.
The solute sample collection device is
an ISCO Retriever II fraction collector.
This device controls the separation of
column outflow for analysis, provides a
measure of column outflow rate, and signals
the computer to record the time when dis-
crete solute samples are collected.
The laboratory apparatus will be used
to measure both total and dynamic porosity.
To measure total porosity, a nonsorbing,
nonreactive tracer similar to the solvent
will be used. The close match between
molecular dimensions and diffusion coeffi-
cients ensures that "all" voids available
to the solvent will be penetrated. The
resulting retention times for migration of
the tracer will enable calculation of total
void volume and total porosity. Tritiated
water is the tracer to be used for this
phase of the experiment.
The use of tritiated water to deter-
mine total porosity is believed to be
superior to traditional drying, weighing,
or bulk density measurement techniques.
Drying of soil either by heating or a
solvent such as acetone has been shown by
some studies to result in alteration of the
micro-structural arrangement of soil par-
ticles, leading to changes in both total
pore space and pore size distribution.
To measure dynamic porosity, a non-
sorbing, nonreactive anionic tracer of much
larger molecular dimensions will be used.
This tracer will penetrate only those voids
capable of accornodating the tracer mole-
cules. Penetration of dead end pores or
pores smaller than the tracer molecule is
not anticipated. The resulting retention
time will be used to calculate a dynamic
porosity for this particular tracer. A
porosity based on elution data is defined
as a dynamic porosity (2,7,12,13,14). It
should be emphasized that a dynamic
porosity measurement for a specific soil
sample will be representative only of the
tested soil and other solutions containing
chemicals with similar characteristics.
Additional experiments must be conducted
for a variety of soil types and solute com-
binations to develop a usable dynamic
porosity data base. Attempts will be made
in this study to select tracers of differ-
ing molecular size to further document the
utility of molecular exclusion principles
to measure effective porosity in clayey
soils.
PROBLEMS ENCOUNTERED
The principle problems encountered to
date have been the development of the
laboratory apparatus. Scaling up chroma-
tographic procedures to the extent required
to perform experiments on soil columns has
required careful documentation and control
of sources of error. Head losses in the
testing apparatus and possible chemical
interferences must be minimized. Simi-
larly, due to the extremely low flow rates
required to conduct this type of experi-
ment, a continuous recording or automated
system had to be developed.
Finally, the selection of appropriate
tracers cannot be made until the mineralogy
of the samples to be tested are well docu-
mented.
RESULTS
Work during the first year of this
project has centered on the development of
the laboratory testing apparatus.
Hydraulic conductivity (k) is measured by
the constant-head permeameter method (8).
Values are automatically calculated, stored
and compared to previous values every five
seconds which permits the detection of
changes in hydraulic conductivity due to
oxidation or microbial growth. Leaks in
193
-------
the testing apparatus are also detected by
this methodolgy.
The apparatus for collection of soil
samples has been constructed and tested.
One set of samples has been collected and
installed in the laboratory testing appa-
ratus to assure that the field sampling
mechanism can be used for the column
studies.
Initial tracer experiments using
Rhodamine dye were conducted on a temporary
apparatus to study the behavior of various
sized particles (table 1) in a chromato-
graphic situation. The apparatus consisted
of an Altex pump, Model 110 A, in series
with a Data Instruments pressure transducer
(100 psi), SSI septum injector, column, a
second pressure transducer (50 psi), and
Gilson Filter Fluorometer, Model 121.
Inlet and outlet pressure transducer
outputs were monitored on a Cole Palmer
dual-pen strip chart recorder. Fluoro-
metric data were collected and stored in a
Hewlett-Packard 3390 A Integrator.
TABLE 1. PARTICLE SIZE AND COMPOSITION
OF TESTED COLUMNS
ParticleSize
size Particle Classification
A urn composition USGS USDA
1 200-500 Ottawa sand Medium Med. sand/
sand fine sand
2 106-250 Ottawa sand Fine Fine sand
sand
3 10-75 Glass beads Silt Very fine
sand/silt
4 5-20 Glass beads Silt/ Silt
clay
Four ranges of particle sizes were
utilized as the column packing in these
experiments.
Rhodamine" WT was used as an organic
tracer in lx!0~3 M NaHCOs buffered water.
Recovery of the tracer was greater than 85%
in all cases. All columns were run at a
minimum of seven different flow rates in
the range of 5 mL/min to 0.1 mL/min. All
runs were conducted in duplicate.
Results to date indicated that the
columns perform as an effective chromato-
graphic tool. Separation of the Rhodamine
WT spike was attained with the expected
Gaussian curve observed at greater pore
water velocities. However, effluent con-
centration curves broadened at lower pore
water velocities (10~3 crn/sec). This
phenomenon was expected since diffusive
transport is more significant at lower flow
velocities. Band broadening is a real-time
measure of diffusion processes under con-
trolled flow conditions.
Batch sorption tests were conducted on
the four column packing materials. Concen-
trations tested ranged from 0.119 ppm to
119 ppm. The tests were conducted for a
period of eight days. The 10-75 urn
diameter glass beads were the only material
which showed a statistically significant
decrease in fluorescence. This trend was
noted only at the end of the eight day
test. The decrease in fluorescence was
observed at three tracer concentrations,
119 ppm, 11.9 ppm, and 0.119 ppm.
Hydraulic conductivities (k) were cal-
culated at every flow rate for each column.
A linear increase in k was noted with
increasing flow rate to a maximum after
which it tailed off (see figure 3). This
trend supports Relyea's (13) statement that
at high velocities the dynamic pore volume
of a sample can decrease. With the dimin-
ished pore volume, a rise in pressure is
necessary to maintain a constant flow
velocity.
The total porosity of all columns
was determined statically according to the
following:
vv = ^column ~ vsolid
where: Vv = volume of voids
^column = total volume of column
vsolid = total volume of solids
[wt. of solid in column bulk
density of sol id]
and TOT = (Vv/VcoiUmn) * 100
where: TOT = total porosity (%)
By this definition total porosity includes
all void space whether interconnected or
isolated. This agrees with the use of the
term by Dotson et al. (7), Coats and Smith
(4), and Corey and Horton (5). Table 2
194
-------
Q 5 - 20 Mm Glass Beads
0 10-75/jm Glass Beads
A 106 - 250 Mm Ottawa Sand
• 250 - 500 Mm Ottawa Sand
1234
FLOW RATE (mL-mm"1)
Figure 3. Hydraulic conductivity versus flow rate.
presents the results for the tested
columns.
TABLE 2. TOTAL AND DYNAMIC POROSITIES
FOR TESTED COLUMNS
vv
void
volume
STOT
total
VDYN
pore nj
volume dynamic
Column (static) porosity (dynamic) porosity
no.
(ml)
(*)
I
2
3
4
35.6
43.1
37.1
29.1
36.5
45.3
38.4
29.1
38.64±.33 40. Oil. 3
37.68+.2S 39.5±0.4
37.64±.30 39.2±0.7
30.49±34 34.0±0.3
The values of dynamic porosity are
equal to or greater than those calculated
for total porosity in these four cases.
The results of the batch adsorption studies
indicate that retention is not a factor.
Therefore, a combination of experimental
error, packing, and pore size distribution
are the likely explanations. The materials
used for Columns 3 and 4 were uniform glass
spheres and exhibited little size variabil-
ity. Consequently, these types of packing
yield little, if any, detectable dead-end
pore space (9). Columns 1 and 2 were com-
posed of Ottawa sand, two to five hundred
times greater in particle diameter than the
glass beads. This large grain size and the
decrease in both sphericity and roundness
contributed to a porosity which is equal to
the total porosity within experimental
error.
Horvath et al. (12) defined three dif-
ferent hold-up times te, tm, and t0, and
corresponding hold-up volumes, Ve • Vm, and
V0 for a variety of solutes in a mobile
phase. These factors will be evaluated by
varying the size of the tracer in a series
of experiments and are graphically illus-
trated in figure 4.
With respect to the preliminary
experiments for Columns 1 and 2, the
chromatographic volume equals the volume of
voids, V0 = Vv. Therefore, all voids
within the columns participate in fluid
flow. The large grain size, the narrow
grain size distribution, and the lack of
both sphericity and roundness of particles
are all factors contributing to the equiva-
lency of dynamic and static void volumes.
Thus, the unsorbed solute, Rhodamine WT,
explored mainly the interstitial volume,
Ve. The intrastitial volume, V-j, was very
small.
Columns 3 and 4 are composed of non-
porous glass beads, a column material
impervious to solute and allowing no intra-
195
-------
VOLUME
0
0
TIME
where Ve - interstitial fluid volume
V, — intrastitial fluid volume
VM - unabsorbed solute volume
VQ - chromatographic hold-up volume
Vv - void volume (statically determined)
VR - volume uf a retarded solute
\l/ - fraction of intraparticulate fluid
space accessible to solute
Figure 4. Schematic illustration of chromatographic volumes.
stitial fluid volume, i.e. V-j = 0. The use
of Rhodamine WT therefore provides an exact
indication of interstitial volume, Ve.
Again, in these two columns V0 is coinci-
dental with Ve (i.e. V0 = Ve = Vv).
The soils chosen for this project are
to be fine grained with a high percentage
of clay. Expected hydraulic conductivities
(1 x 10~7 cm/sec) will be a hundred times
lower than those observed with the four
columns tested to date. Resulting porosi-
ties, therefore, are expected to have con-
siderable intrastitial volumes. Dead-end
pore volumes also will become a recogniz-
able factor with these pore and particle
size distributions over a range of flow
rates. The chromatographic hold-up volume,
V0, will be measured by use of a conserva-
tive tracer, HTO. A second tracer of
larger molecular dimension, i.e. Rhodamine
WT or polyamino acid, will be used to
determine the minimum interstitial volume
available for flow. This information will
provide porosity values as defined by the
molecular dimension of the the two tracer
molecules.
ACKNOWLEDGEMENTS
The authors acknowledge the support of
this study by the Land Pollution Control
Division, Hazardous Waste Engineering
Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio,
through Cooperative Agreement No. CR-
811030-01-0 to the University of Illinois;
Dr. Walter E. Grube, Jr., is the U.S. EPA
Project Officer.
REFERENCES
1. Bear, J., 1979. Hydraulics of Ground-
water. McGraw-Hill Inc., New York,
567pp.
2. Carlsen, L. and W. Batsberg, 1982.
Liquid chromatography in migration
studies. Scientific Basis for Radio-
active Waste Management, 11,
pp715-723.
3. Cheng, W.,1984. Static exclusion
method for determination of specific
pore volume. Analytical Chemistry,
56:11, pp!781-1785.
4. Coats, K. H. and B. D. Smith, 1964.
Dead-end pore volume and dispersion in
porous media. Society of Petroleum
Engineering Journal, 4, pp73-84.
196
-------
5. Corey, J. C. and J. H. Morton, 1968.
Movement of water tagged with 2\\ 3^,
and 180 through acidic kaolinitic
soil. Soil Science Society of America
Proceedings. 32, pp471-475.
6. Daniel, D. E., D. C. Anderson, and
S. S. Boynton, 1984. Fixed-wall vs.
flexible-wall permeameters. Proceed-
ings of the Symposium on Impermeable
Barriers. American Society of Testing
Materials, Denver, Colorado.
7. Dotson, B. J., R. L. Slobad, P. N.
McCreery, and J. W. Spurlock, 1951.
Porosity-measurement comparisons by
five laboratories, Petroleum Trans-
actions. AIME. 192, pp341-346.
8. Freeze, R. A. and J. A. Cherry, 1979.
Groundwater. Prentice-Hall, Inc.,
Englewood Cliffs, New Jersey, 604pp.
9. Graton, L. C., and H. J. Fraser, 1935.
Systematic packing of spheres—with
particular relation to porosity and
permeability. Journal of Geology.
43:8, Part 1, pp785-909.
10. Gray, W. A., 1968. The Packaging of
Solid Particles. Chapman and Hall
Ltd., London, England.
11. Hantush, M. S., 1964. Hydraulics of
wells. In Advances in Hydroscience
(V. T. Chow, ed.), Vol. 1. Academic
Press, New York.
12. Horvath, C. and H. Lin, 1976. Move-
ment and band spreading of unsorbed
solutes in liquid chromatography.
Journal of Chromatography, 126,
pp401-420.
13. Relyea, J. F., 1981. Theoretical and
experimental considerations for the
use of the column method for deter-
mining retardation factors. Radio-
active Waste Management and The
Nuclear Fuel Cycle, 3:2, pp!51-166.
14. Schweich, D. and M. Sardin, 1981.
Adsorption, partition, ion Exchange
and chemical reaction in batch
reactors or in columns—a review.
Journal of Hydrology. 50, ppl-33.
15. Street, K. W., 1984. Determination of
column void volume in HPLC using
fluorescence detectors. Journal of
Chromatographic Science, 22, June,
pp225-230.
197
-------
SOLUBILITY PARAMETERS FOR PREDICTING MEMBRANE-
WASTE LIQUID COMPATIBILITY
Henry E. Haxo, Jr., Nancy A. Nelson, Jelmer A. Miedema
Matrecon, Inc.
Oakland, California 94608
ABSTRACT
The swelling by absorption of waste liquids is a major factor in the loss of prop-
erties that can affect the performance and service lives of polymeric membrane liners.
Swelling can result in softening, loss of mechanical strength and elongation, an in-
crease in permeability, increased tendency to creep under load, and greater suscepti-
bility to polymer degradation. The significant factors involved in the swelling of a
polymeric composition in a liquid are:
- Solubility parameter or cohesive energy density of the polymer relative to
that of the liquid.
- Cross!inking of the polymer.
- Crystal li nity of the polymer.
- Filler, plasticizer, and soluble constituents in the total compound.
Of these, the solubility parameter is probably the most significant factor and the
subject of this research project.
Most of the dissolved organic constituents in a waste liquid have solubility param-
eters in the range of 8-10, which is the same range as the polymers used in the manu-
facture of the membranes. Thus, in order to avoid excessive swelling of a liner by a
given waste liquid, it is desirable to select a polymeric membrane that has a solubility
parameter outside this range or that differs in solubility parameters from the organic
constituents in the waste.
This paper presents an analysis of data from the literature with respect to the sol-
ubility parameters of organic solvents in liquids and of polymeric materials; it also
presents the results of laboratory testing of selected commercial lining materials in
various organic solvents having different solubility parameters.
INTRODUCTION
The compatibility of a proposed poly-
meric membrane liner with the waste to be
contained is one of the principal require-
ments in permitting the design and con-
struction of a waste management facility.
A liner material must be of low perme-
ability and must maintain its integrity to
effectively contain the particular waste
for the required length of time; that is,
it must be compatible with the waste
liquid. In some cases, this requirement
will be for extended periods of time, e.g.
as in landfills; in others, the require-
ment may only be for a few years, e.g.
in an evaporation pond.
The technology involving the use of
polymeric membranes for lining waste man-
agement facilities is relatively new, and
present information in the open literature
regarding the compatibility of liners and
waste liquids is limited. Because waste
198
-------
liquids generally are highly complex mix-
tures with the possibility of detrimental
effects due to the combination of compo-
nents, the compatibility of a lining
material with a specific waste liquid
should be assessed by immersion testing,
such as the EPA Test Method 9090 (EPA,
1984). In this method samples of the
candidate membranes are immersed in repre-
sentative samples of the waste for up to
4 months, and the properties of the mem-
branes are measured as a function of time.
Changes in the membrane and loss of values
of selected properties are indicative of
noncompatibility. Specific protocols
need to be developed to assess long-term
effects.
In assessing the suitability of a
polymeric composition for a product de-
signed for service in contact with liquids
and chemicals, it is common practice in
the rubber and plastics industries (Beer-
bower et al, 1967) to use solubility
parameters in designing or selecting the
polymer compound. The compound is then
exposed to the liquid with which it is
to be in contact and the effects of the
exposure upon the polymeric composition
are measured as a function of time. The
swelling and dimensional changes that take
place and the changes in mechanical prop-
erties, such as tensile strength, elonga-
tion at break, tear resistance, and mod-
ulus, are normally measured. For many
applications the service life will be mea-
sured as the time to a definite percentage
change or loss in a property relevant to
its performance. These criteria are cor-
related with failures in specific products
in simulated service tests or in actual
service. In the case of polymeric mem-
branes used as liners, the correlation of
failure with changes in specific prop-
erties has not been developed. Among the
various properties, the swelling or loss
in weight of a membrane liner in service
appears to be one of the most useful tools
in assessing its compatibility with wastes.
Associated with these changes in weight
are changes in most of the physical prop-
erties, such as tensile strength, elonga-
tion at break, tear resistance, modulus,
hardness, and permeability.
The comparison of solubility param-
eters of a polymeric membrane liner with
that of a waste liquid and its constitu-
ents has been suggested as a predictive
measure of the swelling that might take
place in a membrane liner [Haxo 1981;
Matrecon, 1983 (p. 177)]. Applicability
of the solubility parameters concept to
estimating service lives of liners, how-
ever, has not been quantified; this is
one of the principal objectives of this
project.
This paper is a preliminary report
on the work that is being conducted in a
study of the possible use of solubility
parameters to predict membrane-waste
chemical compatibility and service life.
It includes presentation of pertinent data
from the literature and the results of ex-
perimental work which show the relevance
of this concept in predicting long-term
compatibility of polymeric lining materials
to waste liquids.
OBJECTIVES
The principal objectives of the pro-
ject are:
1. To develop a method of predicting
the long-term compatibility of
membrane lining materials with
different hazardous wastes and
hazardous materials and estimat-
ing service lives under different
conditions.
2. To explore the use of solubility
parameters and other indexes that
have been used in the polymer in-
dustry to select polymer composi-
tions and design products that are
in contact with different chemi-
cal species, e.g. oils, fuels,
lubricants, and other fluids.
Though the principal thrust of this project
is the determination of the feasibility of
using solubility parameters as a means of
determining the swelling and the ultimate
service life of polymeric membranes, other
factors will be investigated as they prob-
ably will be incorporated into the model
for estimating service lives of membranes
under different service conditions.
TECHNICAL APPROACH
Waste Liquids in Contact with Liners
This project is concerned principally
with polymeric membranes in contact with
waste liquids that contain organics,
either dissolved in water or in mixtures
199
-------
of organics. Two conditions that would be
encountered by a liner in contact with
these wastes are schematically illustrated
in Figure 1.
Aqueous waste liquid
with dissolved organics
• Membrane
Aqueous waste liquid
with dissolved organics
,:,3 ; Orgamc^mixture '
Condition 1
Condition 2
Figure 1. Two conditions that membranes
could encounter when in con-
tact with waste liquids as
liners for waste management
facilities.
In the first condition, the waste
liquid that contacts the liner consists of
water with dissolved organics and probably
some inorganics. Though inorganics in a
waste liquid may affect organic solubility,
it is the organic constituents of the waste
liquid that are of importance as they are
potentially soluble in the polymeric mem-
brane. The ultimate absorption of an or-
ganic chemical by the membrane liner is
determined by the relative solubility char-
acteristics of the organic chemicals and
the liner. Once the organic chemical,
either volatile or nonvolatile, has dis-
solved in the liner it will diffuse into
and swell the liner. If the organic is a
volatile material, it can evaporate into
the pores of the soil on the downstream
side.
In the second condition, the liner
would be in direct contact with organic
mixtures. Such a condition has been en-
countered in the field when the organic
solvent concentrations having a higher
specific gravity than the waste liquid ex-
ceed their solubility limit in water and
collect on the liner. In this case, the
layer would consist of a mixture of sev-
eral organics which can partition into the
lining material. The concentration of
organics is considerably higher in the
second condition and, consequently, the
degree of swelling can be higher and
subsequent transmission should be greater.
This project addresses primarily the
matter of solubility of organics in lining
materials as determined by their relative
solubility parameters. The question of
dilute solutions of organics and partition-
ing is also considered in the study of
dilute solutions of organics in water and
in mixtures of organics.
Swelling of a Polymeric Membrane on
Exposure to Waste Liquids
Though most membranes during exposure
to an organic waste absorb constituents
of the waste liquids and swell, highly
plasticized membrane compositions can lose
plasticizer and shrink. These two proces-
ses can operate simultaneously so that, in
some cases, the plasticizer can be extract-
ed from the membrane and, simultaneously,
the organic constituents from a waste can
be absorbed by the membrane to result in
either a net swelling or loss.
The compatibility testing of membrane
liner materials for waste management fa-
cilities, e.g. EPA Test Method 9090 (EPA,
1984), has been concerned principally with
the effect upon liners of immersion in the
waste liquid or leachate which is to be
impounded and the amount of swelling that
takes place. Swelling can result in
deterioration of many of the properties of
polymeric membranes.
It should be noted, however, that
properties of polymeric membranes other
than tensile strength, elongation at break,
tear resistance, puncture resistance, and
gain in permeability can be affected by
solvents and waste liquids without showing
much swelling. Of particular importance
are the effects on semi crystal1ine mem-
branes which, depending on the polymer,
under simultaneous exposure to waste
liquids and mechanical stress, can be sub-
ject to environmental stress-cracking and
rupture.
The major factors involved in the
swelling of polymeric membrane compounds
in a liquid a re:
- Solubility parameter of the polymer.
200
-------
- Crosslinking of the polymer.
- Crystal 1inity of the polymer.
- Filler content of the compound.
- Plasticizer content of the compound.
- Soluble constituents in the
compound.
The solubility parameter is probably
the most important factor and, as noted
above, is used by polymer scientists to
measure the compatibility of a polymer
with a liquid with which it may be in
contact.
Crosslinking of a polymer or a rubber
reduces its ability to swell in a liquid.
The amount of swelling of a crosslinked
polymer in a good solvent for the raw pol-
ymer can be used as a measure of the degree
of Crosslinking: the greater the crosslink-
ing, the less the swelling.
Crystal 1 inity of a polymer acts much
like Crosslinking to reduce the ability of
a polymer to dissolve. Highly crystalline
polymers, such as high-density polyethyl-
ene, will not dissolve in gasoline, even
though they are both hydrocarbons and bas-
ically similar.
Two factors in liner composition that
also contribute to swelling are the amount
of filler used in the recipe, e.g. carbon
black, silica, or clay, and the amount of
plasticizer. The latter is extractable by
organic solvents, but only slightly ex-
tractable by water.
Generally, rubber and plastic com-
pounds contain minor amounts of water sol-
uble inorganic salts. They enter the
compound via the polymer itself, e.g.
catalyst traces, salt used in floccula-
tion, etc, and via small amounts in the
various compounding ingredients, e.g. many
of the non-black fillers contain small
amounts of water-soluble constituents.
These water-soluble salts can cause swell-
ing by diffusion of water into the mass by
the driving force of osmosis.
Solubility Parameters
Solubility parameters have found wide
use in determining the solubility of poly-
meric materials in various solvents. Some
of the many applications are reviewed by
Barton (1975), such as in the paint and
coatings industry. Also of particular in-
terest are the uses of this parameter in
studying the plasticization of polymers,
in preparing rubber blends, and in the de-
signing of rubber and plastic compositions
for contact with various oils, hydraulic
fluids, and gasoline (Beerbower et al,
1964, 1967). The latter application is
similar to that used in assessing the
swelling and performance characteristics
of liners that might be exposed to vari-
ous wastes containing organics.
The solubility parameter (&) and co-
hesive energy density (C.E.D) are concepts
related by the following equation (Hansen,
1967, p. 104):
6 = (C.E.D. )V2 = (AE/VJ1/2 (1)
where
AE = the energy required to
vaporize one mole of
material
Vm = the molar volume.
Thus, 6 is a measure of the potential
energy of any material with respect to
its energy in an entirely disassociated
form, that is, free of any inter-molec-
ular interactions. Intuitively, two
different solvents of exactly equal
potential energies should be mutually
miscible in all proportions with no loss
or gain of energy. This model of solu-
bility, termed the solubility parameter
model, was developed by Hildebrand
(Hildebrand and Scott, 1950, p. 119) and
may be expressed by the equation:
AEmix = Xi X2
(2)
where
AEm-jx = the energy of mixing
Xi 2 = the volume fractions
of components 1 and 2
6l 2 = the solubility param-
eters of components
1 and 2.
Clearly, the mathematical model agrees
with intuition in concluding that equal
solubility parameters imply no energy
change on mixing.
201
-------
The potential energy of solvents may
be simply expressed as 6, but is in fact a
sum of energies due to several different
types of molecular interaction. These in-
clude dipole-dipole interactions, London
dispersion forces, hydrogen-bonding ef-
fects, and at very close distances, repul-
sive effects. These energies are approxi-
mately additive [Hildebrand and Scott,
1950 (p. 56); Garden and Teas, 1976
(p. 428); Hansen, 1967 (p. 104)]:
Etotal =
E2 + E2 + . . .
(3)
It follows from equation (1) that:
4tal = 6l + 6l + 6l + • ' ' W
Consideration of the individual contribu-
tions of the solubility parameter compo-
nents becomes quite important in deter-
mining the solubility of complex systems
such as polymers. These do not behave in
the "ideal" manner assumed in construction
of the solubility parameter model and con-
sequently solubility is sensitive to var-
iations in the component solubility param-
eters, not just the overall solubility
parameters .
In order to properly describe the sol
ubility of polymers, models more complex
than Hildebrand's solubility parameter
model are required. The most important
model of this general form was proposed by
Hansen (1967) and is termed the three-
dimensional solubility parameter model.
It is written as:
total
' 6
2
6P+
(5)
where
5total = tne Hildebrand solubility
parameter
6d = the contribution to the
total solubility parameter
due to intermolecular
London dispersion forces
6p = the contribution due
to intermolecular dipole
interactions
6n = the contribution due to
intermolecular hydrogen-
bonding.
Approaches taken by various researchers
are described in Beerbower et al (1964),
Gardon and Teas (1976), and Van Krevelen
and Hoftyzer (1976). A generally useful
model will probably require parameters de-
fining polymer crosslinking and crystal-
linity as well as polymer solubility
parameters, and may well not be amenable
to a simple graphic presentation.
Type and Composition of Polymeric Membranes
Polymers used in the manufacture of
lining materials include rubbers and plas-
tics differing in basic chemical composi-
tion and structure, solubility character-
istics, chemical resistance, and the mode
in which they deteriorate and ultimately
fail. These polymers can be classified
into four general types:
- Rubbers (elastomers) that are gener-
ally crosslinked.
- Thermoplastic elastomers that are
not crosslinked.
- Thermoplastics that are generally
not crosslinked.
- Thermoplastics that have a relative-
ly high crystalline content, such as
the polyethylenes.
Table 1 lists the types of polymers
that are used in the manufacture of mem-
brane lining materials and shows the ab-
breviations used in this paper. The pol-
ymers most frequently used in liners have
been PVC, CSPE, CPE, IIR, EPDM, CR, and
HOPE. The thickness of commercial pol-
ymeric membranes for liners ranges from
0.5 to 3.0 mm (20-120 mil), with most in
the 0.75 to 1.5 mm (30-60 mil) range.
Most polymeric membranes are based on
single polymers, but blends of two or more
polymers are being developed and used in
liners. Also, different grades (varying in
molecular weight, branching, or crosslink-
ing) of a given type of polymer can be
used. Such variations can affect the solu-
bility characteristics of the membrane.
Generic classifications based on individual
polymers have become increasingly diffi-
cult, even though one polymer may predomi-
nate. All polymers are compounded with
auxiliary ingredients which serve different
purposes and may also significantly affect
solubility characteristics as well as per-
meability and other physical properties.
202
-------
TABLE 1. POLYMERIC MATERIALS USED IN LINERS
Polymer
Butyl rubber
Chlorinated polyethylene
Chlorosulfonated
polyethylene
Elasticized polyolefin
Elasticized polyvinyl
chloride
Epichlorohydrin rubber
Ethylene propylene
rubber
Neoprene
Nitrile rubber
Polyester elastomer
Polyethylene,
low-density
linear low-density
high-density
Polyvinyl chloride
Polyvinyl chloride
oil resistant
Thermoplastic elastomer
Abbrevia-
tion
IIR
CPE
CSPE
ELPO
PVC-E
ECO
EPDM
CR
NBR
PEEL
LDPE
LLDPE
HOPE
PVC
PVC-OR
TPE
Type of
compound3
XL
TP
TP,XL
CX
TP
XL
XL.TP
XL
XL
CX
CX
CX
CX
TP
TP
CS
Thick-
ness,
mm
0.9-2.3
0.5-1.0
0.6-1.0
0.6
0.8-0.9
1.67
0.5-1.6
0.5-1.8
0.8
0.18
0.3-0.8
0.5-0.8
0.5-3.1
0.3-0.9
0.8
1.0
Fabric
reinforced
Yes, no
Yes, no
Yes, no
No
Yes
Yes, no
Yes, no
Yes, no
Yes
No
No
No
No
Yes, no
Yes, no
No
aTP = Thermoplastic; CX = partially crystalline; XL = crosslinked.
The basic compositions of the differ-
ent types of compounds are shown in Table
2. The crosslinked compositions are usual-
ly the most complex because they contain a
crosslinking system. Thermoplastics, ex-
cept for CSPE compounds, contain no cura-
tives; although thermoplastic as supplied,
CSPE liners contain crosslinking agents
that allow the polymer to crosslink during
service. Liners based on crystalline pol-
ymers have the simplest composition and
generally consist of the polymer, a small
amount of carbon black for ultraviolet
protection, and antidegradants.
Several of the components of a liner
formulation can be affected during service
when the liner is either immersed in the
liquid or exposed to the weather. Low
molecular weight fractions in the original
polymer or blend can be lost. The oils and
plasticizers are potentially extractable
and, in some cases, biodegradable; antide-
gradants can be extracted. Loss or change
TABLE 2. BASIC COMPOSITIONS OF POLYMERIC
MEMBRANE LINER COMPOUNDS
Composition of compound
type, parts by weight
Component
Cross-
linked
Thermo-
plastic
Crystal'
line
Polymer or alloy 100 100 100
Oil or plas- 5-40 5-55 0-10
ticizer
Fillers:
Carbon black 5-40 5-40 2-5
Inorganics 5-40 5-40
Antidegradants 1-2 1-2 1
Crosslinking
system:
Inorganic 5-9 a
Sulfur 5-9 ... ...
aAn inorganic curing system that cross-
links over time is incorporated in CSPE
liner compounds.
203
-------
in any of these components can affect prop-
erties and, thus, compatibility or dura-
bility of the compound.
Most of the polymeric membrane liners
currently manufactured are thermoplastic.
Even if the polymer in the crosslinked form
is more chemically resistant (such as CPE
and CSPE), it is generally supplied as a
thermoplastic because it is easier to make
reliable seams and repairs in the field.
Membranes based on thermoplastic polymers
can be heat-sealed or seamed with a solvent
or bodied solvent. Semi crystal 1ine sheet-
ings are generally seamed by various ther-
mal welding or fusion methods.
RESULTS
Solubility Parameters of Solvents and
Other Organic Chemicals
Comprehensive tabulations of solubil-
ity parameters for common solvents and
other organic chemicals have been made by
Barton (1975, 1983). Many more data are
available in other references, and discre-
pancies exist among the data for a given
solvent. A review and analysis of these
data are being made in the performance of
this project and the most reasonable
values will be selected for use. Typical
solubility parameter values for some com-
mon solvents of different chemical types
are presented in Table 3.
Solubility Parameters of Polymers Used in
Manufacture of Membrane Linings
The solubility parameter of a polymer
is generally measured by immersing speci-
mens in solvents of different solubility
parameter values and observing the swelling
or the dissolution of the polymer when the
parameters of the two are equal, as illu-
strated in Figure 2. Most of the data pre-
sented in Table 4 for different polymers
were obtained in this manner on pure or
slightly crosslinked compositions.
Comparison of the Solubility Parameters of
Common Sol vents and Polymers
Figure 3 shows the solubility param-
eters of common solvents and some of the
polymers that are used or potentially use-
ful in the manufacture of lining materials.
A distribution of the solubility parameters
of solvents is also shown by the type of
solvent. These data show that a high
proportion of the common solvents and pol-
ymers fall within a solubility parameter
range of 8-10. This indicates that a large
degree of swelling would take place when
polymers are in contact with waste liquids
that contain these solvents and possibly
dissolution of the polymer will occur.
Swelling of 12 Polymeric Membranes in Sol-
vents of Different Solubility Parameters
In a series of preliminary experi-
ments, a variety of membranes were im-
mersed in a series of 9 solvents, differ-
ing in solubility parameters (6.8-14.5).
The results of these experiments are pre-
sented in Table 5. The results generally
show that, for each material, there is
essentially a swelling maximum. However,
the data are scattered and precise solu-
bility paramater values cannot be set.
Additional solvents need to be used in the
testing to narrow the values obtained.
There is a substantial difference among the
polymers in different solvents, with the
crystalline type materials showing signif-
icantly lower swelling. Since some of the
liner compositions, e.g. PVC, contain sig-
nificant levels of extractables as plasti-
cizers, data for these materials show a
loss in weight due to the extraction of the
plasticizer.
Swelling of Laboratory-Mixed Compounds of
Known Compositions
In view of the possibility of plas-
ticizers and fillers having effects upon
the solubility parameters and the fact
that some of the liners are probably
blends, a series of five polymers were
compounded in simple compositions without
plasticizer or filler. These polymers
included a butyl rubber, a CSPE, an EPDM,
a neoprene, and a nitrile rubber. All were
lightly crosslinked, except the CSPE com-
pound (Table 6).
Specimens of these compounds were im-
mersed in the same series of solvents as
used with the membrane liners discussed
above. The results for 28-day immersions,
presented in Figure 4, show curves similar
to those in Figure 2, except there are in-
sufficient data points. They also show
that the CSPE composition disintegrated in
the solvents of solubility parameters of
8.8, 9.8, and 10.4. These data are based
upon solubility parameters calculated from
cohesive energy densities.
204
-------
TABLE 3. SOLUBILITY PARAMETER VALUES3 FOR SELECTED SOLVENTS
6, (cal cm-3)1/2
Name
Alkanes
n-Octane
Cyclohexane
Aromatic hydrocarbons
Toluene
o-Xylene
Tetrahydronaphthalene
Halohydrocarbons
1 ,1-Dichloroethylene
Ethylene dichloride
Carbon tetrachloride
Tetrachloroethy lene
1,1 ,2,2-Tetrachloroethane
Ethers
Epichlorohydrin
Tetrahydrofuran
Ketones
Acetone
Methyl ethyl ketone
Di ethyl ketone
Aldehydes
Furfural
Esters
Ethyl acetate
Di-n-butyl phthalate
Tributyl phosphate
Tricresyl phosphate
Dioctyl phthalate
Nitrogen compounds
Pyri di ne
Quinol ine
Formamide
Sul fur compound
Carbon disulfide
Monohydric alcohol
Ethanol
1-Propanol
1-Octanol
Carboxylic acid
Acetic acid
Phenol
m-Cresol
Polyhydric alcohol
Ethylene glycol
Water
6
7.6
8.2
8.9
8.8
9.5
9.1
9.8
8.6
9.3
9.7
11.0
9.1
9.9
9.3
8.8
11.2
9.1
9.3
8.8
8.4
7.9
10.7
10.8
19.2
10.0
12.7
11.9
10.3
10.1
10.2
14.6
23.4
«t
7.6
8.2
8.9
8.8
9.8
9.2
10.2
8.7
9.9
10.6
10.7
9.5
9.8
9.3
8.9
11.9
8.9
9.9
8.9
11.3
8.9
10.7
10.8
17.9
9.9
13.0
12.0
10.3
10.5
11.1
16.1
23.4
«d
7.6
8.2
8.8
8.7
9.6
8.3
9.3
8.7
9.3
9.2
9.3
8.2
7.6
7.8
7.7
9.1
7.7
8.7
8.0
9.3
8.1
9.3
9.5
8.4
5.3
7.7
7.8
8.3
7.1
8.8
8.3
7.6
6P
0.0
0.0
0.7
0.5
1.0
3.3
3.6
0.0
3.2
2.5
5.0
2.8
5.1
4.4
3.7
7.3
2.6
4.2
3.1
6.0
3.4
4.3
3.4
12.8
8.1
4.3
3.3
1.6
3.9
2.5
5.4
7.8
«h
0.0
0.1
1.0
1.5
1.4
2.2
2.0
0.3
1.4
4.6
1.8
3.9
3.4
2.5
2.3
2.5
3.5
2.0
2.1
2.2
1.5
2.9
3.7
9.3
2.1
9.5
8.5
5.8
6.6
6.3
12.7
20.7
a25°C solubility parameters: 6 = Hildebrand parameter; Hansen para-
meters: 6^ = total ; 6d = dispersive; 6p = polar; 6n = hydrogen-
bonding. Metric unit MPa1/2 = 2.045 x (cal cm"3)1/2.
Source: Barton (1975).
205
-------
Solubility parameter, 5 of solvent
Figure 2. Equilibrium swelling as a func-
tion of the solubility parameter
of the solvent for a linear
(uncrosslinked) and network
(crosslinked) polymer [from Van
Krevelen and Hoftyzer (1976)].
Swelling of Membranes Based on Different
Polymers in Dilute Aqueous Solution of an
Organic Chemical
As indicated previously, the most
probable condition that liners will encoun-
ter is a dilute aqueous solution in which
organics are dissolved constituents. In
this case, it is necessary for the organic
chemicals that are solutes in the water
to partition with the polymeric membrane.
This problem is similar to that encountered
in the partitioning of nondegradable toxic
organics dissolved in water that enter the
biota and the food chain. In that case,
water-octanol partitioning coefficients
have become very useful (Leo et al, 1971),
and are being explored as they might be
applied to dissolved organic waste con-
stituents and polymeric membrane liners.
Partitioning coefficients for dissolv-
ed organics between water and polymeric
membrane materials have not been located
in the literature. Matrecon made an ini-
tial study (Haxo et al, 1984) of immersing
TABLE 4. SOLUBILITY PARAMETERS3 OF VARIOUS POLYMERSb
Number of
Polymer
Butadi ene/acry 1 oni tri 1 e,
75/25
Butyl rubber
Ethylene propylene
rubber
Neoprene
Polyethylene
Polytetrafluoroethylene
Polypropylene
Polyvinyl chloride
Chlorinated rubber0
Chlorosulfonated
polyethylene0
Poly vinyli dene chloride
Silicone rubber
Average
9.38
7.84
7.95
8.85
7.94
6.2
8.1
9.57
• • •
• • •
12.4
7.4
Range
9.25-9.50
7.70-8.05
7.90-8.00
8.18-9.38
7.87-8.10
• • •
• • •
9.48-9.70
8.5-10.6
8.1-9.6
12.2-12.6
7.3-7.55d
determinations
4
5
2
4
4
1
1
3
• • •
• • •
2
2
a(cal cm-3)V2.
bSource - Sheehan and Bisio (1966)
°Source - Burrell (1975).
^Source - Hauser et al (1956).
206
-------
6 7 8 9 10 11 12 13 14 15 16 17 18
i i i i i i i i i i i i
a) Different polymers
Teflon IIR PP NBR
EPDM CR PVC
Silicone rubber CSPE
P.E . CPE
Saran
b) Ranges by type of solvent
alcohols
ethers
chlorinated solvents
aromatics
, aliphatics . , esters
ketones
c) Distribution of-common solvents
10 11 12 13 14 15 16 17 18
Solubility parameters 6, (cal cm"3)1/2
Figure 3. Solubility parameters of common solvents and polymers
(data from Barton, 1975, and Burrell, 1975).
a variety lining materials in a dilute
solution of tributyl phosphate (TBP). This
organic liquid has reported solubility
parameters ranging from 8.5 to 9.0 and a
solubility in water of less than 0.1%.
Preliminary data from this study are pre-
sented in Table 7. The data show that,
on long exposures, though the concentration
of TBP in the water is low, the absorption
of TBP and possibly some water by the mem-
branes can reach significant values in
which the properties are greatly deteri-
orated. The data also show that polymers
that have solubility parameters differing
from the TBP swell less than those that
have solubility parameters in the range of
8.5-9.0. Also, the CPE that was cross-
linked swelled much less than the thermo-
plastic CPE. Those membranes that were
semi crystalline swelled the least. Such
results indicate that factors in addition
to solubility parameters must be consider-
ed in estimating swelling and ultimately
estimating service life.
Immersion of the same liners in 100%
TBP resulted in high absorption within a
day and a disintegration of PVC and CPE
membranes within 14 days, as shown in
Table 8. The results of the long-term
exposures at 0.1% concentration and the
short-term at several days follow the
same trends, but the magnitudes differ
greatly in some cases; for some liners
there were losses in weight probably due
to plasticizer extraction. The data in-
dicate that there are differences in
partitioning coefficients at different
concentrations and some solubility of the
plasticizers in the water.
A proposed study in this project, in
which samples of liners will be immersed in
dilute aqueous solutions of organics and
tested, should clarify these discrepancies.
This study should also yield information
with respect to the partitioning of dis-
solved organics between the solutions and
various polymeric membrane liners.
207
-------
TABLE 5. SWELLING OF MEMBRANES AFTER 28 DAYS OF IMMERSION IN VARIOUS ORGANIC CHEMICALS AT ROOM TEMPERATURE
N3
O
00
Solvent and solubility
Membranes
Polymer3
IIR (44)
CPE (100)
CPE (154)
CSPE (148)
ECO (178)
ELPO (172)
EPDM (94)
HOPE (184)
HDPE-A (181)
PB (221)
PVC (150)
PVC-OR (144)
Type
XL
XL
TP
TP
XL
CX
TP
CX
CX
CX
TP
TP
Extract-
ables,
11.79
17.42
11.57
6.17
7.63
4.01
14.1
0.73
2.09
3.68
32.73
30.97
Iso-
octane
6.8
64
-10
2
2
70
27
-2
5
12
...
1
-3
Normal
octane
7.6
79
-10
3
4
85
34
0
6
14
19
0
-1
Cyclo-
hexane
8.2
133
5
18
108
124
67
5
8
27
47
4
1
Xylene
8.8
102
100
89
>200d
114
61
44
8
23
22
8
2
Acetone
9.8
-4
-2
46
4
-3
2
6
1
1
...
83
117
parameter
Tetralin
9.8
136
149
159
>700d
160
59
86
9
22
...
60
37
, % weight
Tetra-
chloro-
ethane
10.4
109
294
>500d
>1000d
120
46
6
9
20
28
392
492
increase
Normal
propyl
alcohol
11.9
-4
-9
1
-2
-3
3
-5
0
0
• * •
-4
-8
Methyl
alcohol
14.5
-1
-7
2
6
-1
1
-8
0
0
• * •
-8
-13
DI
water
23.2
lb
lc
46
56
286
0.46
...
...
* • •
...
36
46
aMatrecon liner serial number in parentheses
b!74 days.
C49 days.
^Specimen too deteriorated to weigh.
635 days.
-------
TABLE 6. POLYMER COMPOSITIONS TESTED FOR SOLUBILITY PARAMETERS
Ingredient
Polymer (parts by weight)
Generic name Trade name Butyl CSPE EPDM Neoprene Nitrile
Butyl rubber
Chlorosulfonated
polyethylene
Ethylene propylene
rubber
Neoprene
Nitrile rubber
Zinc oxide
Magnesium oxide
Stearic acid
MBT .
TMTDS
Sulfur
Antioxidant
Accelerator
Wax
Molding conditions:
Temp., °C
Time, min
Polysar 301a
Hypalon 45b
Nordel 1040b
Neoprene Wb
Hycar 1042^
Protox 168
Maglite D
F300
Captax
Tuex
• • •
Flexamine
Pennzone E
Carbowax 4000
100
2.5
0.75
1.5
1.25
160
40
100
4.0
1.5
160
d
100
2.5
0.75
1.5
1.25
160
40
100
• • •
5.0
4.0
100
5.0
• * •
1.0
• • •
3.5
1.0
0.5
160
30
160
30
aPolysar Corporation.
bE.I. dupont de Nemours and Company, Inc.
CB.F. Goodrich Company.
dMolded and cooled in press; the others were crosslinked.
In the second condition that a liner
might encounter in service (shown in Figure
1, Condition 2), in which mixtures of or-
ganics are in.contact with liners, some-
what similar conditions to those with or-
ganic mixtures probably would exist when
the minor organic constituents in the waste
are nonvolatile and have different solubil-
ities in the organic mixture and in the
liner. The major organic constituent of
the mixture will tend to be absorbed in the
liner and the minor constituents will par-
tition between the organic mixture and the
polymeric membrane.
In the case of membranes based on
plasticized compositions, the organic
solvent layer above the membrane will prob-
ably extract plasticizers such as occurred
with the liners immersed in 100% TBP. If
the liner is a plasticized PVC composition,
loss of plasticizer will cause the membrane
to stiffen and become brittle. These param-
eters will be measured during the project.
ACKNOWLEDGMENTS
The work reported in this paper was
performed under EPA Contract 68-03-3213
with the Hazardous Waste Engineering
Research Laboratory, U.S. Environmental
Protection Agency, Cincinnati, Ohio.
The EPA project officer is Mr. Vincent
Salotto.
209
-------
600
400
200
0
600
400
200
0
1400
1200
1000
800
600
400
200
0
Neoprene
7 8 9 10 11 12 13 14 15 789 10 11 12 13 14 15
Solubility parameters 5, (cal cm"3)1/3
Figure 4. Swelling of five lightly cross-
linked nonfilled, polymeric com-
positions in nine solvents of
different solubility parameters.
REFERENCES
1. Barton, A.F.M. 1975. Solubility
Parameters. Chemical Reviews. Vol.
75, No. 6.
2. Barton, A.F.M. 1983. Handbook of
Solubility Parameters and Others
Cohesion Parameters. CRC Press, Inc.,
Boca Raton, Florida.
3. Beerbower, A., D.A. Pattison, and G.D.
Staffin. 1964. Predicting Elastomer-
Fluid Compatibility for Hydraulic
Systems. Rubber Chemistry and Tech-
nology, 37, 246-260.
4. Beerbower, A., L.A. Kaye, and D.A.
Pattison. 1967. Picking the Right
Elastomer to Fit Your Fluids, Chemical
Engineering, December 18, 1967.
5. Burrell, H. 1975. Solubility Param-
eter Values. In: Polymer Handbook.
Brandrup and Immergut. J. Wiley and
Sons, New York.
6. EPA. 1984. EPA Test Method 9090.
"Compatibility Test for Wastes and
Membrane Liners." Cited in the
Federal Register, Vol. 49 - No. 191.
(NTIS No. PB 85 103-026).
7. Gardon, J.L., and J.P. Teas. 1976.
Solubility Parameters. In: Treatise
on Coatings, R.R. Myers and J.S. Long
(eds.), Vol. 2, Part II, Marcel
Dekker, New York.
8. Hauser, R.L., C.A. Walker, and F.L.
Kilbourne, Jr. 1956. Swelling of
Silicone Elastomers, Ind. Eng. Chem.
48, 1202.
9. Hansen, C.M. 1967. The Three-
Dimensional Solubility Parameters -
Key to Point Component Affinities:
In Solvents, Plasticizers, Polymers,
and Resins. J. Paint Technol., Vol.
39, (505), pp. 104-117.
10. Haxo, H.E. 1981. Testing of Mater-
ials for Use in Lining Waste Disposal
Facilities. In: Hazardous Solid Waste
Testing, First Conference, eds., R.A
Conway and B.C. Malloy. ASTM Special
Technical Publication 760. ASTM,
Philadelphia, PA. pp. 269-292.
11. Haxo, H.E. 1983. Analysis and
Fingerprinting of Unexposed and Ex-
posed Polymeric Membrane Liners. In:
Proceedings of the Ninth Annual Re-
search Symposium: Land Disposal, In-
cineration, and Treatment of Hazard-
ous Waste. EPA-600/9-83-018. U.S.
EPA, Cincinnati, Ohio. pp. 157-171.
12. Haxo, H.E., R.S. Haxo, N.A. Nelson,
P.O. Haxo, R.M. White, and S. Dakes-
sian. 1984. Final Report: Liner
Materials Exposed to Hazardous and
Toxic Sludges. U.S. Environmental
Protection Agency, Cincinnati, Ohio.
(NTIS No. PB 85-121333).
13. Hildebrand, J.H., and R. Scott.
1950. The Solubility of Non-elec-
trolytes, 3rd edition. Reinhold
Publishing Company, New York.
14. Leo, A., C. Hansch, and D. Elkins.
1971. Partition Coefficients and
Their Uses. Chemical Reviews,
Vol. 71, No. 6, pp. 525-554.
210
-------
TABLE 7. EFFECTS ON PROPERTIES OF MEMBRANES3 ON IMMERSION IN A DILUTEb AQUEOUS
SOLUTION OF TRIBUTYL PHOSPHATE FOR 522 DAYS
Polymer Butyl
Type of compound XL
Matrecon number 44
Initial thickness, mil 63
Analytical properties:
Weight gain, %
Extractables, original, %
Extractables, exposed, %
Gain in volatiles, %
Change in extractables, %
Physical properties6:
Final thickness, mil
Tensile strength, % retention
Elongation at break, % retention
Stress at 100% elongation,
% retention
Tear resistance, % retention
Hardness change, points
Puncture test:
Stress, % retention
Elongation, % retention
21.9
11.8
12.3
2.2
+0.5
64.3
107
115
74
* • •
-2
73
126
CPE
TP
77
30
107.2
9.1
5.7
42.2
-3.4
47.9
10
155
6
14
-60
20
127
CPE
XL
100
36
34.4
17.4
15.2
19.7
-2.2
40.9
63
79
46
29
-20
85
125
CSPE
TP
55
33
31.6
4.1
4.0
20.7
-0.1
37.9
48
79
80
39
-14
112
131
ELPOC
TP
36
23
6.7
5.5
6.3
2.1
+0.8
20.2
48
74
70
79
-3
101
142
EPDM
XL
91
32
5.2
23.6
22.1
4.3
-1.5
36.8
93
106
81
105
-1
107
111
Neoprene
XL
90
35
41.4
21.5
16.9
24.8
-4.6
43.5
49
82
15
34
-25
66
119
PEELd
TP
75
7
4.5
2.7
3.3
4.1
+0.6
6.6
89
101
90
90
-iof
77
100
PVC
TP
59
33
46.2
35.9
26.1
11.9
-9.8
36.4
31
89
28
23
-33
48
133
HOPE
CX
105
32
0.56
* * •
0
...
31.5
88
101
92
81
-1
101
107
LDPE
CX
108
31
0.51
2.1
2.7
0
+0.6
30.9
104
102
103
86
-20
92
108
PP
CX
106
33
0.18
* • •
...
0
* • *
32.2
97
105
110
91
-2
94
42
aAll materials are unreinforced.
^A saturated solution of <0.1% TBP in deionized water.
cElasticized polyolefin.
dPEEL = Polyester elastomer.
eData for tensile, elongation, S-100, and tear are the averages of measurements made in both machine and transverse
directions.
to curling of sample, magnitude of change as reported is probably too great.
-------
TABLE 8. PERCENT
BUTYL PHOSPHATE
CHANGE IN WEIGHT FOR LINER SAMPLES IMMERSED IN TRI-
(TBP) IN 2 CONCENTRATIONS AT AMBIENT TEMPERATURES
Weight change, %
Polymer
Butyl
CPE
CPE
CSPE
CSPE
ELPO
EPDM
EPDM
HOPE
LDPE
Neoprene
Polyester
PVC
PVC
PVC
Liner
number
44
77
100
6R
55
36
83R
91
105
108
90
75
11
59
89
Type of
compound
CX
TP
XL
TP
TP
CX
XL
XL
CX
CX
XL
CX
TP
TP
TP
In
1 day
-0.6
789
79.1
15.0
13.1
6.6
-5.5
-4.5
0.1
0.4
98.2
5.6
1129
986
1090
100% TBP
14 days
-1.6
a
74.6
21.3
22.9
15.0
-7.9
-14.7
0.2
0.9
92.6
7.2
a
a
a
In 0
500 days
21.9
107.2
34.4
30.0
31.6
7.3
6.6
5.2
0.6
0.5
45.4
4.6
54.7
42.9
49.0
.1% TBP
1000 days
23.1
120.6
37.5
30.1
31.7
9.7
9.8
5.9
0.5
0.5
41.1
4.6
52.8
40.7
47.5
aToo disintegrated to weigh.
15. Matrecon, Inc. 1983. Lining of Waste
Impoundment and Disposal Facilities.
SW-870 Revised. U.S. EPA, Washington,
D.C. 448 pp. GPO #055-00000231-2.
16. Sheehan, C.J., and A.L. Bisio. 1966.
Polymer/Solvent Interaction Parameters.
Rubber Chemistry and Technology, 39,
pp. 149-192.
17. Van Krevelen, D.W. and P.J. Hoftyzer.
1976. Properties of Polymers, Their
Estimation- and Correlation with Chem-
ical Structure. Elsevier Scientific
Publishing Company, Amsterdam, the
Netherlands.
212
-------
ANALYSIS OF LEACHATE PRODUCTION
IN CLOSED HAZARDOUS WASTE LANDFILLS
Randy R. Kirkham, Scott W. Tyler,* and Glendon W. Gee
Pacific Northwest Laboratory
Richland, Washington 99352
ABSTRACT
Hazardous wastes disposed of in landfill sites may continue to drain for several
years after site closure. Leachate sources include waste fluids as well as meteoric water
trapped in the landfill during construction and operation. Waste fluid may be released
via barrel degradation, and interstitial fluid may be released during subsidence and/or
compression of waste materials. Water may also continue to enter the landfill through
structural faults. Predictions of rates and amounts of leachate production can be
developed if the hydraulic parameters and/or specific-yield values for the hazardous waste
and backfill materials are known. A literature search was completed, which showed that
limited hydraulic parameters and specific-yield information is available. Unit-gradient
and specific-yield modeling approaches were evaluated for use at hazardous waste
landfills. Specific yield was determined for two data sets: one collected by a
commercial hazardous waste site operator and provided to us by the state regulatory agency
and the other collected by the authors at a hazardous waste site located in New York
State. Specific-yield values of approximately 10 to 16% were calculated for the first
data set. These values are within the range of reported values for similarly designed
municipal waste landfills. A specific yield of 25% was calculated for recent hourly
leachate level data collected by the authors at the New York State site.
INTRODUCTION
Leachate levels rise in closed
hazardous waste landfills as waste mate-
rials dewater and as meteoric water
trapped during construction and operation
drains under gravitational influences.
This leachate collects on top of the liner
and is pumped or drained by gravity via
leachate collection systems. Under guide-
lines developed pursuant to the Resource
Conservation and Recovery Act (RCRA), the
depth of leachate in the cell above the
liner should not exceed 305 millimeters
(mm) [1 foot (ft)]. If leachate in the
cell lies above the 305-mm standing level,
operators will be required to pump or
drain the cell until it can be demon-
strated that the guideline level has been
achieved. If the cover or liner is not
functioning properly, additional fluid may
be introduced to the landfill. If models
can be developed to estimate the fluid
drainage rate from a properly functioning
closed landfill, deviations from the
predicted leachate drainage rate may be
used as indicators of cover or liner
failure. Few data have been reported on
the hydraulic properties of hazardous
wastes in closed landfills that would help
in the development of these models.
Previous work on leachate production
from sanitary landfills has emphasized the
concept of field capacity to determine the
amount of leachate produced at a land-
fill (18, 7). Field capacity is defined
as the amount of water retained by a
porous material after gravity drainage
ceases; therefore, in closed landfills
*Currently on educational leave of
absence at the Desert Research
Institute, Las Vegas, Nevada.
213
-------
where the waste is saturated, field
capacity refers to the moisture content
after the landfill is drained. Hence, the
maximum volume of fluid that may drain
from a closed landfill is the difference
between the saturated porosity and the
field capacity and is denoted as the
specific yield or drainable porosity. The
specific yield, therefore, is essential in
determining the total amount of leachate
that may be drained from a landfill.
PURPOSE
The purpose of this project was to
estimate the specific yield of typical
hazardous landfilled material. This
information can be used in conjunction
with leachate level data to estimate the
total amount of leachate that may be
drained from a landfill.
APPROACH
Our project was divided into three
main tasks: 1) to assess the availability
of data concerning the hydraulic proper-
ties of waste and backfill material, 2) to
review the conceptual models applicable to
investigating the drainage characteristics
of wastes and soil in a landfill environ-
ment, and 3) to select an appropriate con-
ceptual model to estimate leachate produc-
tion from typical hazardous waste sites.
With respect to this last task, it should
be noted that much of the historical data
on leachate production and level used here
were collected by hazardous waste site
operators and do not necessarily meet
standard quality assurance guidelines.
Assessment of Data Availability and
Sources
In order to model the flow of water
through landfill wastes and soil, several
types of data are necessary. To describe
the flow of liquid through landfill
materials, data must be obtained on the
physical packing and disposal techniques
practiced in hazardous waste cells. In
most landfill scenarios, flow of liquid
will occur through saturated and partially
saturated media. To represent these types
of flow, data concerning the saturated
hydraulic conductivity and porosity, and
the effects of desaturation on the capil-
lary pressure and hydraulic conductivity
are required. In addition, landfilled
material may also deform and consolidate.
Geotechnical data must be used to deter-
mine the effects of loading and consolida-
tion on the saturated void ratio and
saturated hydraulic conductivity.
Information on hydraulic properties
of wastes and soil drainage are found in
the literature from a wide range of disci-
plines (e.g., chemical, civil, agricul-
tural, and geotechnical engineering;
biology; geology; soil physics; and soil
chemistry). A comprehensive literature
review was undertaken using computer-
assisted data bases. The data bases
surveyed included NTIS (for government
research reports), COMPENDEX (for engi-
neering documents), and BIOSIS (for
biological and environmental publica-
tions). Key words, such as landfill,
drainage, landfill cells, liner, leachate,
conductivity, and permeability were used
to identify titles and abstracts appli-
cable to this study. In addition, owners
and operators of hazardous waste landfills
were contacted to determine their disposal
techniques and types of waste forms han-
dled at the facility. Results of this
literature review are summarized in
Table 1.
Conceptual Models
Simplifying the assumptions about the
complexity of the flow was necessary in
order to model the behavior of hazardous
waste landfills. The two major assump-
tions made for this analysis were that
1) all liquid flow is vertical and 2) the
hydraulic properties of the waste mate-
rials are uniform throughout the landfill.
The rationale and implications of these
assumptions are discussed below.
The reduction of the flow geometry
from three dimensions to a one-dimensional
(1-D) analysis is dictated by the
increased complexity and computational
costs associated with 2-D and 3-D flow
models. Although landfills are 3-D in
nature, the assumption of flow only in the
vertical direction may be valid for land-
fills of regular geometry receiving uni-
form areal recharge. The assumption may
not be valid in landfills where surface
soils (covers or daily backfill) or sur-
face slopes result in increased runoff in
certain areas of the landfill and ponding
of precipitation in others. In addition,
horizontal hydraulic gradients at the
landfill sidewalls or the presence of a
214
-------
TABLE 1. REPORTED HYDRAULIC AND GEOTECHNICAL PROPERTIES OF HAZARDOUS WASTES
Waste type
Municipal waste
Municipal waste
Papermill sludge
Flue gas
Desul funtzaion (FGD)
sludge
Solvay soda ash
sludge
FGD sludge
Municipal waste
FGD sludge
Lead/zinc mill
tail ings
Uranium mill tailings
Municipal waste
to
>-• Uranium mill tailings
Ui Coarse
Medium
Fine
Fly ash
Municipal waste
Municipal waste
Spent oil shale
Municipal incinerator
residue
Phosphate tailings
Coal mine wastes
Haste clays from
phosphate mining
Fly ash/soil mixture
Red mud (aluminum
tai 1 ings)
* - particle density
# - compression index
** - bulk density
Saturated parameters
Compacted Grain
KsftT density Partially saturated parameters size
(cm/sec) Porosity (kg/m3) K(e) e(H) 9RES Field Cap. data
50% 45%
340.0 10-14%
1.0 x 10"4
1.0 x 1C"8
1.0 x 10'4
1.0 x 10'b
2.71* Yes
46-57% 1040-1280 Yes
20-35%
1.0 x 10'j: 2.72-2.52*
1.0 x 10~b
3.4 x 10"f 41-57% 2.88-3.02* Yes
5.0 x 10'5
2.2 x 10'4 44% 1.63**
Yes Yes 30-40%
6.3 x 10"3 43.3% 1.48-1.57** Yes Yes 7.6% Yes
2.3 x 10~3 45.8% 1.28-1.50** Yes Yes 9.0% Yes
6.7 x 10'7 66.0% 0.90-1.10** Yes Yes 31.0% Yes
8.3 x 10'3 0.99-1.50** Yes Yes
5.0 x 10"2 Yes Yes
Bulk density
versus field
capacity data
287.0 28.6%
4.1=6.9 x
ID'5
1.2 x 10~4
2.0 x 10~b
1.4 x lO'3
7.22 x 10-b
1.0 x 10-4,
1.0 x 10"8
1.0 x 10'3
1.5 x 10'5
1.2,20 x
io-7
Consolidation parameters Shear
Cv % Vol strength
(crrr/sec) E(a) reduction data Source
(13)
(19)
(1)
Yes Yes (8)
2.87* Yes Yes (12)
Yes Yes (6)
Yes (24)
0.75-1.05* Yes (2)
(9)
(15)
(22)
(11)
(2)
(7)
(18)
Yes Yes (17)
(14)
(23)
(5)
1.0-6-OxlO-4 (10)
(16)
0.09-6.9xlO"2 (?1)
-------
shallow water table may produce signifi-
cant components of horizontal flow.
The second major assumption of uni-
form hydraulic properties ignores the
effects of waste heterogeneity. Typical
hazardous waste landfills in the United
States handle a wide variety of solid
wastes ranging from contaminated soils and
sludges in bulk form to drummed waste and
polychlorinated biphenol-contaminated
transformers. In addition, soil material
is often placed between the layers of
drums to allow vehicle movement within the
landfill cell during operation. Modes of
deposition of the wastes range from neat
upright or horizontal placement of drums
and careful mapping of their location to
haphazard disposal of drums and containers
by dragline or overhead crane. Bulky
waste material is often dumped and spread
by bulldozer or front-end loader. In all
cases, waste properties cannot be expected
to be uniform throughout the landfill.
For this study, however, the focus is
not on a complete description of fluid
movement throughout the cell, but on that
amount of fluid draining to the leachate
collection system. The effect of the
drainage collection system is to average
the local drainage from various portions
of the cell. In similar problems of
agricultural drainage (4), simplified
models of flow mechanisms can be employed
because the details are lost in the
process of averaging over large areas. As
a result of a review of the available
research on flow in hazardous waste cells,
two types of modeling approaches were
chosen for application: unit gradient and
specific yield.
Unit Gradient
The unit gradient approach (20) is
applicable to soil or bulk waste cells
(i.e., those cells receiving a uniform,
soil-like waste material). Such cells
might be found at private generator/
disposer sites where cells are used
strictly for a specific kind of waste
(e.g., flue-gas desulfurizing sludge from
a coal-fired electric utility plant).
These waste types would be expected to
behave like soils, and drainage from these
materials has been analyzed as such (2).
In order to assess the validity of
this approach, we compared this unit
gradient analysis with a fully implicit
1-D numerical ground-water flow code,
UNSAT1D (3). The model was run using 50
equally spaced node points over the 10-m
depth of a simulated landfill. The water
table was set at the bottom of the fill.
The resulting drainage from the
bottom of the simulated waste cell is
shown in Figure 1. As can be seen, both
solutions exhibit a period of constant
drainage (stage one drainage) followed by
a rapid decrease in drainage. At later
times, the drainage from either solution
decreases slowly. The total change of
fluid stored in the landfill at the end of
1 year (the length of time for numerical
simulation) estimated by the analytical
solution was 15% greater than that pre-
dicted by the numerical model. Although
the analytical method overestimates the
length of time required for stage one
drainage, the solution closely matches the
drainage characteristics at later times,
with the drainage rates after 1 year
virtually identical between the analytical
and numerical solution. Therefore, only
the analytical approach need be used in
homogeneous bulk waste cells.
Specific Yield
The second approach is for landfill
cells that may not contain primarily soil-
like material. Currently, many commercial
waste sites dispose of drummed, solidified
waste. Leachate will be stored in the
large voids between the drums as well as
in the pore spaces of any backfilled soil.
In most cases, the large voids will con-
tain the most leachate. Under drainage
conditions, these large voids will easily
dewater, leaving the leachate in the back-
filled soil behind. The concept of free-
draining pore spaces (voids) can be used
to model fluid drainage in drum disposal
cells.
The specific yield of a porous medium
is defined as the volume of drainable pore
space per unit volume of porous medium.
In a drum disposal cell, the drainable
pore space primarily consists of the void
spaces between the drums. The volume of
the interdrum voids may be estimated using
the number of drums in the cell, the
amount of backfill soil and bulky waste,
216
-------
1.25
D Unit Gradient Method
O Finite Difference Method
0 50 100 150 200 250
Time in Days
300
350 400
Figure 1. Comparison of drainage rates from Yolo light clay using unit
gradient analysis and numerical analysis.
and the overall cell volume. Given an
efficient underdrain system (one that will
cause little flow blockage), leachate
levels in the cell may be predicted using
the specific-yield approach.
Using the assumption that the porous
material has sufficient time to drain to
near equilibrium, the specific yield is
given by the following equation:
sy =
LFV
x 100
(D
where Sy is the specific yield, Q is the
amount of leachate pumped out of the
bottom of the landfill, and LFV is the
volume of landfill drained. This Sy value
includes the volume of porous and solid
materials in the landfill volume. Often
the Sy of only the porous material is
required and is obtained by dividing the
Sy for the entire landfill volume by the
ratio of solids to porous, soil-like
materials.
PROBLEMS ENCOUNTERED
Limited information on hydraulic
properties of bulk wastes (e.g., contami-
nated soils, sludges), layered materials
of different hydraulic properties, and
solid materials prevent useful application
of the unit gradient approach. The
specific-yield values, when properly
selected for a given waste material,
appear most promising for determination of
leachate production.
Leachate production data collected at
a commercial hazardous waste landfill
required extensive data manipulation to
reduce variability in order to develop an
estimate of specific yield or drainable
porosity of the waste material. The range
of specific yields calculated compared
favorably with those reported for munici-
pal waste landfills. Nonequilibrium
levels in the standpipes affected the
analysis i n two ways:
1. It was necessary to use the maximum
monthly recorded leachate level in
each subcell to estimate the actual
leachate level in the subcell. Daily
variations in the recorded levels
were the results of leachate levels
not equilibrated because of slow
standpipe recharge.
2. Hydraulically connected standpipes
often did not respond in a similar
manner to leachate pumping. This may
be a result of nonequilibrium condi-
tions or measurement error.
217
-------
RESULTS
The models evaluated represent sim-
plified conceptual models of hazardous
waste landfills. Under actual field
conditions, no one model may be completely
applicable. The models chosen are
designed for use under a variety of land-
fill configurations and waste types.
Table 2 outlines the key points of the
models and the typical data needed for
thei r application.
Site Description Specific Yield Analysis
(Commercial Site)
For this study, we analyzed data from
a (<2 acre) hazardous waste landfill site
located in a humid environment. The site
has many waste cells; each cell is
designed to minimize leachate migration
into the surrounding environment. The
cells are excavated in native clay and
lined with both compacted clay and flex-
ible membrane liners. Figure 2 illus-
trates the general plan of each of the
subcells in Cell #3. Data collected for
Cell #3 at this site are analyzed in this
report. This cell is divided into five
subcells; each subcell is hydraulically
separated using clay berms to allow the
segregation of certain materials.
The data collected by the site
operators consisted of daily leachate
level measurements in standpipes for each
subcell and monthly total leachate volumes
pumped from each cell. Two leachate
levels were recorded for each day for each
subcell: the level before pumping and the
level measured immediately after pumping
stopped. Level measurements were taken in
standpipes using a weighted string, and
pumping volumes were measured using
totalizing flow meters.
TABLE 2. SUMMARY OF CONCEPTUAL MODELS
Model
Unit gradient
or finite
difference
Application
Bulk waste forms, known
hydraulic properties, uniform
initial water content profile,
deep landfill
Resulting data
Drainage flux
versus time, water
content profil e
versus time
Specific yield
Unknown hydraulic properties,
drummed or rapidly draining
waste forms
Leachate level
versus time
58
59
62
iStandpipe Pair
Numbers Indicate Subcell Notation
Figure 2. Plan view of Cell #3.
218
-------
To collect leachate generated during
landfilling operation, leachate collection
networks are located in each subcell.
Each network consists of two perforated
concrete standpipes connected to one
another by a perforated 0.10-meter (m)
clay pipe. The standpipes were con-
structed as the subcell was filled, so
that at no time did the standpipe extend
more than 1.21 m above the working floor
of the landfill. The bottom liner was
sloped toward each standpipe to facilitate
drainage. Leachate was pumped from one
standpipe in each subcell, and the
leachate level was measured daily in each
standpipe.
Waste in each of the cells consisted
of both bulk (i.e., sludges, contaminated
soils) and drummed [i.e., 0.2-m
(55-gallon) drums] material. Drums were
stacked upright as closely together as
possible. Sludge waste material was
unloaded into a bermed area built to
accept the amount of material coming in on
a particular day. Bulk waste material was
also placed in each cell using a front-end
loader. It was assumed that a volume
ratio of two parts bulk waste to one part
drummed waste existed in Cell #3 as esti-
mated by landfill operators. At the end
of each daily shift, at least 0.15 m of
cover material were placed over all
exposed drums. This cover material ranged
from natural clays and sands to neutra-
lized wastewater sludges. On completion,
the cell was sealed with a minimum of
0.91 m of compacted clay, 0.45 m of
uncompacted clay or clay/soil mixture, and
0.15 m of top soil, and then seeded with
perennial rye and bluegrass.
Cell #3 received waste from the
summer of 1979 to the fall of 1982, when a
temporary cover was installed to reduce
infiltration. Final closure was completed
in September 1983. Leachate was removed
from the standpipe collection system
during cell operation using a portable
pump and/or vacuum trucks. Beginning in
February 1984, an automatic pumping system
was installed that consisted of sump pumps
actuated by leachate level.
Results for the Commercial Site
This analysis focuses on interpreting
the existing leachate production data from
Cell #3 to determine the hydraulic proper-
ties of the waste material. To analyze
the leachate production from a mixed-waste
disposal site, we used field data to esti-
mate the specific yield or drainable
porosity. These data are extremely
important because they can be used to
estimate the total amount of drainable
leachate in the landfill. In estimating
the specific yield, the leachate level in
the landfill is assumed to have reached
equilibrium when the monthly maximum value
occurs. In cases where the leachate pump-
ing rate is small in comparison to the
volume of leachate contained in the land-
fill, the assumption of instantaneous
drainage appears valid. The temporary cap
is believed to have significantly reduced
fluid input to the cell. Assuming that
fluid input to the cell is negligible, the
leachate level data taken after September
1982 should have been primarily influenced
by the leachate pumping. The data indi-
cate that the volume of leachate pumped
declined rapidly after the temporary cover
was installed. Table 3 shows the monthly
maximum leachate levels in each of the
five subcells. These data show clearly
decreasing leachate levels in response to
leachate pumping. Table 3 also indicates
that the leachate levels in all but
Subcell #62 decreased at a nearly constant
rate, whereas the leachate production rate
dropped off in an exponential fashion.
One possible cause for the linear
decline in the leachate level was due to
the variable geometry of each subcell as
it drained. Because the sidewalls of the
cell slope, removing a given volume of
leachate at high leachate levels will not
lower the level as much as when the
leachate level is low. If the cell walls
were vertical, the leachate level decline
would be directly proportional to the
leachate production rate. Given the
sloping sidewall, however, we expected to
see a nonlinear behavior in the leachate
level/production ratio. To produce the
linear trend in leachate level found in
each subcell of Cell #3, a leachate
production rate that decreases with time
is required (see Table 3). A second
possible cause of the trend toward a
linear decline in the leachate level may
be nonequilibrium conditions in the
standpipes. As the leachate level falls,
so does the recharge rate to the stand-
pipes because the hydraulic gradient is
reduced. In the later stages of dewater-
ing a cell, the leachate levels recorded
in standpipes may not represent the actual
219
-------
TABLE 3. LEACHATE DATA FROM SUBCELLS OF CELL #3 MONTHLY MAXIMUM
LEACHATE LEVEL AND VOLUME PUMPED SUBCELL
Subcells
Date
Sep 82
Oct 82
Nov 82
Dec 82
Jan 83
Feb 83
Mar 83
Apr 83
May 83
Jun 83
Jul 83
Aug 83
Sep 83
Oct 83
Nov 83
Dec 83
Jan 84
#58
7.01
6.92
6.49
6.95
6.52
5.55
5.73
5.73
5.00
4.51
4.15
4.08
3.99
3.87
#58A
7.07
6.07
6.52
6.71
6.25
4.57
4.45
4.57
4.60
4.48
3.29
3.35
3.60
3.54
#59
2.62
2.87
1.89
2.50
2.87
1.68
1.77
1.71
0.79
0.46
0.30
0.49
1.16
0.88
#59A
2.23
2.96
1.68
2.38
2.93
2.56
1.49
1.80
1.13
1.07
0.73
1.13
0.67
0.55
#60
1.43
1.40
1.40
1.43
1.34
1.07
1.37
0.85
2.38
0.46
0.34
0.49
0.40
0.43
(m)
#60A
1.83
1.43
1.55
1.52
1.25
1.52
0.98
1.46
3.81
0.52
0.43
0.30
0.34
0.40
#61
1.92
1.92
1.92
1.52
1.43
1.71
1.40
1.49
1.22
1.10
0.85
0.52
0.61
0.64
#61A
1.89
1.98
1.74
1.58
1.28
1.58
1.58
1.40
1.68
1.58
0.98
0.76
0.76
0.67
#62
2.68
5.49
7.07
4.39
3.78
7.35
3.84
2.53
1.95
1.55
1.49
1.04
2.04
1.77
#62A
6.77
7.77
3.14
7.04
6.74
6.04
6.64
5.91
4.54
3.51
2.83
2.74
2.47
2.07
Total
m
pumped
363.40
155.58
185.49
147.63
111.81
109.68
90.85
90.85
72.65
60.51
120.52
81.88
58.67
46.61
38.20
36.93
37.01
levels in the landfill but will be lower
because the leachate will not flow into
the standpipes as quickly.
To test these two possibilities, we
calculated the specific yield over two
time periods. The first time period was
from September 1982 to September 1983, the
second from October 1983 to January 1984.
The first of these periods reflects the
time during which the temporary cover was
in place. The second period covers the
time the final cover was in place. Cover
integrity and collection system efficiency
can be determined by examining data from
these time periods. If cover leakage
occurred from September 1982 to September
1983, the specific yield calculated during
this period would be higher than that
calculated after permanent closure. This
increased specific yield would happen as a
result of infiltrating water offsetting
the reduction in the leachate level
brought about by pumping. Also, the
question of equilibrium conditions can be
answered. If leachate levels in the
standpipe are lower than the actual waste
leachate level, then the specific yield
calculated during these times would be
Tower (less fluid drained from a unit
decline in head) than that calculated when
equilibrium conditions existed.
Using Table 3 data, calculations show
that the total saturated cell volume in
September 1982 was 47,969 mj. In
September 1983, the saturated cell volume
was 22,782 m3. The total leachate volume
pumped from September 1982 to Septem-
ber 1983 was 1,649 mj. The resulting
specific yield is calculated as
1,649 m3
S = o-
y 47,969 - 22,782 rn
S = 0.065 or 6.6%
The same method was used to determine
the specific yield during the period from
October 1983 to January 1984. The total
leachate volume pumped during this period
was 158 m3. The specific yield is calcu-
lated as
S. =
158 m
y 18,340 - 16,893 nT
S = 0.11 or 11.0%
The results of these analyses indi-
cate that the specific yield was 6.6%
during the first time period and 11.0%
220
-------
during the second time interval. The
increase in specific yield (after the
final cover was installed) indicates that
the temporary cover was as effective in
reducing infiltration as the final cover.
Leakage through the temporary cover would
have resulted in a higher calculated
specific yield because infiltration would
tend to offset the effects of leachate
pumping. The increase in specific yield
during the second time period also indi-
cates that nonequilibrium conditions did
not significantly affect the results. In
this case, nonequilibrium conditions would
have caused leachate levels measured in
the standpipes to be lower than those in
the landfill, and the specific yield
calculated under these conditions should
be lower than that calculated from actual
leachate levels.
The rise in the specific yield value
is most likely the result of the inhomo-
geneities of the wastes and the subcell
variations hidden by averaging the
saturated waste volumes.
The specific yields determined in
this analysis suggest only an averaged
value over the entire landfill. They do,
however, represent a range of values for
this type of waste. Because approximately
two-thirds of Cell #3's volume is taken up
by soil or porous waste, the overall
specific yield calculated for the landfill
is two-thirds of the specific yield of the
bulk waste. Using the specific yields
calculated for the entire landfill volume
of 6.6% and 11% [Equation (1)], the bulk
waste specific yields would be 10% and 17%
for the two time periods investigated.
These figures are in the same range as
those reported for municipal waste.
Specific yields reported for municipal
waste range from 12% to 21% (7, 22).
Initial estimates of specific yield or
drainable porosity of hazardous waste
landfills may be calculated from existing
data for municipal waste with a reduction
factor accounting for the volume of waste
containing no drainable porosity (e.g.,
drums, transformers) as estimated from
waste inventories.
Site Description Specific Yield Analysis
(Research Site)
An existing hazardous waste site in
New York State was selected for installa-
tion of an automated water level recorder*
to help determine a typical specific yield
value for mixed waste. The automated
water level recorder is used to determine
the changes in water level associated with
leachate removal. This waste site has an
ineffective clay cap, which is scheduled
for replacement during the summer of 1985.
The liner is constructed from two layers
of clay separated by gravel with a leak
detection system located in the gravel.
No detectable leakage from the waste site
has occurred. At the end of August 1984,
a leachate level of 3.6 m was observed
above the top clay liner. The leachate
level is being monitored as leachate is
removed from the waste site in 25.0 m3
(6600 gallon) increments.
Results for the Research Site
Figure 3 illustrates the changes in
water levels after eight leachate removal
events. A specific yield of 25% was
calculated for the period August 30 to
October 1, 1984. Calculation of specific
yield for each pumping event varies
greatly. This variation is primarily due
to initiation of pumping before measured
leachate levels reach equilibrium within
the landfill. A calculation of specific
yield using the initial leachate level
(August 30, 1984) and the maximum leachate
level achieved after each of eight sequen-
tial pumping events yields the following
respective values: 13%, 11%, 19%, 19%,
22%, 24%, 19%, and 25%. The first two
lower values may be caused by an overesti-
mation of the soil volume drained, as the
engineering drawings indicate the 3.6-m
water level is near the upper clay cap
layer. Also, the variation may be caused
by heterogeneity of the bulk waste.
Subsequent measurements in October 1984
indicate specific yields of 23 to 27%.
These higher specific yields may indicate
additional water input from vertical
infiltration through the clay cap. Pump-
ing will cease during winter months, at
which time cover and bottom liner permea-
bilities will be estimated from the change
in water level.
*Product of Handar, Inc., Sunnyvale,
CA 94089.
221
-------
Leachate Removed in 25 m Amounts
245
250 255 260 265
Day of Year (September 1 984)
270
275
Figure 3. Water level at waste site as a function of pumping.
Concluding Remarks ACKNOWLEDGEMENTS
Based on our results, estimating the
total amount of drainable leachate in a
cell appears possible by choosing an
effective specific yield that accounts for
the volume of the waste that does not
contribute any drainable pore space (e.g.,
drums, transformers). Estimates of non-
drainable volume can only be obtained from
landfill operators and should be required
information at all new sites. From the
analysis of data collected at a commercial
waste landfill, the specific yield of bulk
waste ranges from 10 to 17%. Hourly
measurements of leachate levels at a
research waste landfill in New York State
indicated several days may be required for
water levels to reach equilibrium. A
specific yields of 25% was calculated for
an oil-contaminated sandy waste material
containing minimal solids.
Although the research described in
this article has been funded wholly or in
part by the U.S. Environmental Projection
Agency under a related services agreement
with the U.S. Department of Energy
(Contract DE-AC06-76RLO 1830) to Pacific
Northwest Laboratory, it has not been
subjected to Agency review and therefore
does not necessarily reflect the views of
the Agency, and no official endorsement
should be i nferred.
Special thanks to Mr. Jonathan
Herrmann, EPA project officer, and
Ray Cowen of the New York State Department
of Environmental Conservation for their
assistance in site selection and data
collection.
222
-------
REFERENCES
1. Anderslund, 0. B. and R. W. Laza,
1972. Permeability of high ash
papermill sludge, Journal of the
Sanitary Engineering Division - ASCE,
SA6 pp927-936.
2. Bond, F., 1980. Modeling the Fixed
FGD Sludge - Conesville, Ohio,
Phase 1. EPRI CS-1355, Electric
Power Research Institute, Palo Alto,
California.
3. Bond, F. W., C. R. Cole, and P. J.
Gutknecht, 1982. Unsaturated
Groundwater Flow Model (UNSAT1D)
Computer Code Manual. EPRI CS-2434,
Electric Power Research Institute,
Palo Alto, California.
4. Bresler, E. and G. Dagan, 1983.
Unsaturated flow in spatially
variable fields, 2. Application of
water flow models to various
fields. Water Resources Research,
19:2, pp421-428.
5. Cowherd, D. C., 1977. Geotechnical
characteristics of coal mine waste."
Proceedings of the Conference on
Geotechm'cal Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
pp384-406.
6. Cunningham, J. A., R. G. Lukas, and
T. C. Anderson, 1977. Impoundment of
fly ash and slag - a case study."
Proceedings of the Conference on Geo-
technical Practice for Disposal of
Solid Uaste Materials. American
Society of Civil Engineers, New York,
pp227-245.
7. Fungaroli, A. A., 1971. Pollution of
Subsurface Water by Sanitary Land-
fills. EPA/SW-12rg, U.S.
Environmental Protection Agency,
Washington, D.C.
8. Hagerty, D. J. and C. R. Ullrich,
1977. Engineering properties of FGD
sludges." Proceedings of the
Conference on Geotechnical Practice
for Disposal of Solid Waste
Materials. American Society of Civil
Engineers, New York, pp63-70.
9. Hardcastle, J. H., D. L. Mabes, and
R. E. Williams, 1977. "Physical
properties of Pb-Zn mine-process
waste. Proceedings of the Conference
on Geotechnical Practice for Disposal
of Solid Waste Materials. American
Society of Civil Engineers, New York,
pp!03-117.
10. Keshian, B., C. C. Ladd, and R. E.
Olson, 1977. Sedimentation-consoli-
dation behavior of phosphatic clays."
Proceedings of the Conference on
Geotechnical Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
ppl88-209.
11. Klute, A. and D. F. Herrmann, 1978.
Water Movement in Uranium Mill
Tailings Profiles - Final Report.
Technical Note ORP/LV-789/8, U.S.
Environmental Protection Agency,
Washington, D.C., 98pp.
12. Kulhawy, F. M., D. A. Sangrey, and C.
S. Grove, 1977. Geotechnical
behavior of solvay process waste."
Proceedings of the Conference on
Geotechnical Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
ppll8-135.
13. Leckie, J. 0., J. G. Pacey, and C.
Halvadakis, 1979. Landfill manage-
ment with moisture control. Journal
of the Environmental Engineering
Division - ASCE. EE2, pp337-355.
14. Maynard, T. R., 1977. Incinerator
residue disposal in Chicago. Pro-
ceedings of the Conference on Geo-
technical Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
pp773-792.
15. Nelson, R. W., P. R. Meyer, P. L.
Oberlander, S. C. Sneider, D. W.
Mayer, and A. E. Reisenauer, 1983.
Model Evaluation of Seepage from
Uranium Tailings Above and Below the
Water Table. NUREG/CR-3078,
PNL-4461, U.S. Nuclear Regulatory
Commission, Washington, D.C.
223
-------
16. Parker, D. G., S. I. Thornton, and C.
W. Cheng, 1977. Permeability of fly
ash stabilized soils. Proceedings of
the Conference on Geotechnical
Practice for Disposal of Solid Haste
Materials. American Society of Civil
Engineers, New York, pp63-70.
17. Pilz, J., 1982. The Saturated and
Unsaturated Shear Strength Spent Oil
Shale.Master's Thesis, Colorado
State University, Fort Collins,
Colorado.
18.
19.
20.
Remson, I., A. A. Fungaroli, and
A. W. Lawrence, 1968. Waster
movement in an unsaturated sanitary
landfill. Journal of the Sanitary
Engineering Division - ASCE, SAZ,
pp307-317.
Rovers, F. A. and G. J. Farquar,
1973. Infiltration and landfill
behavior. Journal of the Environ-
mental Engineering Division - ASCE,
EE5, pp671-691.
Sisson, J. B., A. H. Ferguson, and M.
Th. van Genuchten, 1980. Simple
method for predicting drainage from
field plots. Soil Science Society of
America Journal, 44, ppl!47-1152.
21. Somogyl, F. and D. H. Gray, 1977.
Engineering properties affecting
disposal of red muds. Proceedings of
the Conference on Geotechnical
Practice for Disposal of Solid Waste
Materials.
Engineers,
American
New York
Society
ppl-22.
of Civil
22. Straub, W. A. and D. R. Lynch, 1982.
Models of landfill leaching:
moisture flow and inorganic
strength." Journal of the
Environmental Engineering Division -
ASCE, EE2, pp233-2250.
23. Vick, S. G., 1977. Rehabilitation of
a gypsum tailings embankment."
Proceedings of the Conference on
Geotechnical Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
pp697-714.
24. Zimmerman, R. E., W. M. Chen, and A.
G. Franklin, 1977. Mathematical
model for solid waste settlement.
Proceedings of the Conference on
Geotechnical Practice for Disposal of
Solid Waste Materials. American
Society of Civil Engineers, New York,
pp210-227.
224
-------
PRECISION AND RELIABILITY OF LABORATORY PERMEABILITY MEASUREMENTS
John L. Bryant and Andrew Bodocsi
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
A preliminary collection of laboratory permeability test data was gathered from nine
sources to create a data bank suitable for statistical analysis. The collected data also
yielded a survey of the most commonly used permeameters and testing methods for clay liner
permeabilities.
This paper describes the methods used to collect and organize the data. Tentative
findings are discussed. Consideration is given to the degree of variability found in
replicated permeability tests, the question of sample equilibration, parameters of sample
preparation which may lead to sizable errors in permeability determinations, and the amount
of variability of permeabilities that may be found in location-to-location sampling from
within a source of liner material.
INTRODUCTION
Clay liners have been used for many
years as barriers below hazardous waste
disposal sites. The currently accepted
permeability requirement for these liners
is 10-7 cm/sec (4). The permeability of
liner materials is usually determined in
geotechnical laboratories under varied
conditions and by various testing methods
as, for example, by use of fixed wall
versus triaxial permeameters. Conse-
quently, there is uncertainty about the
precision and reliability of these
measurements.
PURPOSE
The objectives of the project were:
a. to collect a limited set of
available data on permeability measurements
and to describe the most commonly used
permeability procedures;
b. to analyze the collected data in
order to determine the degree of
variability likely to be found among the
results of permeability testing, and to
identify sources of errors in the testing
procedures and estimate their significance;
c. to recommend a statistical
protocol for future clay liner permeability
tests.
APPROACH
Selection of Data Sources
Permeability data were collected from
laboratories that were active either in
research or design of clay liners. Nine
data sources were identified, based on
established contacts, past and on-going
research, reputation, and publications.
These included four EPA contract sources,
two government laboratories, two private
consulting firms, and one independent
university source.
Collection of Preliminary Data
Data sources were contacted to determine
the availability of permeability data. A
request was then made to each of the
selected sources to provide all suitable
permeability data, clay characterization
data and test specification data.
As of October 1984, complete data sets
have been received from six of the sources;
a partial set from one source; and only
published reports and theses, but no raw
data, from two. Four sets of data have
been reduced for computer manipulation, and
have been analyzed. Data reduction is in
progress for the remaining data sets.
Data Organization and Reduction
Two types of data sheets were used to
225
-------
organize the technical information.
For each clay for which permeability
data had been received, a "Clay Type Sheet"
was filled out, containing a description of
the clay and information on its origin,
batch number, Atterberg limits,
classification, texture, sand and silt
content, Proctor density and optimum water
content, pH, ion exchange capacity, and
specific gravity.
For each permeability test for which
data had been received, a "Permeability
Test Data Sheet" was filled out, containing
information on the type of clay used,
initial density, water content, void ratio
and degree of saturation, type of
permeameter used, type of permeability
testing apparatus used, types of permeants
used and their characteristics, and
criteria for terminating the tests.
Attached to each Permeability Test
Data Sheet were the corresponding raw
permeability test data.
After the data had been organized,
they were filed on the computer and were
available for analysis.
DISCUSSION OF RESULTS
The following points will be
discussed, and examples given below.
a. Although it is somewhat situation-
specific, the variability of permeability
test results using standard aqueous
leachate, standardized compaction, and
testing of replicate samples, taken from a
uniformly moisterized batch of homogeneous
clay soil, is reasonably small relative to
other sources of error discussed below.
b. Based on the relatively few
permeability tests available, which employ
standard leachate over a very large number
of pore volumes, there is evidence that
equilibrium may be considerably more
difficult to establish than is commonly
supposed. Furthermore, the error in a
permeability determination which may be
caused by premature termination may be
large relative to the degree of
within-batch variability discussed in point
a). This implies that the length of
testing time typically employed by
geotechnical firms may be insufficient.
c. The permeability of a soil sample
is extremely sensitive to the parameters of
its preparation, such as the molding water
content at which it is compacted. Minor
variations in these parameters can cause
changes in permeability which are large
relative to the accuracy of the test
itself. Thus, errors introduced during the
moisturization step of a permeability
determination will not be small relative to
the within-batch variability discussed in
point a), at least in tests conducted at
levels of the parameters where the
sensitivity is great. Moreover, in the
actual construction of a liner, molding
water content and compactive effort cannot
be strictly controlled. Under these
conditions, it is meaningless to perform a
single test at a nominally specified
molding water content and standard
compactive effort. Instead, the bounds
within which molding water content and
compactive effort can be guaranteed to be
maintained during the construction of the
liner must be determined. Then, the
proposed material should be tested at
perhaps three molding water contents,
spanning the range which can be maintained
during construction, and conducted at the
least level of compactive effort that can
be guaranteed during liner application.
d. It is important to assess the
degree of uncertainty of permeability
determinations due to the nature of the
test itself; it is equally important to
estimate the amount of variability of
permeabilities that might be expected to
occur in location-to-location sampling from
within a proposed source of liner material,
in order to judge the amount of sampling
required to ascertain the overall
acceptability of the soil within the
source. Data from geotechnical firms,
which provide information concerning this
location-to-location component of
variation, indicate that permeabilities of
samples taken from various positions within
a site exhibit far more variability than do
the replicate tests on uniform clays
observed in the research setting, even when
the samples are relatively homogeneous with
respect to their Uniform Soil Classifi-
cations, Atterberg limits and grain size
information.
Further discussion of these points is
divided into three parts: Errors and
Variation Associated with Compaction and
Testing, Errors and Variation Associated
with Moisturization and Molding Water
Content, and Location-to-Location
Variability.
Errors and Variation Associated with
Compaction and TeVting
To observe the degree of variation
that may be contributed during sample
compaction and subsequent testing, samples
226
-------
of a uniform clay soil which are
identically prepared and moisturized are
required. An example of such data comes
from twelve rigid cell tests of Faceville
clay conducted by Peirce (5) as part of a
comparative study of the effects of chemical
leachates. The soil required for the tests
was moisturized to a molding water content
of 23.7% (optimum moisture content was 23%)
in a common batch, then compacted into cells
and permeated with water (0.01 N CaS04) for
321.75 hours, or on average 2.0 pore
volumes. Then chemical leachates were
introduced into the cells and the tests were
continued.
Figure 1 shows the base 10 logarithm
of the permeability of each sample over the
last 100 hours of water permeation. These
permeabilities were averaged over the
eleven readings made on each sample during
that period. Logs of the permeabilities are
plotted rather than raw permeabilities
since experience has shown that the degree
of variability of log permeabilities is
more consistent over different clay types
than is the degree of variability of the
untransformed permeabilities. This is not
surprising, since one would expect the
standard deviation of permeabilities to be
roughly proportional to their mean, so that
in terms of raw permeabil ities, less
permeable soils would exhibit less
variability than more permeable ones.
Working with logged data compensates for
this effect.
The twelve log permeabilities range
from about -7.7 to -7.5 (2.0xlO~8 < k <
3.2xlO'8) with an average of -7.59 (2.6 x
10"°) and a standard deviation of 0.061.
Figure 2 shows the behavior of one of
the twelve tests, and is typical of the
others. After brief instability, the
permeabilities became fairly consistent at
about 1/2 pore volume of flow. For the
particular test shown, permeabilities appear
to be increasing slightly during the last
100 hours of the test, but this effect seems
quite small. A statistical test to detect
systematic time trends across the twelve
different tests is in fact significant (p <
0.01), but this formal result is practically
meaningless, because the tests were run
concurrently in the same lab and are jointly
subject to minor changes in environmental
conditions such as temperature and temporary
fluctuations in the main pressure source.
Furthermore, the estimated magnitude of the
trend was only on the order of a 0.02
increase in log permeability, which could
be attributed to such minor causes as a 2°C
change in temperature. From a practical
point of view, the tests appear to be in
equilibrium. Based on the data reviewed to
date, had these tests been conducted by a
geotechnical firm with the intention of
determining the permeability of a soil, they
would have been terminated much sooner than
the 321.75 hours allowed by Peirce.
The results of Figure 1 give the
impression that replicated permeability
tests from within a uniformly moisturized
batch should not exhibit a great deal of
variability, since the range of twelve such
replications was only about 0.2 of an order
of magnitude. If, however, equilibrium is
not achieved, additional error is
contributed which may be large in comparison
to this variability, as will be seen below.
At the conclusion of 321.75 hours of
water permeation, chemical leachates were
added. Figure 3 shows the behavior of the
tests. Each test is labeled with an
identification number (21-28; 35-38). White
dots at 271.75 hours represent the final
water permeabilities analyzed above. White
dots at 1020.9 hours represent the final
permeabilities averaged over the last 100
hours of testing (970.9 to 1070.9 hours).
Tests 36 and 38 were continued for 2391.33
hours. Black dots represent 100 hour
average permeabilities (technically, the
points plotted with black dots are medians
rather than means, because occasionally an
isolated permeability reading in the raw
data was extremely high or low, having an
undue influence on the average. This did
not occur during the final intervals marked
by white dots).
The first impression given by Figure 3
is that addition of chemical leachates
affected permeability, causing an average
increase of about 0.44 orders of magnitude,
or 275%. In fact, this is not the case;
four of the cells were used as controls with
water continued as permeant, and these tests
underwent permeability increases similar to
the others. This becomes clearer from
Figure 4, which shows the increase in log
permeability in each test.
Statistical tests to detect an effect
of the chemical leachates relative to water,
which are based on the differences shown in
Figure 4, do not indicate that the chemicals
have increased permeability. This is not
too surprising, because in this experiment
the chemicals were so diluted that many pore
volumes would have to be passed before all
soil in the cells could be chemically
affected.
Possible reasons for the observed
increase are:
a. a change in laboratory environment:
227
-------
i e o o o oo o1 o o o o o1
-7.7 -7.6 -7.5 L
2-0x 108 2.5 x 108 3.2 x 10'8k(cm/s)
Figure 1. Permeabilities of Faceville Clay after 321.75 hours on water
(0.01N CaS04).
-7.4 r
-7.5
-7,6
•*•
50 100 (150 200 250 f300 MRS
1 PV 2 PV
Figure 2. Variation in the permeability of Faceville Clay during
321.75 hours of testing with water (0.01N
228
-------
Logw(k)
-7.0
-7.1
-7.2
-7.3
-7.4
-7.5
-7.6
-7.7
£^£t-.:--.^-.---.:„:
. •-•
RIGID WALL SAMPLES
FeCI3 Tests 21-24
Ni(NO3)2 Tests 25-28
H2O Tests 35-38
100T500 .
PV=2.0 PV=5.4
t
1000!
PVM3.2
1500
HOURS
Figure 3. Variation in the permeability of Faceville Clay in
rigid mold permeameters.
.50
Increase
in
Log1Q(k)
.40
.30
o 24
o 22 o27 o35
o26 o36
o25
o21 °38
o 23
o37
o28
1 1 1
O ^
-T° ^ °M -J
0 ro X O
0) O GC
n •» i
— &. I —
* 0
O
Figure 4. Increase in Log^(k) between 321.75 and 1070.0 hours
of flow for Faceville Clay.
229
-------
This seems extremely unlikely. The effect
of temperature fluctuation on viscosity
could not have caused an effect of the
magnitude observed. Furthermore, pressure
gauges were recalibrated each day.
b. sample disturbance caused by
depressurization when permeant was switched
from water to chemical: This also seems
unlikely, because no switchover occurred in
the control tests. Further, careful
inspection of individual graphs for each
test indicate that perhaps the increase
began slightly before switchover;
c. slow approach to complete
saturation, accompanied by increased
permeability: Peirce conducted two triaxial
cell tests on Faceville clay in addition to
the rigid-wall tests discussed here, one
using Fed3 as a permeant and the other
Ni(N03)2. Vacuum was applied to these two
tests to remove air. Formal statistical
comparison between triaxial and rigid-wall
results is ambiguous, because the tests were
run on separate batches of moisturized clay,
were prepared by different technicians and
were not run concurrently. In addition, the
triaxial specimens were compacted
considerably more densely than were the
rigid specimens. Nevertheless, an informal
comparison is useful. Both triaxial tests
start with log permeabilities in the
vicinity of -7.1 and decrease to virtually
equal equilibria at about -7.29, as shown in
Figure 5. There have been few triaxial
tests in the data that were reviewed to
date, but in most cases, their behavior in
coming to equilibrium appears to be better
than is the case with rigid cells.
d. time-dependent transition to a more
flocculated state, associated with
thixotropic hardening: Mitchell, et al.,
(3) have demonstrated that aging of certain
clays prior to permeability testing may lead
to increases in permeability consistent with
the results seen in this study. Such aging
could conceivably occur during the test
period.
e. side-wall erosion: This is a
distinct possibility in the rigid-wall
tests, but it does not explain why their
initial permeabilities start well below
those of the triaxial tests and equilibrate
at roughly the same level.
The failure of rigid-wall tests to
equilibrate rapidly has been noted in other
data sets as well, including two of four
control tests of Brown and Anderson (1) in
their comparative study of the effects of
organic leachates, and to a lesser degree,
in other tests conducted by Peirce. There
are two obvious implications:
1. In comparative studies, it is
necessary to include control tests in order
to provide statistical confirmation of the
effect of chemical permeants relative to
water.
2. In settings in which the goal is
the estimation of the permeability of a
soil, the number of pore volumes of flow
often seen is inadequate.
In the rigid-wall tests of Faceville
clay, the average log permeability increased
from -7.59 to -7.15, or over 7 times the
initial standard deviation, even though the
behavior of the tests was reasonably stable
over the first two pore volumes. Even after
as many as 10 pore volumes, the equilibrium
of several of the tests is questionable.
Furthermore, the variability of the
determinations at 1070.9 hours is, from
Figure 3, somewhat greater than at 321.75
hours, having a standard deviation of .097,
although arguably this variability among
tests would diminish somewhat as all tests
reach equilibrium. This standard deviation
gives a measure of the dispersion of
permeability determinations which is
relevant under the assumption that the
introduction of permeants had little or no
effect relative to the standard leachate.
To aid in interpretation, a normal
probability distribution contains about 95%
of its probability within two standard
deviations of its mean. Thus, under the
assumptions that the permeability
determinations are at least roughly normally
distributed on the log scale, that the tests
are at or near equilibrium, and that the
sample standard deviation of 0.097 provides
a reasonably precise estimate of the true
standard deviation, this would imply that
within-batch errors would generally be
within ± 0.2 of an order of magnitude.
Log permeabilities are moderately
correlated with void ratio (.82 correlation
at 321.75 hours, .79 correlation at 1070.9
hours; p < .01 in both cases) as shown in
Figure 6. This indicates that variation in
compactive effort may explain a significant
portion of the variability observed in the
permeabilities of the twelve tests. On the
other hand, other causes could lead to
differences in void ratio, such as slight
inhomogeneity of the soil itself.
Errors and Variation Associated with
Moisturization and Molding Water Content
The sources of error discussed in the
previous section are associated with the
processes of sample compaction and the
actual permeability test itself. Additional
sources of error which arise during the
230
-------
FLEXIBLE-WALL
SAMPLES
RIGID WALL SAMPLES
FeCI3 Tests 21-24
Ni(NO3)2 Tests 25-28
H2O Tests 35-38
100
500
1000
1500 HOURS
Figure 5. Permeabilities of Faceville Clay - comparison of rigid-wall and
flexible-waH samples.
-7.5
(0
l_
3
o
10
CM
CO
**
to
o
-7.6
-7.7
0 /
.74 .76 .78 .80 .82 VOID RATIO
Figure 6. Permeability of Faceville Clay at 321.75 hours versus void ratio.
231
-------
moisturization of soil do not enter into the
variability of the twelve rigid-wall tests of
Faceville clay, because the soil required
for these tests was moisturized in a common
batch.
The molding water content at which a
clay soil is compacted may have a profound
effect on its permeability. This is
illustrated in Figure 7, which shows the
sensitivity of the water (0.01 N CaSO^
permeability of Calcareous Smectite to
molding water content. These data are due
to Brown and Anderson (1). In the vicinity
of the optimum water content, the effect is
acute. Permeability falls by roughly three
orders of magnitude as molding water content
increases from 15.9% to 23.5%, or on average
about 0.4 of an order of magnitude per 1%
water content. A second example, shown in
Figure 8, (2) is similar in shape to data for
other clays examined by Brown and Anderson.
Roberts (6) comments that the curve of Figure
8 is typical.
The steepness of these curves in the
vicinity of the optimum water content is
directly related to the degree to which a
permeability determination may be in error,
since the determination of molding water
content is itself subject to error. As an
example, consider a situation in which it is
desired to estimate the permeabilty of the
soil in Figure 8 at a molding water content
of 19% (roughly 1.5% above optimum) and
suppose for the sake of argument that the
permeability-molding water content
relationship shown in Figure 8 is exact.
The slope of the curve at that point is
roughly -0.46 orders of magnitude per 1%. It
follows that if the actual water content
used in the test differs by as little as
0.5% from its nominal value of 19%, an error
as great as 0.23 orders of magnitude (or
since 10-23 = 1.70, an error of 70%) would be
incurred in the permeability estimate, even
if the permeability test itself were
perfectly accurate. While an error of this
magnitude may in certain circumstances be
considered to be acceptable, it is not
negligible relative to the variability of
permeability measurements made on samples
compacted from a common batch of moisturized
clay. Thus, using the Faceville clay tested
by Peirce as an example, it is seen from
Figure 3 that all twelve final log
permeabilities lie within the interval
(-7.0,-7.4), whose half-width is less than
the error incurred by a 0.5% error in molding
water content in the hypothetical example.
This example shows that (except for
tests conducted near that molding water
content required to minimize permeability)
errors in water content determination may
play a significant role in explaining
variability of permeability measurements
obtained on independently moisturized
specimens.
Location-to-Location Vari ability
The sources of error and variation
discussed above relate to inconsistency of
sample preparation and test methods as
applied to essentially identical soil
samples. A separate component of variance
which warrants consideration is due, not to
errors in testing, but rather to
inhomogeneity of soils to the degree which
may be expected to occur within a potential
source of liner material.
The analysis to be described in this
Section was conducted to obtain a rough
estimate of the total variability of
permeability tests performed on samples
selected from various locations within a
proposed source of liner material. Such an
estimate is needed, for example, to provide
judgment on the amount of sampling required
in order to obtain sufficiently precise
estimates of such characteristics as the
median permeability throughout the proposed
source or to perform statistical hypothesis
tests concerning its suitability for use as
a liner material.
The amount of variability found among
such test results should depend rather
strongly upon the uniformity of the soil
source in question. Thus, in this analysis
we considered the variability of permea-
bility determinations taken from within soil
sources which were geographically contig-
uous, each of which was confined to a single
Unified Soil Classification, with Atterberg
limits and grain size data restricted to
relatively narrow bounds. Analysis of
variance (8) was used to estimate the amount
of variability within each source.
Data were provided by an independent
soil testing firm. They consist of 23
permeability tests of samples collected from
four different sites, all of which employed
rigid-wall permeameters and constant head
methods. Water was used exclusively as the
permeant. There was no replication of
testing on soils from the same location.
That is, all samples were taken from
different bag samples within the four sites.
The number of pore volumes passed
during the tests varied from less than .05
pore volume to more than one pore volume,
232
-------
-5
-6
-7
-8
-9
15 17 19 21 23 25
MOLDING WATER CONTENT (%)
(OPTIMUM WATER CONTENT: 21.5%)
Figure 7. Permeability of calcareous smectite versus molding water content.
1 x 1Cf5
5 x 1fJ6
o
V
CO
io~6
o _7
~ 5 x 107
S 1 x 1fJ7
LLJ
5 x
DC
UJ
O.
OPTIMUM WATER CONTENT
STATIC COMPACTION
KNEADING COMPACTION
15 17 19 21 23 25 27
MOLDING WATER CONTENT (%)
Figure 8. Permeability versus molding water content (Mitchell, 1976).
233
-------
TABLE I
UNIFIED SOIL
SITE CLASSIFICATION
SI CL
S2 CL
S2 CL-CM
S3 CL
S3 CL-CM
S3 SM-SC
S4 CL
NUMBER OF
SPECIMENS
IN GROUP
5
2
6
2
2
2
_4
23
RANGE IN
LIQUID
LIMIT
25-33
30-32
18-20
35-41
18-20
19-20
24-33
RANGE IN
PLASTICITY
INDEX
8-12
15-17
5-6
18-23
4-6
5-6
10-17
RANGE IN
PERCENT
FINES
67-90
64-72
52-58
80-80
50-51
37-49
65-68
AVERAGE
LOGi0
PERMEAB"
ILITY
-5.97
-8.44
-7.49
-8.14
-7.
-6.
64
55
-7.82
with perhaps one-quarter pore volume being
typical.
Within the four sites, samples may be
grouped according to their Unified Soil
Classification as shown in Table I.
The permeability of each specimen was
computed based on its total flow. Then, the
within-group variance of log
permeabilities, pooled over the seven soil
groups of Table I, was computed to provide
a rough numerical measure of the degree of
variability that may be expected in
sampling at different locations from within
soils of uniformity comparable to those
summarized in Table I.
The within-group variance was 0.22,
corresponding to a standard deviation of
/.22 = 0.47. For purposes of interpre-
tation of this result, it may be compared
to the standard deviation of 0.097
exhibited by the log permeabilities of
Peirce's rigid-wall tests on Faceville clay
over the last 100 hours of his tests. Thus
on a logarithmic scale, the dispersion of
the permeabilities from location-to-
location within the soil groups is over
four times as great as that exhibited by
the final permeabilities graphed in Figure
3.
Using analysis of covariance (8) it is
possible to correlate a portion of the
variation of permeabilities within the soil
groups to factors having to do with
variation of the parameters under which the
tests were carried out (e.g., molding water
content and void ratio) although sufficient
data are not available to do this in a very
rigorous fashion. After adjustment, the
within-group standard deviation is 0.39.
It may be argued that these tests ai~e
not comparable to Peirce's in that the
volumes of permeant passed were not
sufficient to establish equilibrium.
Nevertheless, these data indicate that
considerable variability in permeabilities
is to be expected in location-to-location
sampling, even within soil sources which are
reasonably uniform with respect to Atterberg
limits and grain size information. Further,
the point is made that the level of
variability that might be observed in data
generated in the research environment, with
uniform test clays, may have little
relationship to that observed in situations
where location-to-location variability of
permeability within a proposed source of
liner material need be considered.
TENTATIVE CONCLUSIONS
Even when standard leachate (0.01N
CaS04) is used as permeant liquid, the
establishment of equilibrium may require
several pore volumes of flow. In Peirce's
Faceville series, permeabilities nearly
tripled after having appeared to be nearly
stable over the first two pore volumes.
Rigid permeameters appear to be less
inclined to rapid equilibration than do
triaxial permeameters.
The variability of properly
equilibrated permeability tests, which are
replicated under uniform conditions, appears
to be reasonably small in cases where
standard leachate is used as permeant
liquid. The amount of between-test
variability exhibited in the Faceville
series is reasonably representative of tests
which are replicated within uniformly
moisturized batches of homogeneous clay
soil. Replication on independently
moisturized batches involves an additional
component of variance, whose magnitude
depends strongly on the degree to which
molding water content varies, the slope of
the permeability-water content curve in the
neighborhood within which tests are
conducted, and the homogeneity of the soil
from batch-to-batch.
For permeants other than standard, the
variability of test results appears to be
234
-------
quite specific to the particular clay-
chemical combination under investigation.
The permeability of a specimen is very
sensitive to the water content at which it
is molded. Unless a test is conducted at a
density and moisture content which closely
approximates actual field conditions, the
results of the test may completely
misrepresent future performance of the
liner material.
Location-to-location variability of
the permeabilities of samples collected
from within a source of liner material may
be considerable, even when such samples
appear to be similar in terms of Atterberg
limits and grain size information.
ACKNOWLEDGEMENT
The authors wish to acknowledge the
assistance of Jonathan Herrmann, the
Technical Project Monitor, Walter Grube of
USEPA, and Richard McCandless of the
University of Cincinnati at the Center Hill
Research Laboratory. They especially wish
to thank all individuals and organizations
who contributed the necessary permeability
data and test information.
REFERENCES
1. Brown, K.W., and Anderson, D.C. Effects
of Organic Solvents on the Permeability
of Clay Soils. EPA Grant No. R
806825010, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1983.
2. Mitchell, O.K. Fundamentals erf SoiJ
Behavior. John Wiley and Sons, Inc.,
New York, New York, 1976.
3. Mitchell, O.K., et al., "Permeability
of Compacted Clay," Journal of the
Soil Mechanics and Foundations Division,
91, No. SM4, July 1965.
4. Public Law #98-616, November 9, 1984.
Hazardous and Solid Waste Amendment of
1984, Section 3004, Paragraph "0":
Minimum Technological Requirements.
5. Peirce, J. Jeffrey. Effects of
Inorganic Leachates on Clay Liner
Permeability. (Unpublished Report)
Department of Civil and Environmental
Engineering, Duke University, Durham,
North Carolina.
6. Roberts. D.W. Soil Properties,
C1assification, and Hydraulic
Conductivity Testing, Draft Technical
Resource Document for Public Comment
(SW-925). USEPA, Cincinnati, Ohio,
March 1984.
7. Walpole, Ronald E. and Myers, Raymond H.
Probability and Statistics for Engineers
and Scientists, 2nd ed. Macmillan
Publishing Co., Inc., New York, New
York, 1978.
8. Winer, B. J. Statistical Principles in
Experimental Design, 2nd ed., McGraw-
Hill, New York, New York, 1971.
235
-------
SIMULATING LANDFILL COVER SUBSIDENCE
Harry J. Sterling
Environmental Services
Daniel International Corporation
Greenville, South Carolina 29602
Michael C. Ronayne
ATEC Associates, Inc.
Louisville, Kentucky 40205
ABSTRACT
The subsidence response of a landfill cover system's clay layer was modeled at
reduced scale using a geotechnical centrifuge. The prototype structures for the clay
model layers were the multi-layered covers presently being tested as part of an EPA-
sponsored investigation of hydrologic isolation. These structures consisted of three 2 ft
(0.61 m) layers, with a compacted clay layer at the bottom, then sand, and finally top
soil. The clay layer was modeled in the laboratory at 1/48, 1/32, 1/24, and 1/16 scales,
and was surcharged with lead shot to simulate the overburden due to the other two layers.
The scope of the work was limited to clay layers subjected to "slump", a type of subsi-
dence characterized by localized void spaces beneath the covers. Data were collected to
investigate the effects of cavity width and clay moisture content. Qualitative descrip-
tions of the failure mechanisms were developed.
Clay layer behavior was very dependent upon both the moisture content and cavity
width. Field cavity widths of 4 ft (1.2 m) were successfully bridged by all models tested
at that width, but failures and extensive cracking occurred for many models simulating
both 8 ft (2.4 m) and 12 ft (3.7 m) field cavities. Collapse of the samples at the 8 ft
width occurred at optimum moisture content (which also was the plastic limit of the clay
soil), while sample collapse at the 12 ft width occurred at moisture contents 3 percent
below optimum. Qualitatively, soil cracking features were similar to those observed and
reported at field sites investigated by other workers.
The centrifuge modeling procedure has promise as an investigative tool to determine
the physical mechanisms of subsidence and appears capable of testing a wide variety of
different configurations for subsidence response in order to determine optimum landfill
cover designs.
INTRODUCTION
The most common and often the most
damaging failure mechanism that has been
observed in low-level radioactive waste
landfills is the inability of cover systems
to adjust to short- and long-term subsi-
dence of the underlying waste and backfill
materials (5). Volume reduction and subsi-
dence of the trench fill can result from
the following processes:
• collapse of structural bridges formed
by random dumping of waste containers.
• collapse of voids in the backfill soil
that remain after poor soil compaction.
• filtering of backfill soil into
unfilled void spaces between waste
containers.
• rupturing, degradation and collapse of
waste containers.
• percolation of liquid wastes from
ruptured containers.
236
-------
• dissolution of solid wastes in rup-
tured containers.
• chemical degradation and consolidation
of waste materials.
t pumping of leachate from disposal
trenches.
• consolidation of backfill soil.
Figure 1 illustrates the processes
leading to localized cover subsidence (pot-
holes and slumping). The process of par-
ticular interest in this study is slumping,
steps (d) and (e) of Figure 1, in which the
trench cover bridges a void space. Col-
lapse of the cover will result when the
void space reaches sufficient size. Subsi-
dence can result in a breach of cover
integrity and can jeopardize functional
performance of the cover system. The most
damaging effects of subsidence are signif-
icantly increased infiltration rates due to
ponding and crack formation, direct expo-
sure of the waste cell to precipitation and
surface runoff, release of contaminants,
and accelerated bio-degradation of the
buried wastes (5).
Concern has been expressed that
eventual subsidence will compromise the
hydrologic integrity of landfills
containing hazardous waste components, with
subsequent leaching of toxics into
groundwater regimes. Since some hazardous
waste landfills are known to contain
materials in bulk form, in metal or paper
drums, and possibly in other decomposable
containers; and because it is known that
the interstices among these containers are
not always completely filled with packed
earth, subsidence may occur after some time
has elapsed and the landfilled material
settles and consolidates. The cover system
over a completed landfill forms the barrier
to keep out infiltrating precipitation.
Breaches in this cover may occur if
subsidence of the underlying material
occurs, and causes the cover system to
change in structure from its original
construction. The degree of subsidence and
the corresponding amount of cover damage
and detriment to the impervious cover
system greatly depends on the structural
competence of the underlying landfilled
material over time. This latter factor has
not yet been very well described for modern
hazardous waste landfill structures.
PURPOSE
The main force which causes cover
slumping is the soil cover self-weight.
Centrifuge model testing was selected for
this study because it is the only labora-
tory procedure which can reproduce the
effects of gravity in a reduced-scale model
(1). The use of centrifuge soil modeling
is becoming an accepted method for veri-
fying laboratory soil parameters, geotech-
nical design assumptions and safety factors
(6). It has also been applied to complex
problems of soil response, for which accept-
able solutions have yet to be developed (1).
The basic concept of centrifuge model-
ing is to create a scale model that is
similar in geometry, material properties,
and boundary conditions, to the full-scale
prototype, and to subject the scale model
to a radial acceleration via centrifugation
such that the increase in self-weight
stresses in the model matches those at
corresponding points in the prototype. If
the model dimensions are scaled by a factor
1/N as compared to the prototype dimen-
sions, then an acceleration field of N
times the acceleration of gravity is re-
quired for stress similarity between model
and prototype (2).
Centrifuge testing can also satisfy
geometric, material, boundary loading,
strain, transient fluid flow, and other
similarity conditions in a small-scale
physical model. These conditions must be
verified by experimental evidence, which
may be from prototype physical tests, ana-
lytical or computer models, or centrifuge
model tests at various scales (2).
APPROACH
The behavior of soil structures is
influenced by many complex factors such as:
0 solid-liquid-gas phase relationships and
interactions.
• particle mineralogy, chemistry, size,
grading, shape, and structure.
• pore liquid and gas composition, distri-
bution, and pressure.
• previous stress history.
• boundary conditions.
• construction practices.
237
-------
(a)
SURFACE RUNOFF
;••';* .'.'?2-£Xff:Wt'-'!-'- "••"-*
- ... !,-.-»^..> ., BACKFILL
;«£v
-------
PROTOTYPE COVER
LABORATORY MODEL
PLEXIGLASS WALL
FLANGE
PLAN
SAMPLE CONTAINER
LVOT
AS3E
PLASTER MOLD
SOLENOID ASSEMBLY
CROSS SECTION
Cross Section
20 mil PVC
Figure 2. Plan and cross section of the field (prototype) covers and
the laboratory (centrifuge) models.
-------
The unknown levels of uncertainty and
variability associated with these and other
factors can make it difficult, and in some
cases impossible, to use either theoretical
or empirical methods for predicting field
soil behavior with an acceptable degree of
confidence. As a result, it is often
necessary to test geotechnical designs to
verify that they will function as intended.
The most reliable verification method is to
construct full-scale structures for test-
ing. However, the use of full-scale test-
ing is limited due to cost, time, and ma-
terial handling considerations. Due to
these limitations, scale model testing is
often used for predicting unknown modes of
soil behavior and for verifying geotech-
nical design concepts (6).
Side-by-side comparisons of the proto-
type (field) covers and the laboratory
(centrifuge) models are shown in Figure 2.
The auger access tubes will allow removal
of sand from beneath the field cover for
future full-scale subsidence testing. The
plots were instrumented to monitor soil
moisture using tensiometers, resistance
(gypsum) blocks, thermocouples, tipping
buckets and a weather station. Hydro-
logical isolation of the plots from the
surrounding ground is provided by a PVC
liner underlying the whole cover systein.
The field covers are more fully described
by Warner et al. (7).
The centrifuge used for this study was
a horizontal rotating arm mounted on a
vertical drive shaft powered by an electric
motor. The centrifuge consisted of six
basic sub-units, which were 1) sample con-
tainer and counterweight, 2) rotating arm,
3) power supply, transfer and control sys-
tem, 4) structural frame, 5) sand bag wall
and protective housing, and 6) data collec-
tion and slip-ring assembly.
Figure 3 shows a cross-section of the
centrifuge. The 0.61 m (2 ft) thick proto-
type clay layer was modeled in the centri-
fuge using either a 1.27 cm (0.5 in.), 1.90
cm (0.75 in.), 2.54 cm (1.0 in.), or 3.81
cm (1.5 in.) thick soil sample accelerated
to 48, 32, 24, or 16 g's, respectively.
The inner diameter of the cylindrical con-
tainer in which the soil sample was placed
was 29.2 cm (11.5 in.). Plaster molds were
used to simulate cavities beneath the soil
samples. Several were constructed to be
equivalent to 1.2 m (4 ft), 2.4 m (8 ft),
and 3.7 m (12 ft) diameter cavities in the
field. Lead shot was used above the clay
layer to simulate a top soil overburden.
Lead shot was also used to initially fill
the cavity beneath the layer. A solenoid
valve at the base of the sample container
was activated during the test to release
this shot and thus form the cavity. The
models were circular for radial symmetry.
1.78 m
SANDBAG WALL
SLIP-RING ASSEMBLY
•SHEETMETAL
HOUSING
SAMPLE CONTAINER
Figure 3. Cross section of centrifuge.
240
-------
Surface elevation measurements were
made with a dial-gauge. A template was
temporarily attached to the test capsule,
and used to locate the points to be meas-
ured. It was made from a 38 cm (15 in.)
square sheet of plexiglas, with dial-gauge
guide holes drilled on 1.27 cm (0.5 in.)
centers in a 23 x 23 grid pattern. .Each
hole was identified by a set of reference
coordinates.
During each test, subsidence was re-
corded by monitoring the mid-point deflec-
tion of the soil layer using a probe. A
linear voltage displacement transducer
(LVDT) mounted on the top plate of the
sample container was used to convert the
movement of a surface displacement probe
into a voltage that was linked to a strip
chart recorder via the slip ring assembly.
Physical connection between the LVDT and
the probe was made with a balanced lever,
which minimized the additional load ap-
plied to the soil surface. The elapsed
time and the LVDT output, as read from a
voltmeter, were written down on the chart
record at regular time intervals during the
test and when rapid changes in deformation
were observed.
The data recorded for each test are
listed below:
• model scale factor
• centrifuge speed
• radius to the sample container's
centroid
• cavity geometry
• number of blows for clay layer
compaction
• clay layer weight
• clay layer moisture content values
• lead shot surcharge weight
• initial and subsided elevations of
points on the surface of the clay layer
• time allowed for pore pressure
dissipation
• midpoint subsidence of the clay layer
• photographs of the subsided clay layer
• qualitative description of the test
These parameters were selected based
on published centrifugal modeling studies
(1, 6).
PROBLEMS ENCOUNTERED
Construction of the experimental appa-
ratus and development of the model testing
procedure resulted in a number of problems.
The two most significant problems are
discussed.
The first problem was finding a labo-
ratory testing technique suitable for simu-
lating subsidence. There are generally two
approaches to soil model testing which are
commonly used to fulfill similarity re-
quirements between a prototype and a model.
They include:
1) Build a model at a reduced scale
using the same material as in the prototype
and subject it to the same forces experi-
enced by the prototype, or
2) Build a model at a reduced scale
using a substitute material that is propor-
tionally weaker than the prototype and
subject it to forces that are propor-
tionally weaker than those experienced by
the prototype.
In general, it is easier to reproduce
the forces that act on one body in another
body than to reproduce soil behavior in a
different material. The second technique
has been applied when a known strength
parameter can be assumed to govern soil
behavior. For example, if a soil can be
considered to behave elastically, a model
could be built from material having a pro-
portionally lower elastic modulus than the
prototype material. The forces applied to
the model should be reduced accordingly
(4). While geometric, boundary loading and
stress similarity can be readily approached
with this technique, the actual soil be-
havior is often too uncertain and/or too
complex for material similarity to be ful-
filled. As a result, the first technique
is most commonly used to model soil be-
havior (1), and was the reason for the
selection of the centrifuge technique.
The second major problem was inter-
preting the experimental results for pre-
diction of the prototype cover system's
subsidence behavior. A common procedure
used in geotechnical centrifuge modeling to
determine the relationship between dif-
ferent variables and the behavior of the
soil sample is the "Modeling of Models"
method. With this method, models with
different physical dimensions are accel-
erated so that they are diinensionally con-
sistent. Variability in the behavior of
the samples can then be observed and used
to better predict the behavior of the pro-
totype structure. This second problem was
addressed as follows.
241
-------
Four clay-layer models were con-
structed and tested in the centrifuge with
different thicknesses, but were otherwise
dimensionally consistent. All four had
moisture contents of approximately 27 per-
cent, equal compaction intensity, and were
bridging a 2.4 m (8 ft) equivalent field-
size cavity. The model thicknesses were
1.3 cm (0.5 in.), 1.9 cm (0.75 in.), 2.5 cm
(1 in.), and 3.8 cm (1.5 in.), with corre-
sponding accelerations of 48, 32, 24, and
16 g's. Thus, they all simulated the 0.6 m
(2 ft) field thickness. All four of these
samples bridged the cavity without failure,
but a small vertical displacement was meas-
ured in each after reaching equilbrium.
The relationship between this position and
the scaling factor appeared to be inversely
proportional to the cube of the scaling
factor. Data from these four tests seem to
indicate that the displacement in the field
prototype would be much greater than the
comparable displacement in the laboratory
models, thereby leading to the conclusion
that the prototype cover would fail sooner
or at lower moisture contents than the
models. Because of the limited number of
tests, this conclusion is tentative, and
conclusive verification through more exten-
sive testing is necessary.
Behavior of Model Clay Layers
Figure 4 illustrates the typical out-
put recorded by the strip chart recorder
during the different phases of a test.
Following startup of the centrifuge, the
sample was accelerated for a time period to
allow pore pressure dissipation within the
sample. Then the cavity was created by
activating the solenoid valve and releasing
the lead shot. Three types of responses
then resulted. Some samples failed imme-
diately (Type 1), others reached an equi-
librium displacement which did not change
over time (Type 3), and some samples had a
rapid initial displacement, which subse-
quently slowed down before the final col-
lapse (Type 2).
The following observations of sample
behavior were noted.
• Small displacements were recorded for
every test as the centrifuge was brought
up to speed. These initial drops were
attributed to the re-seating of the
movable components in the sample con-
tainer with increased acceleration.
RESULTS
Soil Properties
Model tests were carried out using a
silty, inorganic clay (CH). Table 1 lists
some properties of this material, which was
from the stockpile used for the field cover
plot study.
TABLE 1. SOIL PROPERTIES
Optimum moisture
content (standard Proctor) 25 %
Maximum dry
density (standard Proctor) 98 pcf
Percent sand 25 %
Percent silt 37 %
Percent clay 38 %
Specific gravity 2.82
Liquid limit 57 %
Plastic limit 26 %
Plasticity index 31 %
TYPICAL TEST PROCEDURE
1.5
E
u
UJ
UJ
o
1.0
0.
CO
o
0.5
0.0
I. START CENTRIFUGE
2. PORE PRESSURE
DISSIPATION
3. CAVITY CREATION
TYPE I BEHAVIOR
TYPE 2
TYPE 3
40 80 120 160 200 240
TIME IN MINUTES
Figure 4. Output of strip-chart recorder
during centrifuge test run
showing all three types of soil
behavior.
242
-------
• During the pore pressure dissipation
period, rapidly decreasing rates of
displacement were observed during all
tests. This behavior was probably
caused by consolidation of the clay
layer.
• Measurable subsidence of all clay layers
was recorded at the instant of cavity
formation. For dry soils and/or small
cavities, the magnitude of subsidence
was small. No apparent changes in the
clay layers were visible. For wet soil
and/or large cavities, immediate col-
lapse resulted.
The majority of tests were conducted
using one of three soil thickness configur-
ations. For each configuration, tests were
conducted for a number of different soil
moisture contents. Figures 5, 6, and 7
show that increasing moisture content
greatly decreased the ability of the clay
layer to withstand subsidence, and Figures
5 and 6 clearly demonstrate that with a
larger cavity a lower moisture content was
needed to induce failure. The 2.4 m field
cavity size for both the 3.8 cm and 2.5 cm
thick laboratory samples had rapid collapse
occur at slightly above optimum moisture
content. The 3.7 m field cavity size for
the 2.5 cm thick sample failed at 2 % below
optimum moisture content.
Circumferential cracks at the upper
surface and radial cracks at the lower
surface were typical of most tests. Some
of the low displacement (equilibrium) tests
exhibited no cracking, with the tendency
toward cracking stronger for the wider
cavity opening for these samples. Many of
the collapsed samples exhibited spalling of
soil from the lower surface.
Finally, there did not appear to be an
observable scaling factor for time from the
tests. If this is true, then failure of
the prototype cover would occur rapidly
following subsidence, since this was the
behavior exhibited by the laboratory
models.
ACKNOWLEDGMENTS
The authors acknowledge the support of
this study by the Land Pollution Control
Division, Hazardous Waste Engineering Re-
search Laboratory, U.S. Environmental Pro-
tection Agency, Cincinnati, Ohio, through
Cooperative Agreement No. CR-810431 to the
University of Kentucky; Dr. Walter E.
Grube, Jr. is the US EPA project officer.
The authors also gratefully acknowl-
edge D.J. Greer, D.M. Metts, J.E. Wilson,
and R.C. Uarner for their technical assis-
tance throughout this study.
2.4 m CAVITY
24 G'S
27.4
" "
26-5
25.1 X
*
25.4
3.7 m CAVITY
24 G'S
23.2
IO 20 30 4O SO 6O 0 IO 20 3O 40 50 60
TIME IN MINUTES TIME IN MINUTES
*
28.4
X
27.3
x— ^
Y
2.
16
. '
4 m CAVITY
G'S
*
25.7
27.0 _
26,*
O 10 20 30 40 50 60
TIME IN MINUTES
* PERCENT SOIL MOISTURE
Figure 5 (left). Responses of 2.5 cm thick clay layers to 2.4 rn field size cavities
being accelerated at 24 g's.
Figure 6 (center). Responses of 2.5 cm thick clay layers to 3.7 m field size cavities
being accelerated at 24 g's.
Figure 7 (right). Responses of 3.8 cm thick clay layers to 2.4 m field size cavities
being accelerated at 16 g's.
243
-------
REFERENCES
1. AI-Hussaini, M.M., "Centrifuge Model
Testing of Soils: A Literature Review,"
U.S. Army Engineer Waterways Experiment
Station Report, Vicksburg, Mississippi,
1976.
2. Cheney, J.A., "An Introduction to Geo-
technical Centrifuge Modeling," Pro-
ceedings, Recent Advances in Geotech-
nical Centrifuge Modeling, University
of California, Davis, 1982.
3. Lutton, R.J., Regan, G.L., Jones, L.W.,
"Design and Construction of Covers for
Solid Waste Landfills", Army Waterways
Experiment Station, Vicksburg, MS,
prepared for Municipal Environmental
Research Laboratory, Cincinnati, OH,
Report # EPA-6QO/2-79-165. August,
1979.
4. Park, D.W., Summers, D.A., Aughenbaugh,
N.B., "Model Studies of Subsidence and
Ground Movement using Laser Holographic
Interferometry," International Journal
of Rock Mechanics, Mineral Science, and
Geomechanics Abstracts, Vol. 14,
pp. 235-245, Permagon Press, 1977.
5.Skryness, R., "Overview of Design
Considerations for Effective Trench
Covers", Proceedings, Symposium on Low-
Level Waste Disposal, Facility Design,
Construction, and Operating Practices,
Washington, D.C., Sept. 29-30, 1982,
NUREG/CP-0028.
6. Townsend, F.C. and Bloomquist, C.G.,
"Geotechnical Centrifugal Modelling and
University of Florida Experiences,"
Report, University of Florida, Depart-
ment of Civil Engineering, Gainesville,
FL, 1983.
7. Warner, R.C., Wilson, J.E., Peters, N.,
Sterling, H.J. and Grube, W.E., "Multi-
ple Soil Layer Hazardous Waste Landfill
Cover: Design, Construction, Instru-
mentation, and Monitoring," Proceed-
ings, Tenth Annual Research Symposium
on Land Disposal of Hazardous Waste,
EPA-600/9-84-007, Cinci nnati, OH,
April, 1984.
244
-------
LEACHATE COLLECTION SYSTEMS - SUMMARIZING THE STATE-OF-THE-ART
Jeffrey M. Bass
Arthur D. Little, Inc.
Cambridge, MA 02140
ABSTRACT
A manual has been prepared to summarize the state-of-the-art in leachate collection
system design, construction, inspection, maintenance and repair as it pertains to system
failure. Leachate collection system failure mechanisms, as well as design, construction,
inspection, maintenance and repair practices which may reduce the likelihood of failure,
are reviewed.
INTRODUCTION
Leachate collection systems, the
drainage systems which remove excess
liquid accumulation above liners at land
disposal facilities, like all other
liquid collection systems, are subject to
failure. Failure of leachate collection
systems at hazardous waste disposal
facilities are of particular concern
since undetected failures may cause
leachate to build up on top of the liner,
which can lead to failure of the liner
system and subsequent contamination of
groundwater. Furthermore, failures which
are detected may be difficult to repair,
and replacement is not simple since
excavation of hazardous wastes would be
required.
In order to minimize the potential
for system failure, it is first necessary
to understand the mechanisms of failure
and then to understand how these mech-
anisms can be avoided. As a first step
toward this a manual was prepared which
summarizes the state-of-the-art in
leachate collection system design,
construction, inspection, maintenance and
repair as it pertains to the prevention
and correction of failure. A schematic
of a typical leachate collection system
is given in Figure 1.
.APPROACH
The Manual was prepared using
information obtained from interviews with
landfill owners and operators, design
engineers, state and Federal regulatory
officials, and equipment vendors, and
from directly applicable and related
literature. Information from the inter-
views included experience with leachate
collection system performance and
recommendations for system design,
construction, inspection, maintenance or
repair. Directly applicable literature
mostly pertained to leachate collection
system design and construction.
Important related literature included
materials on the inspection and main-
tenance of sewers and agricultural
drainage systems.
Information collected was used in
the Manual in two ways. First, exper-
ience with leachate collection system
failure mechanisms found in the inter-
views and the literature was used to
confirm that suspected failure mechanisms
(1,3) are indeed possible in hazardous
waste systems. Second, experience with
leachate collection system performance,
recommendations of experts, and infor-
mation found in the literature were used
to provide guidance on leachate
245
-------
en
§
rn
o
rn
o
o
o
o
o
—l
O
CO
CO
-------
collection system design, construction,
inspection, maintenance and repair. Cap
drainage systems were also discussed
where appropriate.
RESULTS
Confirmation of Potential Failure
Mechanisms
Table 1 provides a listing of 23
cases of failures which have been
reported for actual leachate collection
systems. Types of facilities included in
the survey were municipal, co-disposal,
industrial and hazardous waste landfills.
When the facility type is not given or
when the information is based on general
experience from facilities of all types,
the facility type in Table 1 is listed as
"not specified". The cause listed in
Table 1 refers to whether the failure was
due to errors in design, construction, or
operation. When a cause could not be
determined based on the information
provided in the interview, the cause is
listed as being "undetermined". The
final column in Table 1 provides comments
on the failure mechanism noted. Three of
the mechanisms are described with the
phrase "not clogged". In these cases,
the biological or chemical mechanism was
active in the facility but a reduction in
flow due to clogging was not noted.
Table 2 summarizes the experience
with failure mechanisms by facility type.
The table indicates that most of the
experience is with municipal and co-
disposal/industrial facilities and that
only two examples were found at hazardous
waste facilities. There are
fewer hazardous waste than municipal and
co-disposal facilities to draw from, and
companies may be less likely to provide
information on problems in hazardous
waste leachate collection systems. This
observation alone does not demonstrate
that failure is less likely to occur in
hazardous waste leachate collection
systems. Experience with failure mech-
anisms in municipal, co-disposal/-
industrial and unspecified facilities may
also be applicable to hazardous waste
landfills if the cause of failure could
occur regardless of facility type.
Construction errors, for example, are
independent of wastes to be handled, and
can occur at any type of facility.
An analysis of the failure mech-
anisms indicates that all of the mech-
anisms are possible at hazardous waste
disposal facilities; there is no con-
ceptual basis for ruling out any of the
mechanisms for hazardous waste leachate
collection systems. Since the least
experience was found for the biochemical
mechanism, it was tested by conducting a
limited laboratory study to determine the
potential for biochemical clogging in a
hazardous waste leachate environment.
Results for this study indicate that a
typical hazardous waste leachate does not
appear to affect the formation of ochre
in the leachate collection system. A
detailed description of the laboratory
program and results obtained are given in
the Manual.
Design considerations to prevent
failure of leachate collection systems
include pipe location, redundancy,
maintenance features, and other features
to address specific failure mechanisms.
Pipe location and placement is
critical to the performance of leachate
collection systems. Crushing or displace-
ment due to equipment loading and differ-
ential settling is a well-recognized
problem. A pipe is best protected when
it is placed in a trench with careful
consideration given to loading conditions
and proper bedding. The trench provides
added protection for the pipe, especially
during placement of the first lift of
waste when the pipe is most susceptible
to crushing.
Redundancy in design is important to
minimize the effect of failure mechanisms
when they occur. The system should be
able to remove liquid from any point in
the facility by more than one pathway.
One of the primary ways to provide
redundancy is to design collection
laterals such that drainage requirements
can be met by the gravel layer alone if
flow through the pipe is restricted. In
addition, laterals should be spaced so
that if one lateral is totally blocked,
liquid can be removed through an adjacent
lateral.
One of the most important consid-
erations to prevent failure is designing
247
-------
TABLE 1
LEACHATE COLLECTION SYSTEM FAILURES
Failure Mechanism
Facility
Type
Cause
Comments
Pipe breakage
Pipe breakage
Pipe separation
Pipe deterioration
Pipe deterioration hazardous
Tank failure
Capacity exceeded
Capacity exceeded hazardous
Outlet inadequate
Sedimentation
Sedimentation
Sedimentation
Sedimentation
Sedimentation
Sedimentation
Biological growth
Biological growth municipal
Biological growth
Biological growth
Chemical
precipitation
Chemical
precipitation
Chemical
precipitation
Biochemical
precipitation
NS 0 by clean-out equipment if
bends greater than 22°,
general experience
municipal D differential settling,
improper bedding
municipal C joints not glued
NS D problems with ABS pipe,
general experience
0 from acid or solvent
disposed of in wrong cell
co-disposal D leachate holding tank
co-disposal D under-design, other problems
noted
0 periodic rather than
automatic pumping of sump
co-disposal D caused leachate buildup
NS C no filter installed
NS U general experience
co-disposal U in 1 year old system
co-disposal U of gravel layer and pipe
municipal U general experience
NS C general experience
industrial D 100 ft. long biological growth
flushed out under high pressure
U reduction in flow every 2 years;
flushed out
municipal U of filter fabric
co-disposal U on 3/4 inch stone, not clogged
municipal 0 EPA test cell, not clogged
co-disposal U iron oxide, not clogged
co-disposal 0 attributed to waste
characteristics
co-disposal U in leachate collection wells
NS = not specified; C = construction related; U = undetermined;
0 = operation related; D = design related.
248
-------
TABLE 2
SUMMARY OF FAILURES BY FACILITY TYPE
Facility Type
Co-disposal/ Not
Mechanism Municipal industrial Hazardous Specified Total
Pipe breakage
Pipe Separation
Pipe deterioration
Other non-clogging
1
1
-
3
-
-
1
1
1
-
1
_
2
1
2
4
problems
Sedimentation 1
Biological growth 2
Chemical precipitation 1
Biochemical precipitation -
6
4
3
1
TOTAL
10
23
the system to facilitate inspection and
maintenance. Access should be provided
to all parts of the system. This
includes the placement of manholes and
cleanouts so that maintenance equipment
can reach any section of pipe. The
design should consider minimum pipe size,
distance between access points, and
maximum angles negotiable by such equip-
ment. For example, some jet cleaning
equipment requires at least a four inch
pipe and can operate in straight runs up
to 1000 feet in length. In one case,
cleanout equipment is reported to break
pipes at bends greater than 22 degrees.
Finally, leachate collection systems
can be designed to avoid specific
clogging mechanisms. For example,
sedimentation can be avoided by selecting
the proper grain size distribution for
the filter material, and by providing
minimum slope and avoiding hydraulic
perturbations to maintain flow velocity
during leachate collection so that solids
can not settle out. Maintaining flow
velocity will also help prevent buildup
of biological and chemical materials.
Selection of construction materials based
on the wastes to be handled by each cell
of a hazardous waste disposal facility
can help avoid deterioration caused by
reactions with waste leachates.
Construction
Leachate collection system con-
struction according to design specifi-
cations is critical. Specific con-
struction problems which can cause
leachate collection system failure
include:
• the use of materials with
specifications other than the
ones in the approved design;
• foreign objects (e.g., rocks,
clumps of soil, construction
tools) left in drainage system
piping which plug or restrict
leachate flow and may not be
249
-------
removable using designed
maintenance procedures;
• neglecting to install materials
in the location specified in
the design;
• siltation of filter or drainage
layers resulting from improper
upgradient drainage during
construction and/or careless
construction techniques; and
• the use of improper con-
struction equipment or improper
placement procedures which
cause damage to system com-
ponents, such as crushing or
misalignment of pipes due to
loading of heavy equipment.
Construction quality assurance (CQA)
is needed to verify that the completed
leachate collection system meets or
exceeds the design requirements. This
involves monitoring and documenting the
quality of materials used and the con-
ditions and manner of their placement.
CQA serves to detect variations from
design, whether as a result of error or
negligence of the contractor, and to
provide for suitable corrective actions
before wastes are accepted at the
facility. Without proper CQA, problems
with leachate collection systems due to
construction, like the two specific
examples noted in Table 2, would not be
discovered until the system failed during
operation.
In addition to CQA, specific con-
struction methods may be important in
avoiding failure mechanisms. For
example, John Pacy of Emcon Associates
(2) recommends placing pipe in trenches
in small sections and completing each
section before starting on the next in
order to minimize the amount of surface
sediments which is washed into the drain
during construction. Construction
techniques which are important to system
performance should be specified in the
design and monitored by the CQA program.
Inspection
Inspection methods are intended to
determine whether failure mechanisms are
active in the leachate collection system.
Regular inspections include monitoring
flow at outlets or access points,
monitoring leachate level within the
facility, correlating leachate quantity
with rainfall data, and correlating
leachate quality with clogging
indicators, such as the presence of iron
reducing bacteria. Direct methods
include TV or photographic inspection of
drainage pipes and checking the physical
continuity of the pipes with sewer
cleaning equipment. Direct methods are
limited to use in drain pipes, since the
drainage and filter layers are not
accessible to physical inspection.
In addition to regular or periodic
inspections, the following special
inspections should be performed as
appropriate:
• as-built inspection, as part of
the CQA plan, to verify con-
tinuity of the pipe network
after construction.
• re-verification of continuity
after the first lift of waste
is placed.
• testing to locate and diagnose
suspected problems.
Maintenance
Experience with maintenance tech-
niques at leachate collection systems is
lacking. It seems that system repair is
generally used as a substitute for
maintenance. Although maintenance and
repair often involve the same methods,
regular maintenance may be the more
cost-effective option, since a less
intensive effort may be required and
since some problems which can be avoided
by maintenance cannot be easily corrected
once they begin.
The simplest maintenance technique
is flushing of the pipe network. This
can be done by gravity flow using water
or leachate, or by low pressure jet
cleaning equipment. Flushing and jetting
remove sediments and immature precipitate
before substantial buildup can occur.
Standard sewer cleaning devices such as
roto-rooters, snakes, pigs, buckets,
sewer balls may also be used to clean the
collection system. These perform
basically the same function as flushing
and are also a good check on the con-
250
-------
tinuity of the drains since they can
discover crushed or displaced pipes, as
well as areas of more serious clogging.
Control of some failure mechanisms
can also be accomplished by controlling
waste inputs to the extent possible.
Wastes placed in the first lift, for
example, should have a high permeability
and may have special chemical character-
istics. High permeability wastes are
placed in the first lift to facilitate
the flow of leachate to the collection
system. An example would be a carefully
placed layer of upright drums with sand
or soil placed in between the drums.
Special waste characteristics in the
first lift may include wastes of
relatively lower pH, wastes which inhibit
biological activity, and wastes which do
not contain high iron, calcium, mag-
nesium, nutrient, or sediment content.
The chemical characteristics desirable in
the first lift of waste are dependent on
the leachate characteristics expected in
general at the facility, and the failure
mechanisms which are expected to be
active.
Currently, no techniques are avail-
able for the maintenance of drainage or
filter layers. Certain chemical tech-
niques developed for other drainage
systems (SCL gas, hydrogen perioxide,
acid treatment) may be applicable to
drainage layers in leachate collection
systems, although considerable testing
would be required. Compatibility of the
chemicals with the waste disposed of at
the facility must also be considered.
Repair of a leachate collection
system is necessary when the system is
unable to remove leachate from the
facility, causing leachate to accumulate
over the liner. The need for repair can
be determined using the inspection
techniques discussed above. If the
failure of the leachates system is due to
clogging of the pipe, standard sewer
cleaning techniques may be used to
correct the problem, provided the pipe is
accessible via manholes or cleanouts.
When the pipe is not accessible or
when the problem is with the drain or
filter layer, or when maintenance tech-
niques are not effective, replacement of
the failed portion of the leachate
collection system may be required.
Replacement techniques include:
• excavating the failed portion
with repair or replacement
using new materials. The
acceptability of excavation
will depend on the type of
waste in the facility (to
evaluate safety hazards posed
by excavation) and the number
of lifts of waste in place.
• retrofitting a new collection
line adjacent to the failed
portion to remove accumulated
leachate.
0 installing well points through
the waste to remove accumulated
leachate.
ACKNOWLEDGEMENTS
This project has been funded wholly
by the United States Environmental
Protection Agency under contract
68-03-1822 to Arthur D. Little, Inc. The
E.C. Jordan Company of Portland, Maine
provided input to the Manual under
subcontract to Arthur D. Little, Inc.
The Manual is in the EPA review process
at this time (2/85). Mention of trade
names or commercial products does not
constitute endorsement or recommendation
for use. The Project Officer is Mr.
Jonathan Herrmann, Municipal
Environmental Research Laboratory,
Cincinnati, Ohio.
REFERENCES
1. Bass, Jeffrey M., J.R. Ehrenfeld and
J.R. Valentine. "Potential Clogging
of Landfill Drainage Systems." U.S.
EPA, Municipal Environmental Reseach
Laboratory, Cincinnati, OH.
EPA-600/S2-83-1090, February, 1984.
2. Pacey, John, Emcon Associates.
Personal interview. November, 1983,
3. Young, Charles W., T.J. Nunno, M.P.
Jasinski, D.R. Cogley and S.V.
Capone. "Clogging of Leachate
Collection Systems Used in Hazardous
Waste Disposal Facilities," Draft
White Paper, prepared for U.S. EPA,
Research Triangle Park, North
Carolina. June, 1982.
251
-------
VERIFICATION OF LEACHATE COLLECTION EFFICIENCY
WITH PHYSICAL MODELS
Paul R. Schroeder
Environmental Laboratory
USAE Waterways Experiment Station
Vicksburg, Mississippi 39180
ABSTRACT
Two large-scale physical models of landfill liner/drain systems were constructed to
examine the effects that the length, slope and hydraulic conductivity of the drain layer
and the depth of saturation above the liner have on the subsurface lateral drainage rate.
The models have two different lengths and adjustable slope. The models were filled with
a 3-foot sand drain layer overlying a 1-foot clay liner. A 2-inch layer of gravel was
placed under the liner to collect seepage from the clay liner. The design, construction,
instrumentation and preparation are described in detail.
Several drainage tests were run on each configuration of the models by applying
water as rainfall to the surface of the sand layer and then measuring the water table
along the length of the models and the drainage rate as a function of time. The experi-
mental design and test procedures are presented along with some descriptive results.
INTRODUCTION
Background
Landfills have been widely used for
disposal of municipal, industrial and
hazardous solid wastes. Storage of any
waste material in a landfill poses several
potential problems. Among these is the
possible contamination of ground and sur-
face waters by the migration of water or
leachate from the landfill to adjacent
areas. Given this potential problem, it
is essential that the liquids management
technology perform as expected during its
development. It is also essential that
the performance of the technology can be
simulated or modeled with sufficient accu-
racy to design landfills to prevent migra-
tion of water from the facility. The
modeling of the moisture movement through
landfills is also important to review
landfill designs and to determine the ade-
quacy of the design and the limitations of
the liquids management technology.
This study was conducted to test and
verify the liquid management technology
of lateral sub-surface drainage in covers
and leachate collection systems. Specif-
ically, this involves verification of the
Hydrologic Evaluation of Landfill Perfor-
mance (HELP) Model and other regulatory
and technical guidance provisions and pro-
cedures developed by the U.S. Environmen-
tal Protection Agency (USEPA).
The HELP model is a computer model
that generates water budgets for landfills
by performing a daily sequential simula-
tion of water movement into, through and
out of a landfill. The model provides
estimates of runoff, evapotranspiration,
lateral drainage, percolation and depths
of saturation. Lateral drainage is com-
puted in the model as a function of the
average depth of saturation above the
liner, the slope of the surface of the
liner, the length to the drainage collec-
tor, and the hydraulic conductivity of the
lateral drainage layer (1, 2):
0.16
-UU
2(0.51 + 0.00205aL)KyV\aL
+ all
252
-------
where:
Q = lateral drainage rate,
inches/day (in/day)
K = hydraulic conductivity, in/day
y = average depth of saturation,
inches
a = slope, dimensionless
L = drainage length, inches
The USEPA regulatory provisions for
lateral drainage layers require only that
the depth of leachate buildup at the bot-
tom of the landfill should not exceed
I foot (ft) and the construction materials
for leachate collection should be resis-
tant to chemical attack and the physical
forces exerted on them (3). USEPA techni-
cal guidance states that the drainage layer
be constructed to be at least 12 inches
thick at a minimum slope of 2 percent and
to have a hydraulic conductivity of not
less than 1 x 10 centimeters/second
(cm/sec). Also, the drainage tile system
should be of appropriate size and spacing
to efficiently remove the leachate. It is
believed that 4-in diameter (dia) tiles
spaced 50 to 200 ft apart would be ade-
quate for removing leachate (3). The
drainage layer in the cover should have
the same specifications as above except
that tile drainage systems are not neces-
sary although free drainage must be pro-
vided at the perimeter of the cover (3).
Purpose
The objective of this study was to
verify the USEPA technical guidance by
determining the lateral drainage rate as a
function of the hydraulic conductivity,
slope, length and depth of saturation of
the lateral drainage layer in large-scale
physical models. The measured average
depths of saturation, drainage rates and
drainage times in the physical models will
be compared with the results predicted by
the HELP model and its equations for ver-
tical and lateral drainage. For the re-
mainder of this paper the term, "model"
refers to the physical models.
EXPERIMENTAL APPROACH
Model Design and Construction
Two physical models were designed and
constructed to perform the laboratory ver-
ification tests. The models were con-
structed at two different lengths to
permit the examination of the effects of
length on drainage. One model was built
to have a usable depth of 5.0 ft, an
inside width of 5.3 ft, and a drainage
length of 25.4 ft. The other model had
the same depth and width but had a drain-
age length of 52.4 ft.
Both models were constructed of iden-
tical materials. The base of the models
were constructed of steel soil test cars
for mobility studies. These cars are re-
inforced 1/4-in steel tanks that are 27 ft
long, 5.3 ft wide and 2.5 ft deep. The
sides of the cars were extended upward
4 ft with marine grade, 1/2-in plywood
supported by 2-in angle iron. Silicone
sealant was used to fill the joints and
cracks and to seal and prevent leaks.
The models were placed at a two per-
cent slope and fitted with supports for
attaching 5-ft screw jacks at the upper
end. A jacking structure was built to
attach to either model which could raise
the end of a fully loaded model. The jack
could increase the slope of the long model
to eleven percent and the short model to
twenty percent. The slopes of the models
during testing were determined by
surveying.
Three troughs, approximately
12 inches deep and extending across the
width of the models, were placed inside
the lower end of each model. The bottom
trough was used to collect seepage through
the clay liner. The middle trough col-
lected subsurface lateral drainage at the
bottom of the sand drainage layer. The
top trough collected runoff from the sur-
face of the sand. The bottom two troughs
were constructed of 1/4-in steel plates
welded to the steel soil test car. The
top trough was fabricated of galvanized
sheet metal and bolted in place to the
plywood walls. The sheet metal was sealed
to the walls with silicone sealant.
The runoff trough was drained to a
5-ft deep, 32-in dia sump tank via 1 1/2-
in polyvinyl chloride (PVC) pipe. The
water level in the tank was measured with
a steel measuring tape before and after a
rain event to determine the runoff volume.
The drainage trough was also drained to a
5-ft deep, 32-in dia sump tank via 1 1/2-
in PVC pipe. The drainage sump tanks were
automatically pumped out when filled to a
height of about 44 inches, requiring only
253
-------
about 5 minutes to lower the water depth
to about 10 inches. The minimum time re-
quired to fill the tank was about 4 hours.
The seepage trough was periodically
drained into a 4-litre graduated cylinder
to measure the seepage volume. The seep-
age was drained from the trough via an
attached 2-in rubber hose which was sealed
between measurements.
The lateral drainage rate from the
sand layer was determined by two devices,
one for low flow rates and the other for
high flow rates. Low flow rates (below
0.08 gallons/minute (gpm) or below
0.05 inches/hour (in/hr) for the short
model and 0.02 in/hr for the long model)
were determined by measuring the drainage
volume as a function time using a Weather-
tronics Model 6010 Tipping Bucket Rain
Gage. The tipping bucket collected the
water from the end of the PVC pipe drain-
ing the drainage trough and discharged the
water into the drainage sump tank. A by-
pass was placed in the PVC pipe to divert
the drainage directly into the sump tank
during periods of very high flow rates.
High flow rates (greater than 0.08 gpm)
were determined by measuring the water
level in the drainage sump tank as a func-
tion of time using a Weather Measure Cor-
poration Model F553-A Water Level Recorder
powered at 12 volts by a ELCO Model 1060S
Low Ripple Battery Eliminator and Charger.
Both models were fitted with a device
to spray water uniformly across the length
of the surface of the sand drain layer.
This device consisted of a 1 1/2-in PVC
pipe mounted on a 12-ft long aluminum cart
that tracked back and forth on rails that
ran the length of the model. A nozzle
that sprayed a uniform line of water
across the width of the model was attached
on each end of the pipe. Each nozzle
sprayed on one half of the model's sand
surface as the cart moved its maximum dis-
tance in one direction. The height of the
nozzles and the travel limits of the cart
were adjustable to insure that the entire
surface of the sand was sprayed. The
water application rate or rainfall inten-
sity was adjusted by changing the size of
the nozzles or changing the water pressure
applied on the nozzles. The intensities
of flow rates could be varied between
0.02 in/hr to 6 in/hr or 0.06 gpm to
10 gpm. The intensity during a test was
determined by measuring the volume of
water applied as a function of time. A
stopwatch was used to measure the time and
a Badger Meter, Inc. Model Recordall 12
water meter measured the water volume.
The nozzles used in the tests were Flood-
jet nozzles 1/8 K.25 316SS through 1/4
K18 316SS that were manufactured by
Spraying Systems Co.
The major features of the models are
shown in the schematic in Figure 1. The
schematic is a side view of the short
model showing the loaded test cell,
troughs, drain lines, sumps, jacking
structure, rain cart, and instrumentation.
Test Preparation
After construction was completed, the
models were filled with three functional
layers for testing and a layer of fill
dirt as shown in Figure 1. The top layer
was a 3-ft sand layer for lateral drain-
age. The second layer was a 1-ft com-
pacted clay liner designed to minimize
percolation. The third layer was a 2-in
layer of pea gravel to transmit the seep-
age from the clay liner to a drain. The
bottom layer was a 10-in layer of fill
dirt to build up the pea gravel layer to
promote drainage.
Ten inches of fill dirt was compacted
in the bottom of the steel test car to
fill the trapezoidal section of the model
and to reduce the volume of pea gravel
required. The fill dirt was covered by a
30-mil, butyl rubber impermeable membrane,
T-16 manufactured by Firestone. The mem-
brane was glued to the walls of the steel
car to prevent percolation into the fill
dirt. G-580, a synthetic rubber resin
dispersed in solvent and manufactured by
Pittsburgh Paint & Glass, was used as the
glue.
Two inches of washed creek pea gravel
was placed on the impermeable membrane and
in the seepage collection trough. This
layer was designed to have a high permea-
bility and low storage potential in order
to transmit the small volume of seepage
rapidly to the collection trough. The
layer was covered with Bidim Type C-34
filter fabric manufactured by Monsanto to
prevent migration and penetration of the
clay liner into the pea gravel and to pro-
tect the ability of the pea gravel to
transmit seepage. The filter fabric was
glued to the walls of the steel car to
insure that the fabric and pea gravel did
254
-------
JACKING
STRUCTURE
RUNOFF SUMP
TIPPING BUCKET
AND DRAIN SUMP
SEEPAGE
COLLECTION
Figure 1. Schematic of the Short Model.
not move during placement of the clay
liner. Both the gravel and fabric were
wetted during placement of the liner.
Wetting reduced the seepage required to
sufficiently deplete the storage volume of
the pea gravel and rubber membrane for
transmission of the seepage to begin.
A 12-in clay liner was placed above
the seepage transmission layer. Buckshot
clay, a locally found, well-defined clay,
was used. The clay was spread on a con-
crete strip and cut up into small clods
using a scarifier. The clay was then
wetted to a moisture content of about 27%
and placed in the model in two-inch lifts.
Each lift was compacted with a gasoline-
powered, hand-operated compacter manufac-
tured by Wacker. The compacter had a com-
pacting area of about 0.5 square feet (sq.
ft). Additional water was added during
compaction to insure good blending and to
prevent drying. At least four passes were
made on each lift with the compacter at
which time no additional compaction could
be discerned by additional passes.
After placement of the clay liner was
completed, a sample of the top 3 inches
was taken in the short model. Its water
content was 27.1%, its dry density was
94.9 pounds per cubic foot (pcf) and its
coefficient of permeability was 1.67 x
10 cm/sec. A second sample, taken at a
depth of 6 inches while filling, had a
moisture content of 28.3% and a dry den-
sity of 91.0 pcf. Nuclear density probe
measurements indicated a moisture content
of 34.5% and a dry density of 89 pcf. The
moisture content read by the nuclear den-
sity meter is much higher because it in-
cludes water of crystallization and hy-
dration which would not be and was not
measured by the standard moisture content
test. Buckshot clay is known to have a
high content of bound water. In all of
the samples the dry densities were very
high for the moisture content which indi-
cates very good compaction. The soil
testing were performed in accordance with
ASTM accepted procedures as outlined in
Laboratory Soil Testing, EM 1110-2-1906
(4). The nuclear density measurements
were made with a Troxler Electronic Labo-
ratories, Inc. Model 341IB nuclear density
gage using the procedures outlined in its
instruction manual. The same sampling
procedures were used for the long model
with similar results.
While placing the clay liner, a flap
of the T-16 rubber membrane was glued to
255
-------
the walls of the steel car at eight inches
above the base of the clay and four inches
below the eventual top of the clay liner.
This flap extended four inches into the
clay. The purpose of the flap was to pre-
vent seepage from occurring between the
car and the clay by diverting the water
into the clay in the event that the clay
shrank or cracked at the walls.
Six transducers were placed in the
long model and five in the short model.
Consolidation Electrodynamics Corporation
Type 4-312 pressure pickups were fitted in
a brass body with a porous stone such that
the pressure transducer could only measure
pore pressure which in this case was the
head of water on the transducer. The
transducers were placed at the top of the
clay liner and were exposed to the sand
layer. The locations of the transducers
were approximately 3, 6, 11, 16 and 23 ft
from the lip of the drainage trough in the
short model and 3, 6, 11, 23, 35 and 50 ft
from the trough in the long model. The
transducers were placed on the centerline
dividing the width of the models. The
elevations and locations of the trans-
ducers were determined by surveying with
an optical transit.
A 3-ft layer of sand was placed above
the clay liner. The sand was placed in
4-in lifts and each lift was vibrated
until no additional compaction could be
discerned. In general, four passes were
made with a hand-operated, gasoline-
powered sand compacter manufactured by
Wacker. The size of the vibrating plate
on the compacter was about 2 sq ft.
Two different sands were used in sep-
arate tests. One sand was Reid-Bedford
sand with about ten percent Windham clay
coarsely dispersed throughout the sand in
the short model and with about thirteen
percent clay in the long model. The dry
density, as placed, was 106 pcf, the spe-
cific gravity was 2.7, and the laboratory
coefficient of permeability was about 4 x
10~ cm/sec. The porosity was 0.37, the
drainable porosity was about 0.21 to 0.25,
and the wilting point (i.e. the moisture
content at 15 bar tension) was about 0.07.
The effective drainable porosity was much
smaller due to air being trapped in the
sand. The volume of air varied during
testing in the physical model.
The other sand was a locally found,
ungraded washed creek sand. The dry den-
sity, as placed, was 107 pcf in the long
model and 110 pcf in the short model. The
specific gravity of the sand was 2.65, and
the coefficient of permeability was
2.2 x 10~ cm/sec. The porosity was 0.35
in the short model and 0.36 in the long
model. The drainable porosity was about
0.24 to 0.28 and the wilting point was
about 0.04 to 0.05.
Piezometers were installed in the
sand layer to provide a backup method to
measure the depth of saturation above the
clay liner. These measurements were also
used to calibrate the pressure transducers
and to check the accuracy of the transduc-
ers throughout the testing. Piezometers
were constructed by attaching 1 1/2-in
diameter, 4-in long cylindrical porous
stones to 1/2-in PVC pipes of approxi-
mately 4-ft lengths. The piezometers were
installed with the stone end placed flush
with the top of the clay liner and with
the pipe extending straight up through the
sand layer. The depth of saturation was
determined by measuring the depth to the
water surface from the end of the pipe and
then subtracting that value from the total
length of the piezometer. Piezometers
were placed at the locations of the second
and last transducers from the lip of the
drainage trough in the short model and at
the second, fourth and last transducers
from the lip of the drainage trough in the
long model.
Experimental Design
A complete block experimental design
was used to examine the effects of drain-
age length, slope, hydraulic conductivity,
depth of saturation, rainfall intensity
and rainfall duration on the lateral sub-
surface drainage rates. The block design
was selected because it provided the most
data with the least time and expense for
construction and model preparation. Only
two models could be built due to their
costs and only two sands could be examined
due to the unavailability of another of
significantly different permeability.
Several slopes and rainfall events could
be examined quickly since very little time
was required for changing these test con-
ditions. Also, the time requirements and
costs for running an additional test with
256
-------
a different slope or rainfall were less
than ten percent of the requirements for
preparing the model for a different sand.
Additional rainfall events were examined
in lieu of replicates since the lateral
drainage rate as computed by the HELP model
does not directly consider the effects of
rainfall intensity or duration. Also,
since a complete block design was used,
the effect of a change in a variable is
directly examined under multiple test con-
ditions, reducing the need for replicates.
In the model tests two drainage
lengths were compared — 25.4 ft and
52.4 ft. Three slopes were examined —
approximately 2, 5 and 10 percent. Sands
of two hydraulic conductivities were used
— 4 x 10 cm/sec and 2.2 x 10 cm/sec.
Four rainfall events were examined — a
1-hr rainfall at 0.50 in/hr, a 2-hr rain-
fall at 1.50 in/hr, a 6-hr rainfall at
0.50 in/hr and a 24-hr rainfall at
0.125 in/hr. In addition to these drain-
age tests, the sand was saturated, pre-
dominantly from the bottom up for several
test conditions, and then allowed to
drain. Also, water was applied to the
sand for a long period of time (generally
more than 36 hours) at a rainfall inten-
sity which would maintain the average
depth of saturation in the sand at
12 inches.
Test Procedure and Data Collection
Each test was run in one of three
manners. Lateral drainage was measured
either from a pre-saturated sand, from the
sand before, during and after a rainfall
event, or from the sand while attempting
to maintain a constant average depth of
saturation or head at twelve inches. The
test procedure for the three manners were
very similar; only the procedure for
applying the water to the model varied.
The procedures for running a test
were as follows:
1. Prepare model for testing. This
included any preparation of the sand such
as placing or leveling, preparation of
rainfall cart for desired rainfall inten-
sity by changing nozzles, nozzle heights
and end stops, changing the slope of the
model, and draining sump tanks.
2. Initialize test which included
starting the recorders for drainage and
heads, setting the time, draining the
lateral drainage trough, recording test
conditions, measuring initial depth of
water in runoff sump tank and recording
the initial water meter reading.
3. Apply water to the model as
prescribed for the test. This included
adjusting the water application rate and
the spray pattern.
4. Note peculiarities of the test
including unusual environmental
conditions, problems, and variations from
the test procedure.
5. After the rainfall was completed
and runoff ceased, measure the final depth
of water in runoff sump tank.
6. Measure water levels in piezom-
eters periodically.
7. Drain tanks as required.
8. Continue testing until the heads
become so low as to nearly expose a trans-
ducer to air and then stop drainage or
start a new test.
9. Collect data from recorders and
reduce the data.
Prior to calibration of the pore
pressure transducers, water was added
slowly to both ends of the model. In this
manner the sand was saturated primarily
from the bottom up which forced the air
out of the sand. Following the cali-
bration, the sand was drained and the
drainage rate and heads on the liner were
measured continuously while draining.
Drainage from pre-saturated sand was
measured for only four test conditions
since saturation was required only for
calibration.
For the tests in which a constant
average head of twelve inches was main-
tained, water was applied to the model to
increase the head on the liner rapidly to
twelve inches. This water was applied by
raining at an intensity of about 0.5 in/hr.
After reaching a head of twelve inches,
the rainfall intensity was decreased to
the intensity estimated to maintain the
head. The desired intensity was deter-
mined by adding together the evaporation
rate, the rate of leakage and seepage, and
the drainage rate which corresponded to an
257
-------
average head of the twelve inches during a
previous rainfall for the same physical
test conditions. The intensity was ad-
justed during the test if the head did not
remain constant near twelve inches. The
rain was continued for at least 12 hours
after the head remained nearly constant.
The evaporation rate ranged from
about 0.2 to 0.3 in/day while raining.
The rate was dependent on building temper-
ature and on the spray nozzle. The build-
ing temperature ranged from about 58°F to
90°F but for most tests the temperature
ranged between 70°F and 82°F.
Leakage varied directly as a function
of head on the liner. Both cars leaked
but the short model was more watertight
than the long model. The leakage rates
for the short model were about 0.02 in/day
at a head of 20 inches and about 0.01 in/
day at a head of 12 inches while the leak-
age rates for the long model were about
0.12 in/day at a head of 20 inches and
about 0.07 in/day at a head of 12 inches.
Neither model leaked at heads below
6 inches since all of the seams in the
model were above this depth. The long
model leaked more since it suffered
greater deflections upon loading, causing
more seams to rupture and leak. The leak-
age rates, although important from an
overall water balance perspective, have a
negligible influence on the drainage
rates.
The majority of the drainage tests
encompassed measurements before, during
and after a specified rainfall event. In
these tests, a specified quantity of
water, either 48.5 or 291 gallons for the
short model and either 98.5 or 591 gallons
for the long model, was sprayed across the
entire surface of the sand in a specified
period of time, either 1, 2, 6 or 24 hours.
These quantities of water correspond to
about 0.57 inches and 3.4 inches of rain,
respectively for both models. The tests
measured the heads and drainage rates as
the head increased from a starting point
of about 4 inches to a peak value occur-
ring about 30 minutes after the rainfall
ceased and then returned to about 4 inches
when the test was stopped. Several tests
were allowed to continue until the head
was nearly zero inches.
RESULTS
Twelve model configurations were
examined in this study — two different
sands placed in two models of different
length (25.4 ft and 52.4 ft) and tested at
three slopes. Typical results of the
drainage tests for these configurations
are presented in Figure 2. Each curve is
a plot of the drainage rate per unit area
as a function of the average depth of sat-
uration above the liner and shows results
for drainage from a model with a given
drainage length and hydraulic conductivity
at three slopes — 2, 5 and 10 percent.
The results shown here are from tests
using a rainfall duration of 6 hours and a
rainfall intensity of 0.5 in/hr except for
the curve for 2 percent slope in Figure 2a.
This curve shows the results for a rain-
fall duration of 24 hours and a rainfall
intensity of 0.125 in/hr. The results for
other durations and intensities are simi-
lar to those shown here.
Figure 2 shows the effects that inde-
pendently changing each of the four major
parameters (head on the liner or depth of
saturation, slope, hydraulic conductivity
and drainage length) has on the drainage
rate. Each graph shows the effects of
changes in slope and head on the liner
while holding the other parameters con-
stant. Comparisons between Figures 2a
and 2b and between Figures 2c and 2d show
the effects of changes in drainage length.
Comparisons between Figures 2a and 2c and
between Figures 2b and 2d show the effects
of changes in hydraulic conductivity.
Figures 2a and 2b show the results
for fine sand in the short model and in
the long model, respectively. The perme-
ability was measured in the laboratory to
be about 0.005 cm/sec in the short model
and 0.004 cm/sec in the long model but the
results indicate that the permeability of
the sand as placed in the models was
higher. The permeability of the sand was
about 0.006 to 0.007 cm/sec in the short
model and about 0.007 to 0.009 cm/sec in
the long model.
Figures 2c and 2d show the results
for coarse sand in the short model and in
the long model, respectively. The permea-
bility of the coarse sand was measured to
258
-------
e.176-
B.15B-
R 8.125-
A
1
N
e. i ee-
R
A
7
E B. 876-
I
N
/ 8.868-
H
R
a.925-
8 888-
2%
te 15 2e
AVG HEAD CIN>
38
e ee-
B.aa-
e 07-
e.eo-
e.es-
8.B4-
8 83-
B 82-
e.ai-
8.88-
IS 20 26
AVO HEAD
-------
be 0.22 cm/sec in the laboratory but the
results in the model indicate that permea-
bility was much lower. The results for
the short model indicate that the permea-
bility was about 0.017 to 0.025 cm/sec
while the results for the long model indi-
cate about 0.023 to 0.030 cm/sec. The
results show an hysteresis indicating that
the drainage rate was greater while the
head was building than while falling.
Detailed examination of the results
has not been completed at this time; how-
ever, preliminary results indicate that,
as expected, an increase in the head on
the liner, slope or hydraulic conductivity
yielded an increase in the drainage rate
per unit area of surface as did a decrease
in the drainage length. The changes were
similar in magnitude to those predicted by
the equation used in the HELP model when
using the apparent hydraulic conductivity
of the sand as placed instead of the value
measured in the laboratory. The agreement
between the hydraulic conductivity of the
sand measured in the laboratory and the
apparent value determined from the drain-
age results was poor.
SUMMARY
Two large-scale physical models of
landfill subsurface drainage systems were
designed and constructed to have drainage
lengths of 25.45 ft and 52.45 ft. The
models were filled with a 1-ft clay liner
and a 3-ft layer of sand. Drainage tests
were conducted on two sands having permea-
bilities of 0.004 cm/sec and 0.22 c-n/sec.
The models wera tested at slopes of 2, 5,
and 10 percent. The analysis of the
results was not completed at che submittal
time of this paper but the initial exami-
nation of the results indicated that the
permeability of the sand as measured in
the laboratory did not agree well with the
apparent hydraulic conductivity of the
sand as placed in the model. The results
showed the expected trends where an in-
crease in the head on the liner, slope or
hydraulic conductivity yielded an increase
in the drainage rate per unit area of sur-
face as did a decrease in the drainage
length.
DISCLAIMER
Mention of trade names or commercial
products does not constitute endorsement
or recommendation for use by the U.S. En-
vironmental Protection Agency or the
U.S. Army Engineer Waterways Experiment
Station.
ACKNOWLEDGEMENTS
The author appreciates the support
and guidance of Mr. Roy Leach of the USAE
Waterways Experiment Station for his con-
tributions to the design, construction and
preparation of the models. The author
also appreciates the support and guidance
of Mr. Douglas Ammon, Project Officer,
USEPA Hazardous Waste Engineering Research
Laboratory.
REFERENCES
1. Schroeder, P. R., J. M. Morgan, T. M.
Walski, and A. C. Gibson. 1984. The
Hydrologic Evaluation of Landfill Per-
formance (HELP) Model: Volume I.
User's Guide for Version 1, EPA/530-
SW-84-009, U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio.
2. Schroeder, P. R., A. C. Gibson, and
M. D. Smolen. 1984. The Hydrologic
Evaluation of Landfill Perfor-
mance (HELP) Model: Volume II.
Documentation for Version 1, EPA/530-
SW-84-010, U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio.
3. U.S. Environmental Protection Agency.
1982. RCRA Technical Guidance Man-
uals, 40 CFR 264, Washington, DC.
4. Headquarters, Department of the Army.
1970. "Laboratory Soils Testing,"
Engineer Manual EM 1110-2-1906,
Washington, DC.
260
-------
CHEMICAL RESISTANCE OF FLEXIBLE MEMBRANE LINERS
IMMERSED IN SELECTED AQUEOUS SOLUTIONS
Gordon E. Bellen, Principal Investigator
Rebecca Corry, Project Assistant
National Sanitation Foundation
Ann Arbor, Michigan 48106
ABSTRACT
The goal of this project is to develop
screening criteria for evaluating the
chemical compatibility of flexible
membrane liner (FML) materials. These
criteria will be used to make a material/
chemical waste compatiability guide.
Data on the resistance of five common FML's
to nine different chemicals plus water are
being generated during the project. The
data will be compiled and compared with
existing chemical resistance tables (from
trade groups and manufacturers) to form a
general matrix for screening FML chemical
resistance. The use of accelerated test
procedures (high temperature immersions)
will be evaluated by testing at 23°C and
50 °C.
The FML materials represent the three major
types of liner materials: elastomers
(ethylene propylene rubber and
epichlorohydrin); thermoplastics (polyvinyl
chloride and chlorinated polyethylene).
Unsupported materials only are being are
used to focus the study on polymer/chemical
interaction.
The nine chemicals are sodium chloride,
hydrochloric acid, sodium hydroxide,
potassium dichrcmate, methyl ethyl ketone,
phenol, furfural, ASTM #2 oil, and
1,2-dichloroethane. Each represents a
chemical type commonly found in waste
streams. Although an actual waste stream
will usually be a mixture of several
components, this study uses only one
chemical per exposure in water solution.
Inorganic chemicals represented are
solutions of acid, base, salt, and
oxidizer. Organic groups represented are
ketone, aldehyde, organic acid, oil, and
chlorinated hydrocarbon. Some of the
specific chemicals were chosen to
correspond with those being used in
related FML chemical resistance studies.
Initial screening showed that most
materials increased in weight and volume
during a seven day exposure period, but in
many cases returned to very near the
original size and weight after the air-
drying period. It was decided to test
materials as soon as possible after immer-
sion and to keep them from drying out
before being tested. Some of the materials
were destroyed by pure solvents, as
expected. As a result, the majority of
concentrations chosen for further testing
were at or below the water solubility
limit. Three concentrations of each
organic chemical were selected to show the
effect of concentration on material change.
Samples are tested for change in weight,
volume, tensile properties, and tear
resistance after exposure times of 1, 7,
14, 28, and 56 days at each temperature.
Also, weight and volume change in material
samples at each chemical/temperature
combination are measured every four months
in a two year exposure test. The samples
are also visually inspected for excessive
degradation. Where possible, at the end of
the two year test period, the tensile
properties of the exposed samples will be
measured.
Data from these immersion tests are studied
to determine the correlation between
different types of measurements and tests,
and to identify major indicators of
membrane degradation. The effect of
chemical type, temperature, concentration,
and time on material response is studied
both to better understand the mechanism of
change and to evaluate the use of
accelerated test procedures.
261
-------
ASSESSMENT OF SYNTHETIC MEMBRANE SUCCESSES AND
FAILURES AT WASTE STORAGE AND DISPOSAL SITES
Jeffrey M. Bass, Warren J. Lyman, Joseph P. Tratnyek
Arthur D. Little, Inc.
Cambridge, MA 02140
ABSTRACT
Data provided by five vendors, from
27 lined facilities, were analyzed to
determine the factors which contributed to
success or failure of the liner at each of
those facilities. The sites represented a
wide variety of wastes handled, liner
types, geographic locations, facility
ages, facility sizes, etc. Based on the
definitions used in this study, the 27
facilities selected by the vendors had a
total of 12 "failures" at 10 sites. At
four or five of these sites groundwater
contamination apparently resulted from the
failures. Some of the contributing
factors, if not causes, for the failures
include the following:
t Failure to control operations (at an
operating site) so as to safeguard
the liner;
• Poor (or inadequate) design work;
t Failure to use an independent,
qualified design engineer;
• Poor (or inadequate) installation
work;
• Poor or inadequate communication and
cooperation between companies working
on an installation job;
• The use of untrained and/or poorly
supervised installers;
• Failure to conduct (or adequately
conduct) waste-liner compatibility
tests;
t Adverse weather conditions during
installation;
• Use of old dump site, with
contaminated soil, as site for lined
facility;
• Selection of companies (for liner
job) by processes that did not help
ensure that good materials and
workmanship would result;
• Selection of liner material by
process not involving detailed bid
specifications (prepared by design
engineer, not liner manufacturer);
• Facility age (more failures were
associated with the older sites).
Two main elements of membrane liner
success are: (1) a proper philosophical
and conceptual approach; and (2) the
extensive use of quality assurance
programs in all facets and stages of a
facility's operation. Other factors noted
as contributing to success included:
• Overdesign of a system;
t A knowledgeable customer;
• Bidding to specifications;
• Selection of qualified companies;
• Cooperation amongst companies on
liner job;
• Conducting waste-liner compatibility
tests;
• Simplicity of design; and
t Good weather.
262
-------
RELATIONSHIP OF LABORATORY AND FIELD DETERMINED HYDRAULIC
CONDUCTIVITY (K) IN COMPACTED CLAY SOILS
Andrew S. Rogowski-
USDA-ARS, Northeast Watershed Research Center
University Park, PA 16802
ABSTRACT
Details of the field scale clay liner
research facility and liner installation
are given. The facility consists of a
30 x 75' bridge-like platform. It is in-
strumented with a grid of underdrains,
gypsum blocks, infiltration rings, sites to
measure swelling, density probe access
tubes, inflow-outflow recorders, and evap-
oration pans. The clay is from the B-
horizon of a Typic Hapludult (USDA
Classification), or type CL (USCS
Classification) soil, with a laboratory-
determined hydraulic conductivity
K = 1 x 10~8 cm/sec. It was compacted to
90% Standard Proctor using a crawler
tractor pulling a sheepsfoot roller. The
clay liner was placed in three lifts of 6"
uncompacted thickness each. To facilitate
quality control measurements with the nu-
clear density meter, the surface of each
lift was smoothed with a smooth-face steel
wheel compactor. Following the soil
density readings, the lift surface was
scarified by tracking with a bulldozer
before the material for the next lift was
applied.
Instrumentation to monitor moisture
movement in the field facility was based
on a^ priori prototype studies conducted
in TO-feet diameter caissons under roof.
These prototype studies have underscored
the importance of adequate sealing of
interfaces and elimination of liner per-
forations such as for example, wells or
moisture-density access tubes. Prototype
data suggest considerable variability from
point to point. Consequently, a regular
grid was imposed on the surface of the
facility and measurements are being made
at the specific times and locations.
Using geostatistical techniques, areal
distribution of infiltration and possible
outcomes for differing infiltration or
density scenarios can be projected.
A Standard Proctor mould was scaled-up
using one-half of a 40-gallon steel drum.
After soil was compacted into this mould, a
12" diameter ring was placed, which was
filled with water to measure infiltration
rate; the surface outside the ring was also
flooded to provide the double-ring
infiltrometer. The ring is the same size
as will be placed at specified locations in
the field facility. Hydraulic conductivity
values obtained from the scaled-up mould
will be compared with the distribution of
values obtained both from the field
facility and laboratory determinations.
- Contribution from the U.S. Department of Agriculture, Agricultural Research
Service, in cooperation with the Pennsylvania Agricultural Experiment Station, The
Pennsylvania State University, University Park, PA 16802, supported by EPA-ARS
Interagency Agreement Funds: EPA-IAG-DW12930303-01-0.
263
-------
CRITICAL REVIEW AND SUMMARY OF GAS AND LEACHATE PRODUCTION
FROM MUNICIPAL SOLID WASTE LANDFILLS
Frederick G. Pohland
and
Stephen R. Harper
School of Civil Engineering
Georgia Institute of Technology
Atlanta, Georgia 30332
ABSTRACT
The purpose of this project was to
provide a review and critique of literature
and the status of technology pertaining to
the formation, characterization and
treatment of municipal solid waste (MSW)
landfill gas and leachate. This objective
was accommodated by an initial recognition
of the various factors influencing the
applicability and efficacy of various
landfill disposal options and/or leachate
and gas management strategies. Therefore,
the review is organized to reflect bench-,
pilot- and full-scale evaluations;
leachates have been categorized according
to low, medium and high strengths; and
product gases have been differentiated on
the basis of constituents and medium or
high energy content.
Commencing with a brief historical
perspective focusing on the problems
associated with migration of MSW leachate
and gas from landfill disposal sites, the
review proceeds with a characterization
of MSW gas and leachate quality produced
throughout the phases of landfill stabi-
lization and the various leachate and gas
treatment methods. Accordingly, leachate
treatment by external aerobic and anaerobic
biological methods and by _in situ anaerobic
degradation with leachate management are
compared to physical-chemical leachate
treatment by chemical oxidation, precipi-
tation, coagulation, disinfection,
absorption, ion exchange and reverse
osmosis. In addition, methods of landfill
gas collection and treatment by sorption,
refrigeration, and membrane separation
are presented together with the implica-
tions of ultimate disposal of treated
leachates, including an understanding of
leachate/soil interactions.
Biological treatment is shown to be
superior to physical-chemical methods for
raw leachate; aerobic and anaerobic treat-
ment methods were generally capable of
removing 85 to 99 percent of leachate BOD^
and COD as well as most heavy metals. The
physical-chemical processes are shown to
be better suited for polishing of biologi-
cal treatment effluents and for the removal
of refractory organics. Leachate manage-
ment through recycle is also demonstrated
to possess a high potential for process
stability, although further effort is
needed to determine application criteria on
a large scale. Process loadings, opera-
tional constraints, and degree of variance
and uncertainty in reported data and their
interpretation are presented to indicate
that past efforts have not resolved all the
technical challenges associated with gas
and leachate management at landfill
disposal sites, and that additional
investigations are required.
264
-------
HYDRAULIC CONDUCTIVITY OF TWO PROTOTYPE CLAY LINERS
Steven R. Day
Geo Con, Inc.
Pittsburgh, PA 15235
David E. Daniel
University of Texas
Austin, Texas 78712
ABSTRACT
Two prototype compacted clay liners
were constructed at a site near Austin,
Texas, using clays of high and low plasti-
city. The liners were 15 cm (6 inches)
thick and covered an area of approximately
6 meters by 6 meters. The clays were
compacted to 100% of standard Proctor den-
sity at a water content slightly wet of
optimum using a sheepsfoot roller. Small
dikes were constructed around the peri-
meter of the-test area, and water was
ponded on the liners. An underdrain
system was installed to collect liquid
that flowed through the liner and to
measure the quantity of seepage. The
underdrain consisted of an impermeable
geomembrane, freely draining filter fabric,
washed gravel, and perforated collection
pipes. The collected liquid flowed by
gravity to a pit, where the quantity of
flow was measured and was used to calcu-
late hydraulic conductivity.
The actual hydraulic_conductivity of
the two liners was 4 x 10" cm/sec for ,
the clay of high plasticity and 9 x 10"
cm/sec for the clay of low plasticity. A
variety of laboratory permeability tests
were performed. The tests included
permeability measurements on laboratory-
compacted specimens of clay, on samples
of the actual clay liners obtained with a
thin-walled sampling tube, and on hand-
carved blocks of clay obtained from the
actual liners. Field permeability tests
were also performed using single- and
double-ring infiltrometers.
The various laboratory and field
permeability tests showed that: (1)
essentially all of the laboratory tests,
even on undisturbed samples, produced a
measured hydraulic conductivity that was
one thousand times less than the actual
hydraulic conductivity, and (2) field
permeability tests yielded an average hy-
draulic conductivity that was close to the
actual hydraulic conductivity.
The findings of this study raise
important questions about whether labora-
tory permeability tests on compacted clay
are relevant to field problems and re-
inforce previous suggestions that
compacted clay liners may contain numerous
hydraulic defects such as fissures,
slickensides, zones of poor bonding
between clods of clay, and zones of
relatively poor compaction. The desir-
ability of field permeability tests is
evident from the results reported.
Permeability tests on laboratory samples
that were tested at different effective
stresses suggest that clay liners sub-
jected to significant overburden pressure
may retain their hydraulic integrity much
better than clay liners that have little
or no overburden pressure.
265
-------
CLAY-CHEMICAL COMPATIBILITY AND PERMEABILITY TESTING: A REVIEW
Garrie L. Kingsbury
Robert S. Truesdale
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
ABSTRACT
It has been recognized for many
years that certain chemicals can affect
the permeability of clay soils. Several
EPA-sponsored studies of the effects
of selected organic and inorganic
chemicals on clay soils have demon-
strated the potential of some chemicals
to increase clay permeability.
Additionally, recent evidence from a
hazardous waste disposal site in the
United States suggests that the presence
of an organic solvent, which formed a
separate, denser than water phase in
a clay-lined surface impoundment,
contributed to the failure of the liner.
Although several researchers have
suggested explanations for the observed
chemically-induced changes in permea-
bility, our review of the state-of-the-
knowledge in this area concludes that
current theories are inadequate to
predict the extent to which complex
chemical mixtures may affect clay soil
permeability. Qualitative predictions
can be made of the effects of certain
concentration chemical solutions and
nearly-pure solvents on selected clay
soils. Previous reviews have identi-
fied mechanisms that could account for
observed changes in permeability; use
of this information to interpret test
results and to predict chemical effects
on clay permeability are complicated
by the possibility that two or more
of the processes may operate simul-
taneously.
The issues surrounding measurement
and interpretation of permeability and
chemically-induced changes in permea-
bility are complex and EPA is sponsoring
a number of studies on these topics. In
January 1984, a 2-day workshop was
convened to exchange research experience,
discuss current findings, and to
coordinate on-going and future work.
Among the principle topics discussed
were the following:
Need for standard procedures
Importance of sample preparation
Effect of conditions (saturation,
hydraulic gradient, etc.) on
tests results
Standard permeant solutions
Chemistry of inflow and outflow
solutions and soil sample
Criteria for concluding compati-
bility tests without missing
significant changes
Criteria for appropriate test
devices
High priority chemicals for
further testing
At least three different types of
test devices (fixed-wall, flexible wall,
and consolidation cells) are being used
to measure clay soil permeability in
the laboratory and a variety of differ-
ent sample preparation methods and test
procedures have been employed. The
different approaches to measuring the
effects of chemicals on clay permea-
bility and the sensitivity of the tests
266
-------
to a large number of variables has lead
to test results that are often contra-
dictory and difficult to interpret.
There is a divergence of opinion
regarding the most appropriate labor-
atory test device. The large increases
in permeability observed in tests
using fixed-wall apparatus with certain
organic liquids are attributed by some
researchers to side-wall flow resulting
from shrinkage of the clay sample on
exposure to the chemicals. Although
the lateral confining pressure in
flexible wall cells prevents side-wall
leakage, it may also prevent the
development of shrinkage cracks that
could be caused by the chemical
displacement of water from the soil.
The pressure applied to the surface of
a sample in a consolidation cell may
reduce leakage by forcing the sample
into contact with the side-wall. There
is some question about overburden
pressures for use in testing clay liner
material in consolidation cells.
Because of the effect of soil
fabric on permeability and the
difference in soil fabric between
laboratory-compacted and field-
compacted samples, much attention is
being given to appropriate sample
preparation for laboratory testing and
to procedures for conducting tests in
the field.
This review of the status and
results of current research on clay-
chemical compatibility and permeability
testing was conducted as part of the
production of a Technical Resource
Document on design, construction, and
evaluation of clay liners for waste
management facilities.
267
-------
CLAY LINER TRANSIT TIME PREDICTION METHODS AND EFFECTS
OF LEACHATE COMPOSITION ON LINER INFILTRATION
Ashok S. Damle
Scott M. Harkins
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, North Carolina 27709
ABSTRACT
In order to evaluate an existing
clay-lined facility or to design a new
facility, two approaches to defining
the transit time of leachates moving
through a clay liner are recognized.
The first approach is based on the
movement of water or the saturation
front in the liner. An alternative
approach is based on the movement of
leachate chemicals. For this project,
the available prediction methods for
transit time were evaluated for their
applicability to either approach. The
initial degree of liner saturation
(water content), as well as the defini-
tion of transit time, determine the
applicability of a given method.
Five analytical methods, as well
as the numerical solution of the
governing flow and solute transport
equations, were reviewed. Three of
the analytical methods solve both the
solute transport and water flow
equations and are based on the assump-
tion that the liner material is
completely saturated. Because the
difference between molding water content
and saturation are often small in
compacted clays, these methods may be
applied to partially saturated clays
without major error. The other two
methods describe the infiltration
dynamics in unsaturated materials but
do not take into consideration solute
transport.
Limitations in applying these
methods in clay liner design arise from
the data requirements and assumptions
that must be made. For example, all
the transit time prediction methods
assume a homogeneous liner and do not
consider the effect of leachate compo-
sition directly. Numerical methods make
fewer assumptions, but are expensive,
and the accuracy achievable depends
upon the quality of available data.
Numerical solutions also require specific
expertise in available models.
Since clay liners are compacted
wet of optimum (i.e., at a moisture
content slightly greater than the
moisture content at which maximum
density can be achieved) their condi-
tion differs from the completely
saturated (zero air void) clay systems
that are used typically to determine
clay permeabilities. Thus, test data
from unsaturated compacted clay soils
are needed to assess the reliability
of computer models for predicting
pollutant transport in unsaturated clay
soils.
Clay-lined hazardous waste land-
fills accept chemical wastes which
produce leachates containing many
chemical components. These chemicals in
the leachate may contact and migrate
into the clay liner where they could
affect clay structure and ability to
268
-------
control movements of fluids (permea-
bility). Because of this, a companion
investigation was performed to deter-
mine if leachate composition can
significantly change the rate of
liquid movement during infiltration
into unsaturated compacted clay liners.
Three native clay soil samples,
compacted in plexiglass cylinders
(4.25" OD and 5" long) were exposed to
test liquids under a constant 1 foot
head. The solutions tested were 0.01N
CaS04 (control), sulfuric acid (1 per-
cent), sodium lauryl sulfate
(1 percent), and methanol (8 percent).
After several months of exposure, the
soil columns were extruded and sliced.
Analyses were performed to determine
moisture content and chemical com-
ponents at various depths in the soil
column. The rate of liquid movement
into the clay was calculated from
these analyses and any liquid flow out
of the clay was measured.
269
-------
EFFECTIVE POROSITY OF COMPACTED CLAY SOILS PERMEATED WITH
ORGANIC CHEMICALS
Janice W. Green
Arthur D. Little, Inc.
Cambridge, MA 02140
and
Kirk W. Brown and James C. Thomas
Soil and Crop Sciences Department
Texas A&M University
College Station, Texas 77843
ABSTRACT
Total porosity has commonly been
considered as the available pathway for
flow of impounded liquids through
compacted clay soil liners. The
validity of this assumption was
investigated by estimating the fraction
of active total pore space in several
soils for conduction of permeant
liquids. The morphology of three
compacted sandy clay loam soils
permeated with three liquids, e.g.,
water, acetone, and xylene wastes, is
discussed and illustrated. The soils
are differentiated by their dominant
clay mineralogy: kaolinitic, bentoni-
tic, or micaceous. Total porosity was
obtained from bulk density calculations
Effective porosity was estimated from
microscopic observations, which
included estimates from (1) Planimeter
measurements on micrographs from
scanning electron microscopy
(Backscatter mode), (2) Ribbon-traverse
counts on thin sections, observed under
epifluorescent, and (3) Petrographic
light- microscope observation. In some
cases, less than 10% of the total
porosity was effective in transmitting
the liquids. Photomicrographs showing
examples of observations from the
microscopic techniques are presented.
The macromorphology of the soils
after permeation is illustrated in
photographs of excavated field test
cells where compacted clays were exposed
to industrial solvent paint wastes. Two
dyes, one visually detectable and the
other fluorescent under ultraviolet
light, were added to each waste liquid
at the beginning of the study. At the
time of excavation of the simulated
liners, visual observations were made of
the fluid flow path through the
compacted clay layers using both the
visible dye and pigments contained
within the waste liquids. The dominant
macroscopic flow paths through the
kaolinitic and micaceous xylene-waste
permeated soils were planes that were
believed to have been induced by the
compaction process. Microscopic studies
supported the visual observation.
Effective porosity predominated in pores
greater than 125 micrometers average
diameter, which comprised less than 1%
of the total soil volume for these soils
with xylene-based waste.
The bentonitic soil demonstrated a
higher effective porosity in the
presence of xylene but was still less
than 3% of the total soil volume.
Visual observations showed more numerous
soil surfaces dyed within bentonitic
soil than those of the other two
mineralogies, suggesting a greater
effective porosity.
Less than 2% of the total soil
270
-------
volume was effective in transmitting
acetone or water in micaceous and
kaolinitic soils. Nearly 5% of the total
soil volume was effective in transport-
ing both water and acetone in the
bentonitic soil. Soils permeated with
acetone-waste in field cell studies
contained essentially the same degree of
effective porosity as the same soils in
laboratory permeameters.
Total porosity was measured from
portions of each soil taken near the
surface of the compacted bulk, which has
been shown by other investigators to be
the most dense portion of the soil and,
thus, the most restrictive to liquid
flow. Total porosity in each of the
three soil studies was as follows:
bentonitic soil - 14.5%; kaolinitic soil
- 11.6%; and micaceous soil - 13.2%.
The compacted clay soil pore space that
was active in allowing liquid to flow
through the soil was demonstrated, by
the available data, to be substan-
tially less than the pore space
calculated as the total porosity.
271
-------
INFLUENCE OF CONCENTRATIONS OF ORGANIC CHEMICALS ON THE COLLOIDAL
STRUCTURE AND HYDRAULIC CONDUCTIVITY OF CLAY SOILS
Kirk W. Brown and James C. Thomas
Soil and Crop Sciences Department
Texas A&M University
College Station, Texas 77843
ABSTRACT
The impact of concentrated organic
chemicals on the permeability of
compacted clay landfill and pond liners
is now reasonably well documented. The
response of clays to concentrations of
nonpolar compounds which do not exceed
their water solubility levels have been
shown to be similar to that of water. As
the concentration of polar compounds
increase, their impact on the
permeability of clay liners appears to
shift from that of water to that of the
specific organic liquid.
The present study was undertaken to
determine the concentrations at which
this shift in behavior occurs and to
evaluate the mechanisms that regulate the
shift. Three different clays were
suspended in a series of aqueous
dilutions of acetone and ethanol in
hydrometer jars. The percent apparent
clay was used as an index of the
flocculation of the clays. In all cases,
the clays remained dispersed in water,but
as the concentration of the organics
increased, a concentration was reached
above which the clays quickly floccula-
ted and settled to the bottom of the
cylinders. Flocculation occurred at
dielectric constants in the range of 30
to 50 for all the systems. Conductivity
measurements made in fixed wall
permeameters with similar dilutions on
the clays indicated that marked
increases occurred at concentrations
above which the clay flocculated.
The results indicate that when
organics are sufficiently concentrated,
the clay will flocculate and result in
shrinking, cracking and large increases
in hydraulic conductivity. Flocculation
tests or dielectric constants may
possibly be used as a tool to predict
the influence of a given solution on the
behavior of clay liners. However, more
accurate indications of clay liner/
leachate compatibility may be possible
by using other parameters which better
characterize the interaction of clays
and solutions.
272
-------
LABORATORY DETERMINATION OF DIELECTRIC CONSTANT AND
SURFACE TENSION AS MEASURES OF LEACHATE/LINER COMPATIBILITY
Charles I. Mashni, Howard P. Warner, and Walter E. Grube, Jr.
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Research on hydraulic conductivity of
compacted clay soils using various permeant
liquids has shown the trend that as liquids
of lower dielectric constant are employed,
the soil hydraulic conductivity increases.
These observations suggested that there may
be a reliable relationship between the clay
soil's performance as a landfill liner and
readily measured properties of leachate
liquid which the liner must contain. Both
industry and agencies would benefit from
the confirmation of a rapid and easy test
to measure and/or predict liner/leachate
compatibility.
This report compiles, from published
technical reports and laboratory studies,
dielectric constant and surface tension
measurement data for several common organic-
solvents and selected binary mixtures.
Dielectric constant was examined both
because published reports cited this key
parameter in relation to clay soil perme-
ability and a commercial instrument was
available to measure this property of
various liquids.
Surface tension was measured because
of the potential relationship to pore entry
pressure of non-aqueous materials in miner-
al matrices, such as clay soils, and the
relatively easy laboratory measurement.
Data will be presented which demon-
strate the laboratory performance of a
dielectric constant meter and surface
tensiometer using various solvents and
mixtures. Operational precautions and
instrumental peculiarities will be dis-
cussed. Data generated will be compared
with published reports of the hydraulic
conductivity of compacted clay soils using
liquids the same as or similar to those
examined in this study.
273
-------
QUICK INDICATOR TESTS TO CHARACTERIZE BENTONITE TYPE
Andrew Bodocsi and Richard M. McCandless
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Bentonite-soil mixtures are a primary
element of slurry wall waste containment
barriers. With several different bentonite
products commercially available, selection
and specification of a particular bentonite
type is an application-dependent decision.
Cost and technical performance considera-
tions are frequently applied only on a sub-
jective basis. Commercial product descrip-
tions generally categorize bentonites as
ordinary (unaltered), physically altered,
polymer protected or extended, or chemical-
ly resistant types. Currently there are no
known quick field laboratory test protocols
which allow a design engineer's quality
control program to distinquish bentonite
type or to evaluate in-service performance
under chemical or leachate attack.
The purposes of this project are: 1)
to identify a practical indicator test(s)
to distinguish among bentonite types; 2)
to develop a data base which characterizes
various bentonite types in terms of chemi-
cal/leachate impact; and 3) to relate
measured impact to in-service hydraulic
conductivity of the soil-bentonite mixture.
Research to date has focused on ben-
tonite type distinction and chemical/
leachate impact characterization. Ten
conventional laboratory classification and
index tests employed in environmental
science, industry, and geotechnical engi-
neering were selected for study. Data
characterizing the impact of three liquids
on various test parameters were generated
for ten of approximately 42 commercially
available bentonite products.
Preliminary findings on the basis of
limited data obtained to date are: (1)
Tests designed to measure bentonite slurry
gel strength, free swell potential, and
various viscosity parameters appear to have
strong potential for distinguishing among
bentonite types; (2) Bentonite hydration in
a 3:1 water:acetone solution (25% acetone)
enhanced desirable engineering properties
of bentonite relative to water hydration
results in about 85 percent of the tests;
(3) Bentonite hydration in a 1:3 water:
acetone solution (75% acetone) degraded
desirable bentonite properties relative to
water hydration in about 69 percent of the
tests; (4) For bentonites hydrated in
water or acetone:water solutions, chemical-
ly resistant, ordinary (unaltered), and
physically altered amine-treated benton-
ites were distinguishable from the other
types tested.
Future research through 1985 will
involve refinement of bentonite type clas-
sifications, different water:acetone
leachate concentrations, evaluation of
other potentially useful indicator tests,
and bentonite characterization on the
basis of chemical or leachate impact after
full water hydration.
274
-------
ESTIMATING TRANSIT TIMES OF NONINTERACTING
POLLUTANTS THROUGH COMPACTED SOIL MATERIALS
Robert Horton, Michael L, Thompson, and John F, McBride
Department of Agronomy
Iowa State University
Ames, Iowa 50011
ABSTRACT
Pollutant travel times through compacted soil materials cannot be accurately pre-
dicted from the permeability (saturated hydraulic conductivity) alone. Travel time is
also dependent on the effective porosity of the material; i.e., the portion of the total
porosity that contributes significantly to fluid flow. Once permeability and effective
porosity are determined for a selected material, travel times for noninteracting pol-
lutants through specified thicknesses of compacted material at specified hydraulic gra-
dients can easily be predicted. This paper presents a straightforward method of deter-
mining the effective porosity of compacted soil materials and compares measured and pre-
dicted solute breakthrough times for three compacted soil materials.
The determination of effective porosity is based upon the total porosity and the
spread on a log scale in the pore sizes of a compacted sample. Once the total porosity
and pore size distribution information are obtained for a particular sample, the effective
porosity can be determined by using a graphical solution.
Pollutant travel time, T, through a compacted soil (i.e., when pollutant first ap-
pears at the bottom of the liner) can be predicted by T=EL/KI; where E is effective poros-
ity, L is thickness of compacted sample, K is sample permeability, and I is hydraulic
gradient. Chloride travel times through compacted samples of glacial till, loess, and pa-
leosol materials were measured as less than 208 minutes (min.), between 15 and 30 min.,
and between 250 and 380 min., respectively. Predictions of noninteracting solute break-
through for the glacial till, loess, and paleosol materials were 227 min., 32 min,, and
270 min,, respectively. Thus, the predictions are reasonably close to measurements.
Further experiments using thicker soil samples are under way.
INTRODUCTION AND PURPOSE
Compacted clay is frequently used to
line hazardous-waste-disposal landfills.
The purpose of the compacted clay liner is
to restrict the movement of hazardous li-
quid materials out of the landfill. Cur-
rently, soil permeability (saturated hy-
draulic conductivity) is the most common
measurement used to estimate the potential
effectiveness of the liner material (USEPA,
1978). Soil permeability, K (m3/m2/s), is
defined from Darcy's equation:
K= J/I (1)
where J is the fluid flux density (m /m /s)
and I is the hydraulic gradient (m/m),
Soil permeability has dimensions of volume
per unit area of liner per unit time; thus,
it is a bulk parameter related to the areal
average fluid flow in the liner, But know-
ledge of average fluid flow does not allow
the prediction of pollutant breakthrough or
travel times. To predict pollutant travel
times through clay liners, the liner ef-
fective porosity must be known.
Effective porosity, E, has been des-
cribed as that portion of the total liner
porosity that contributes significantly to
fluid flow. Effective porosities are less
than total porosities because some of the
pore space is discontinuous (dead end) and
some of the pore space is so narrow that
fluids in these spaces are essentially im-
mobile. If the fluid flux density through
a clay liner can be determined by using
275
-------
Darcy's equation^ then the mean effective
fluid velocity, VE, can be described by:
V?= KI/E (2)
If the hydraulic gradient is constant and
liner K and E are both known, then the
average pollutant travel distance per unit
time can be estimated. Thus, the time of
first breakthrough, T, of a noninteracting
pollutant can be predicted by the equation:
T = EL/KI (3)
where L is the liner thickness (m),
The Environmental Protection Agency,
EPA, requires that a disposal unit liner
prevent migration during the active life of
a unit (USEPA, 1982). The active life of a
unit containing noninteracting pollutants
can be estimated by using Eq. (3). Know-
ledge of the liner permeability is not
sufficient information to accurately
estimate the length of time of liner
effectiveness. When a disposal unit is
constructed with known liner thickness, L,
and is designed for a known hydraulic gra-
dient, I, measurement of both permeability
and effective porosity are required to es-
timate the active life of the unit with Eq.
(3). Although methods of measuring perme-
ability of compacted clay materials have
received much attention in the literature
(Olson and Daniel, 1981), determination of
effective porosity has received little at-
tention.
The purpose of this paper is to pre-
sent a method for estimating the effective
porosity of compacted clay materials. The
estimates of effective porosity are used in
Eq. (3) to predict transit times of nonin-
teracting solutes through samples of com-
pacted materials. The predictions of tran-
sit time are compared with measured chlo-
ride breakthrough times,
APPROACH
To estimate the travel time of non-
interacting pollutants through a liner by
Eq, (3), the liner's effective porosity
must be determined. We earlier described
effective porosity as that portion of total
porosity that contributes significantly to
fluid flow. We now define "significance"
by defining effective porosity as that por-
tion of total porosity that conducts fluid
faster than the average pore-water veloci-
ty. Large pores conduct fluid faster than
do small pores. Therefore, in a porous me-
dium that exhibits a range of pore sizes, a
range of pore-water velocities will exist
as a fluid passes through, Higher-than-
average pore-water velocities in the larger
pores will be responsible for the first ap-
pearance of pollutants, whereas lower-than-
average pore-water velocities associated
with smaller pores will not be responsible
for first appearance of pollutants. Thus,
the determination of a liner's porosity
versus pore-water velocity distribution is
crucial to the evaluation of the liner's
effective porosity.
The pore-water velocity distribution
can be calculated from the unsaturated per-
meability, K(6), as a function of fluid-
filled porosity,6 . On the basis of theory
developed by Mualem (1976), Van Genuchten
(1980) showed that the unsaturated perme-
ability could be predicted from the total
porosity, the residual porosity (i.e., the
value of porosity when water films lose ef-
fective continuity), and a measure of the
uniformity of pore sizes by the equation:
e-e t
where 6 is the total porosity, QT is the
residual porosity, and m = [l-(l/n)] for
0
-------
we define 0 as equal to total porosity of
the freeze-dried samples and transform the
mercury-intrusion cumulative porosity to a
form similar to the soil-water characteris-
tic curve as follows: 6=6- mercury-
filled porosity. Once the cumulative
porosity curves are transformed, the 6, n,
and a parameters(Eq. 6) can be determined by
using nonlinear regression techniques.
When nonlinear regression techniques
are not accessible, a first approximation
of the 6 and n parameters can be deter-
mined from a graphical display of the
transformed data. The residual porosity,
0 , is approximated as the value of 6
that the cumulative porosity curve asympto-
tically approaches as limiting pore size
decreases. Mercury-intrusion cumulative
porosity curves for compacted soil samples
generally have 6r -values approximately
equal to their non-mercury-filled porosi-
ties after intrusion to 414 MPa (60,000
psi). For soil materials considered for
landfill liners, 9r is generally less than
0,1, The n-parameter is estimated as fol-
lows: The cumulative porosity data are
plotted as 0 vs. log r; next, the absolute
value of the slope, S, of the cumulative
porosity data is calculated at a value of
approximately halfway between 6.r and 0g.
The n-parameter is then determined by using
the equation:
f" exp (0.8 Sp) • (0l)
Sp Sp2 Sp3
where Sp is equal to S/(0S -6r).
Once 0 and n are estimated from the
measured cumulative porosity data, the sam-
ple effective porosity can be determined
from Figure 1 (for a detailed explanation,
see Horton, 1985). With an estimate of the
effective porosity, the time required for
the first appearance of noninteracting pol-
lutants at the bottom of a clay liner can
be predicted by using Eq. (3).
Soils
We collected subsurface materials from
three soils developed in major Iowa parent
materials: till, loess, and paleosol. The
till-derived soil was Nicollet composed of
sandy clay loam materials (Unified Classi-
fication: CL; AASHTO classification: A-4).
The loess-derived soil was Fayette composed
. 1 5
. 1 2
. 09
. 06
. 03
. 00
8g=0.3
1 . 0
1 . 5
2 . 0
. 5
3 . 0
3 . 5
n
Figure 1. Influence of n on effective porosity (E) at various values of 6 (9 - 0.05),
S X*
277
-------
TABLE 1
PARTICLE SIZE DISTRIBUTION, OPTIMUM MOISTURE CONTENT
AND BULK DENSITY OF THE COMPACTED SOIL MATERIALS
Soil
Material
Clay
(<2ym)
Silt
(2-50ym)
V _
Sand
(50-2000ym)
Optimum
moisture
content
Bulk
density
(g/cm3)
Nicollet
Fayette
Clarinda
21
35
44
25
57
35
54
8
21
14
18
27
1.84
1.67
1.45
of silty clay loam materials (Unified clas-
sification: CL; AASHTO classification: A-
7), Clarinda was an exhumed paleosol com-
posed of clay materials (Unified classifi-
cation; CH; AASHTO classification: A-7).
These soils were chosen to be representa-
tive of the kind of materials likely to be
used in construction of landfills in Iowa,
The three soil materials were distin-
guished by their particle size distribution
(Table 1), Nicollet materials had the
least clay and the most sand, whereas Cla-
rinda materials had the most clay and an
intermediate sand content, Fayette mate-
rials, derived from loess, were highest in
silt.
Soil materials were collected by exca-
vation, air-dried, and ground to pass a 2-
mm sieve. A standard moisture-density re-
lationship for each soil material was de-
termined according to ASTM standard test D-
698-78 Method A. The optimum moisture con-
tents and bulk densities are reported in
Table 1.
Porosity
Samples were prepared by compacting
soil materials at moisture contents 1 to 2%
higher than optimum. Compacted samples
were broken into fragments <3 cm . The
fragments were freeze-dried according to
the method of Zimmie and Almaleh (1976),
Cumulative porosity of triplicate
fragments of each soil material was mea-
sured by using a Quantachrome SP-200 mer-
cury porosimeter in four intrusion steps: 0
to 0.1 MPa, 0.1 to 8,3 MPa, 8.3 to 41,4
MPa, and 41.4 to 414 MPa (60,000 psi). Cu-
mulative porosities were plotted as cumula-
tive volume vs. log pore radius,
Chloride Breakthrough Measurements
From the moisture-density relation-
ship, a value one or two percentage points
above optimum moisture was chosen at which
to moisten the soil. The soil was moulded
in a 6,35-cm diameter by 2,54-cm high cy-
linder at the corresponding bulk density.
Each compacted soil sample was placed
in a 7.62-cm diameter acrylic permeameter.
Paraffin, kept at 60°C, was poured between
the wall of the permeameter and the sample
in 0.5-cm increments until the top of the
compacted sample was reached; the paraffin
was allowed to solidify between increments.
In this manner, any gaps formed when the
paraffin cooled were filled by the next
increment, thus minimizing wall effects.
De-aired, saturated CaSO* solution
(adjusted to 0.06% formaldehyde, to control
microbial growth) was introduced into the
permeameter and pressurized from an air-
pressure source. The air was separated
from the solution by means of a rubber mem-
brane within the solution container. Hy-
draulic gradients in the range of 100-200
were used; because of the relatively small
sample height, this required a hydraulic
head of approximately 2,5-5.0 m H20 (3.5 -
7,5 psi). Once the hydraulic gradient was
established, a minimum of 5 pore-volumes
was passed so as to obtain a constant flux
from which to determine the sample perme-
ability.
A de-aired, 0,lN^CaCl2 solution with 0.016%
278
-------
Acid Fuchsin, a red dye, and 0.06% formal-
dehyde was used as the tracer solution for
the solute breakthrough studies. The dye
gave a visual test for any wall effects.
The tracer was exchanged for the CaSO^
solution and allowed to leach through the
soil sample under the same hydraulic
gradients. The leachate was collected in
equal-volume increments with a fraction
collector; a minimum of 3 pore-volumes was
collected. The leachate was analyzed for
chloride concentration with an automatic
titrator,
PROBLEMS ENCOUNTERED
The problem of greatest concern was
whether soil porosity was altered during
sample drying preceding mercury porosimetry
measurements. Our method of determining
effective porosity depends on both the cu-
mulative pore size distribution and the
total porosity. Yet, to determine pore
size distribution by mercury porosimetry,
all water must be removed from the sample.
Some of the total porosity in the samples
was lost when water was removed by freeze
drying (Table 2).
TABLE 2 TOTAL POROSITIES OF COMPACTED SOIL
MATERIALS BEFORE AND AFTER DRYING
Soil
Materials
Nicollet
Fayette
Clarinda
Before
drying
0.29
0.40
0.46
After
drying
0.28
0.40
0.40
As Table 2 shows, some loss of poros-
ity occurred in the Nicollet and Clarinda
samples during freeze drying. The greatest
decrease in porosity was with Clarinda soil
materials, the materials highest in clay
content. Total porosity in Fayette soil
materials was not diminished by freeze dry-
ing. The significance of the losses in
total porosity to calculations of effective
porosity does not seem to be large.
RESULTS
Only the n-parameter is required to de-
scribe the unsaturated permeability.
Figure 2 presents a family of curves over a
range in n from 1,1 to 5, The curves show
that the spread on a log scale of pore
sizes decreases as n increases. Thus, it
becomes clear that the theory provided for
calculating unsaturated permeability
considers the spread in pore size
distribution as the predominant factor.
.001 .01 .1 1
PORE RADIUS (urn)
100
The function used to describe the cu-
mulative porosity curve (Eq, 6) has two em-
pirically determined parameters,a and n.
Figure 2. Theoretical cumulative porosity
curves as influenced by the n-
parameter (a= 1.0, S= 6/9 ).
s
Figures 3, 4, and 5 show the cumula-
tive porosities for Nicollet, Fayette, and
Clarinda soil samples, respectively. The
discrete, plotted symbols represent values
determined by mercury porosimetry. The
solid curves represent the least-squares
fit of Eq, (6) to the measured data. In
all cases, the model does a reasonable job
of matching the measured data. The a, n,
6r and 0 parameters for each model are pre-
sented in Table 3.
Equation (4) describes the unsaturated
permeability of soil once the n-parameter
is known. Figure 6 presents a family of
curves displaying the influence of the n-
parameter on unsaturated permeability. As
n increases, the relative permeability
tends to decrease more gradually as the
soil begins to desaturate. This indicates
that, the larger the n-parameter, the larg-
er is the fraction of porosity with rela-
tive importance for conducting fluids. As
n increases, one can expect the effective
porosity to represent a larger portion of
the total liner porosity.
Figure 1 presents a family of curves,
over a range in 6S from 0.3 to 0.5, dis-
279
-------
.27 -
.01 .1 1
PORE RHDIUS (uM)
Figure 3. Measured and curvefitted cumu-
lative porosity for compacted
Nicollet soil samples.
. 09
FRYETTE COMPRCTED
. 001
.01 .1 1
PORE RflDIUS (uM)
10
Figure 4. Measured and curvefitted cumu-
lative porosity for compacted
Fayette soil samples.
o
g
(L
.36
. 27
. IB
CUflRINDR COMPRCTED
. 001
.01 .1 1
PORE RHDIUS (uM)
100
Figure 5.
Measured and curvefitted cumu-
lative porosity for compacted
Clarinda soil samples.
TABLE 3 PARAMETERS USED WITH EQUATION (6)
TO DESCRIBE THE CUMULATIVE PORO-
SITY OF THE SOIL MATERIALS.
Soil
Materials
Nicollet
Fayette
Clarinda
* * *
a n 6
r
4.08 1.52 0.00
0.88 2.11 0.05
1.22 1.98 0.08
6
s
0.28
0.40
0.40
*/
— a, n, and 8 were determined by least-
squares nonlinear regression of mercury
intrusion porosimetry data and Eq. (6).
— 9 was specified as the total porosity of
tne soil samples after freeze drying.
l.E-5 .0001 .001 .01
Figure 6.
Theoretical relative perme-
ability (K ) curves as influenced
by the n-parameter (S
8/6 ).
playing the influence of n on the effective
porosity when 6r is constant at 0.05, If
similar sets of curves were drawn for 9r =
0.0 and 6r = 0.10, only small differences
from Fig, 1 would be observed; i.e., the
maximum difference in E would be 0,007.
Thus, Fig, 1 can be used to obtain esti-
mates of effective porosity of compacted
clay liners with 6g between 0.3 and 0.5,
between 0.0 and 0.
3.0.
and n between 0.1 and
The method developed in this paper for
estimating the effective porosity of com-
pacted clay liners depends only on parame-
ters describing liner pore structure;
280
-------
TABLE 4
MEASURED PERMEABILITIES AND TIME RANGES OF FIRST CHLORIDE
BREAKTHROUGH AND PREDICTED EFFECTIVE POROSITIES AND NON-
INTERACTING POLLUTANT BREAKTHROUGH TIMES FOR THE COMPACTED
SOIL MATERIALS.
Soil
Material
Nicollet
Fayette
Clarinda
Permeability
(m3/m2/s)
5 x ID'10
2 x 10"8
1 x 10~9
Measured Cl
breakthrough
time (min)
208
between 15
and 30 min
between 250
and 380 min
Effective
porosity
0.04
0.08
0.08
Predicted
breakthrough
time (min)
227
32
272
i.e., QS, 6rj and n-parameter. Cumulative
porosity measurements provide the basis for
determination of the appropriate parame-
ters. By using the pore structure parame-
ters presented in Table 3, effective poros-
ities for the Nicollet, Fayette, and Cla-
rinda soil materials are estimated from
Fig, 1 as 0.04, 0.08, and 0.08, respective-
ly. Once effective porosity is estimated,
pore fluid velocity and noninteracting pol-
lutant travel time can be calculated for a
compacted liner with a known permeability,
hydraulic gradient, and thickness.
Table 4 presents measured values of
permeability and times of first chloride
breakthrough, estimated effective poros-
ities, and predicted times of first break-
through of noninteracting pollutants for
compacted Nicollet, Fayette, and Clarinda
samples, Nicollet and Clarinda samples
both had permeabilities equal to or less
than the preferred value of 1 x 10
m /m /s. The breakthrough measurements
were reported for time ranges because ef-
fluent volume increments were collected for
analysis of tracer concentrations. The es-
timated effective porosities were between
15 and 20% of the total porosities of the
compacted samples. The predicted noninter-
acting pollutant breakthrough times agreed
quite well with the measured chloride
breakthrough times. Predicted breakthrough
times for the Nicollet and Fayette samples
were slightly later than the measured
breakthrough time ranges. The predicted
breakthrough time for the Clarinda sample
was within the measured breakthrough time
range. Because the clay fractions of the
compacted samples had a net negative charge
and chloride is an anion, it is reasonable
to assume that chloride would breakthrough
the samples sooner than a noninteracting
pollutant (i.e., tritiated water).
Future work will include breakthrough
measurements using tritiated water as the
tracer. The thickness of the compacted
samples will be increased for further
breakthrough studies. The pore size dis-
tributions of other compacted materials
suitable for landfill liners will be fur-
ther investigated. Perhaps evidence can be
acquired to show that the n-parameter of
compacted liner materials has a narrow
range (i.e., 1.5 to 2.5) and, thus, indi-
cate that the effective porosities repre-
sent a narrow percentage range of the liner
total porosities (i.e., 15 to 25%). Such
findings will enable easy, yet reasonable,
estimates of the active lifetimes of hazar-
dous-waste-disposal units,
EXAMPLE
The theory presented provides the
basis for estimating effective porosity of
compacted clay liners. To predict the tra-
vel times of noninteracting substances
through the clay liner, the hydraulic gra-
dient, liner permeability, liner thickness,
and effective porosity are required. As an
example of how the theory can be used, we
will make such a prediction. The informa-
tion that we assume is: liner thickness = 1
m, hydraulic gradient = 1,33, liner perme-
ability = 1 x 10~9m3/m2/s.
If we follow only the Darcy equation.
Eg, (1), fluid flux density is 1.33 x 10~*
nr/m2/s. However, the fluid flux density
does not represent the fluid velocity
within the liner. We know that a portion
of the total porosity is responsible for
conducting fluid at velocity higher than
281
-------
the average and that the remaining portion
of the porosity conducts fluids at veloci-
ties lower than the average value. What is
important in pollutant transport is the
portion of porosity that conducts fluid at
higher velocities. To determine the ave-
rage effective fluid velocity, we must
first estimate the effective porosity.
Effective porosity is determined on
the basis of theory previously presented.
By using the mercury porosimetry data for
the Clarinda materials presented in Fig, 3
and Eq. (7), the n-parameter is determined
as 2.0, 9r as 0,08, and 6 as 0.40. The
effective porosity then, is estimated to be
0.08 by using Fig. 1. From Eq. (3) the
travel time for noninteracting substances
through a Clarinda clay liner in this exam-
ple is predicted to be 2 years. Therefore,
under these conditions, we predict that af-
ter 2 years, noninteracting pollutants will
reach the bottom of the clay liner.
Although the concentration of the noninter-
acting pollutant may be low because of
lateral mixing by diffusion, the initial
fraction is predicted to appear after 2
years. In this example, if the product of
liner permeability and hydraulic gradient
(i.e., fluid flux density, J) is mistakenly
used as the fluid velocity, the pollutant
travel time is calculated to be 23.8 years.
If the fluid velocity is mistakenly assumed
to be equal in all soil pores, the nonin-
teracting pollutant travel time is calcula-
ted as 9.5 years. Thus, travel time for
noninteracting pollutants is significantly
overestimated if effective porosity is not
taken into account. The active lifetime of
hazardous-waste-disposal units may be over-
estimated if liner effective porosity is
not properly estimated and used in the pre-
diction of pore-fluid velocity,
ACKNOWLEDGMENTS
The information in this report has re-
sulted from research funded in part by the
United States Environmental Protection
Agency, under Cooperative Agreement CR-
811093, to Iowa State University, The
USEPA Project Officer for this study is Dr.
Walter E. Grube, Jr. We thank Dr. Grube
for providing helpful comments during the
preparation of this paper. We thank J.
Barmettler for writing computer codes to
make our calculations and figures.
This report is Journal Paper No. J-
11707 of the Iowa Agriculture and Home
Economics Experiment Station, Ames, Iowa.
Project No. 2556.
REFERENCES
1, Horton, R., 1985, Method to estimate
effective porosity and pollutant break-
through time for compacted soil mate-
rial. Submitted to Soil Science
Society of American Journal.
2, Mualem, Y., 1976, A new model for
predicting the hydraulic conductivity
of unsaturated porous media. Water Re-
sources Research, 12, pp513-522,
3, Olson, R, E. and D. E. Daniel,
1981. Measurement of the hydraulic
conductivity of fine-grained soils.
In: T. F. Zimmie and C, 0. Riggs
(eds), Permeability and Groundwater
Contaminant Transport, ASTM STP 746,
pp!8-64.
4. USEPA, Federal Register, Monday
December 18, 1978. pp58946-59028.
5. USEPA, Federal Register, Monday July
26, 1982. pp32274-32388.
6. Van Genuchten, M. Th., 1980. A
closed form equation for predicting the
hydraulic conductivity of unsaturated
soils. Soil Science Society of America
Journal, 44, pp892-898,
7. Zimmie, T. F, and L. J, Almaleh,
1976. Soil Specimen Preparation for
Laboratory Testing. ASTM STP 599,
pp202-215.
282
-------
WASTE-LINER COMPATIBILITY STUDIES
Warren J. Lyman, Arthur D. Schwope,
Joseph P. Tratnyek and Peter P. Costas
Arthur D. Little, Inc.
Cambridge, MA 02140
ABSTRACT
Three different studies were undertaken focusing on the compatibility of synthetic
liners (principally flexible membrane liners [FML]) with hazardous wastes. Each of the
three studies is described herein under its final report title. In Study I, qualitative
and quantitative information on the chemical resistance of FML materials was collected from
vendor and technical publications. This information was compiled into a computer data base
comprising about 3,000 data fields on 23 liner materials and over 500 chemicals. Criteria
for assessing this information on a common basis were developed, with separate criteria for
olefinic and non-olefinic liner materials. On the basis of these criteria, normalized
ratings of chemical resistance were developed for each chemical/ material pair for which
there were data. The Study report contains these ratings (over 1,300 of them) as well as a
listing of the raw data from which they were derived.
In Study II, a search was made for laboratory test methods that were, or could be,
used to ascertain the compatibility of flexible membrane liners with hazardous wastes.
Twenty-seven identified methods, from the U.S. and several foreign countries, were examined
and compared. Each pertinent test is described in the Study report. The Study report also
discusses the philosophical basis of these tests and the extent to which they can
realistically be expected to reflect short- and long-term liner performance.
In Study III, a research effort was undertaken to evaluate routes to chemical/liner
compatibility estimation. Proven estimation techniques might be used, for example, in
prescreening FML materials to be tested for compatibility with a waste, or estimating the
resistance of an existing FML to a new chemical in the waste. This study showed that one
approach using solubility parameters can provide useful information in many situations.
Solubility parameters were shown to be useful in three different ways, in solubility
parameter maps, statistical correlations, and with the use of the exchange energy density
parameter.
STUDY I - COMPATIBILITY OF FLEXIBLE
MEMBRANE LINERS WITH CHEMICALS AND WASTES
INTRODUCTION
At present there are four primary and
approximately ten secondary base polymeric
materials from which flexible membrane
liner (FML) sheeting is made. In lined
RCRA facilities, the FML liners may be
exposed to a variety of chemicals some of
which will be known in advance, and some
which will not have been identified or may
be added later. The process of selecting
the most appropriate liner material (for a
new facility) or of checking to see if a
modified waste stream (containing new
chemicals, for example) will be compatible
with the in-place liner can be a
283
-------
time-consuming and expensive undertaking.
Selection of a liner material for a new
RCRA site, including confirmatory
compatibility tests, could take, for
example, over six months and cost several
thousand dollars.
Over the last several years chemical
compatibility tests have been conducted by
the liner industry, by those organizations
which are contemplating installation of a
waste impoundment facility, and by various
government agencies. Although these
groups have conducted a significant amount
of testing, the use and/or comparison of
their results has been difficult because
there have been no standard procedures for
testing or reporting the findings.
Furthermore, the findings are scattered
throughout the technical literature,
government reports, vendors' brochures,
and proprietary and unpublished
literature.
In spite of the above limitations,
there is a clear need to compile the
available data so that maximum use can be
made of past work. A compiled and
evaluated data base on waste-FML
compatibility may be of use if preliminary
screening for FML selection and in
determining compatibility for in-place
FMLs to be exposed to new waste
compositions. Such a data base would also
become a valuable tool in evaluating
different test protocols that have been
used, in identifying major data gaps, and
in the development of FML-chemical
compatibility prediction schemes.
PURPOSE
The purpose of this study was to
gather data on the compatibility of FML
and various chemicals and waste liquids,
and to analyze and report these data on a
common basis. The intent of this
compilation of chemical compatibility
information is to assist liner material
vendors and purchasers, and those involved
in reviewing permit applications for lined
facilities, in ensuring that - at least
from a chemical compatibility standpoint -
the most appropriate material is installed
at a waste site.
APPROACH
The study was composed of three
principal components:
(1) Collection and compilation of
liner-waste compatibility
information;
(2) Evaluation of information; and
(3) Development of a liner-waste
compatibility matrix.
The information was collected from
the open literature and from brochures and
data sheets obtained from the liner
industry. Additional data were sought
from European investigators, but the yield
was minimal. While the focus was on
liners fabricated from synthetic polymers
(FMLs), asphalt liners were also
addressed. Both qualitative and
quantitative compatibility data were
collected.
The information was compiled using
commercially available database software.
The format allowed easy review of all
available data on a given liner-chemical
or liner-waste pair at once and the
formulation of an overall judgment as to
the compatibility of the liner material
with the waste. The results were
summarized in the form of a compatibility
matrix, organized such that all chemicals
within a given family of chemicals (e.g.,
aldehydes, ketones, acids) are reported
together. This may facilitate
extrapolation of the resistance
information to liner-waste combinations
for which no data are available.
Finally, all the information (test
data and qualitative ratings) were
tabulated and included as an appendix to
the main report (Schwope et_ al_., 1984).
This will allow others (who may wish to
independently evaluate compatibility)
access to the original data.
RESULTS
The qualitative data, usually
provided by FML vendors as FML-chemical
compatibility charts, typically assign one
of two or three ratings to each
FML-chemical pair (for specified
conditions, eg., temperature). Examples
of compatibility ratings include the
following:
Company A
S = Laboratory data or field experience
indicate satisfactory service likely
U = Not recommended, unsatisfactory
service likely
284
-------
Company B
Company D
0 = No effect
M = Moderate effect
S = Severe effect
E = Less than
10% change
in volume,
ultimate
elongation
or tensile
strength
Company C G = Less than
30% change
S = Satisfactory in volume
L = Limited application P = More than
possible 30% change
L) = Unsatisfactory in volume
These ratings were all converted to a
normalized basis using a rating ranging
from "a" to "d" where "a" is the most
resistant and "d" is the least resistant.
The combinations "ab" and "be" were used
for a few cases where the vendor's
description of the degree of compatibility
appeared to be intermediate between our
classification ratings. Ratings of "a",
"ab" and "b" are taken as indicative of
acceptable resistance or compatibility;
ratings of "be", "c" and "d" are taken as
indicative of unacceptable resistance.
Table 1 provides, for example, the
normalized ratings from three high density
polyethylene (HDPE) vendors for seven
chemicals. Note that there is essentially
complete agreement among the vendors in
this case. One would expect, perhaps,
significant variability for a polymeric
product that for specifc purposes, is
produced with varying amounts of fillers
and other compounding agents.
The evaluation and use of the
quantitative data available were
undertaken with the thought that some
overall performance standard might be
defined in terms of chemical permeability.
There are, unfortunately, insufficient
data relating FML physical property
measurements to chemical permeability;
thus, it was necessary to additionally
consider what might be acceptable changes
in some physical properties. Table 2
provides a listing of the chemical
resistance criteria finally selected.
(Note that HDPE has been assigned a
separate criteria).
The criteria developed for assessing
the chemical resistance of liner materials
are based on the technical judgment of the
investigators and are in general
concurrence with the criteria that have
been suggested by other investigators. A
discussion of references may be found in
the Study I report (Schwope £t aK_, 1984).
The criteria, however, have neither been
proposed to, nor adopted as standards by,
the EPA or any independent organization.
They are believed to represent a good
starting point for developing chemical
compatibility criteria for liner
materials. The values in Table 2 are not
a guarantee nor are they intended to be
used as a rule by which liners are either
accepted or rejected for any given
application. They are a set of values
intended to provide a perspective from
which information can be analyzed.
Readers may also use the raw data provided
in an appendix to the final report
(Schwope et aJL, 1984) to derive their own
conclusions.
Based on the normalized ratings
reflecting qualitative data and the
criteria for the evaluation of
quantitative data (Table 2), an overall
rating was assigned to each chemical-liner
(or waste-liner) pair for which data were
available. The ratings use the "+" and
"-" symbols; one (either + or -)is
assigned to represent compatibility
conclusions derived from quantitative test
data, and a second (also either + or -) is
assigned to represent compatibility
conclusions derived from qualitative
information. A vertical bar, "|", is used
to separate the two ratings for each pair.
When data gaps are considered, the total
number of ratings possible is eight (Table
3).
In all, the project report (Schwope
et^ al., 1984) provides ratings for
approximately 550 chemicals (or wastes)
and 23 FML materials. In addition special
discussions are given on FML-water
interactions, environmental stress
cracking, and the effect of chemical
concentration on permeation rates through
FMLs.
285
-------
TABLE 1. COMPARISON OF THE CHEMICAL RESISTANCE INFORMATION
FROM THREE HOPE VENDORS
Chemical
Carbon tetrachloride
Glycerol
Ethyl ether
An i 1 i ne
Ethanol
Sulfuric Acid, <30%
Nitric Acid, <30%
Normalized
Vendor X
be
a
be
a
a
a
a
Rating (at 20-25°C)*
Vendor Y
be
no data
be
a
a
a
a
Vendor Z
be
a
be
a
a
a
a
*Ratings range from "a" most compatible to "d" least compatible; see
text for discussion.
TABLE 2. CRITERIA USED TO DEFINE CHEMICAL RESISTANCE IN THIS STUDY
Property (units)
Permeation rate (g/m2d)
Change in weight (%)
Change in dimensions (%)
Change in tensile strength(%)
Change in % elongation
at break (%)
Change in 100% or 200%
modulus (%)
Change in hardness (points)
Resistant (+) Not Resistant (-)
HOPE Other FMLs HOPE Other FMLs
<8.6 <8.6 >8.6 >8.6
< 3 <10 > 3 > 10
si <10 >1 >10
<20 <20 >20 >20
<20 <30 >20 >30
<30 -- >30
,10 - >10
*The "+" and "-" symbols are the symbols used in the final compatibility
matrix.
286
-------
TABLE 3. RESISTANCE RATINGS USED FOR FML COMPATIBILITY DATA
Rating
Test Data
Qualitative Information
+|+ Good resistance to and
chemical
+| Good resistance to and
chemical
|+ No test data but
+|- Good resistance to but
chemical
-j+ Poor resistance to but
chemical
-j Poor resistance to but
chemical
|- No test data but
-|- Poor resistance to and
chemical
Favorable ratings
No qualitative ratings
Favorable ratings
Unfavorable ratings
Favorable ratings
No qualitative ratings
Unfavorable ratings
Unfavorable ratings
STUDY II - TEST METHODS FOR DETERMINING
THE CHEMICAL WASTE COMPATIBILITY OF
SYNTHETIC LINERS
INTRODUCTION
It is well known that no one
synthetic liner is resistant to all
chemicals, that most facilities receive a
variety of wastes, and that the nature of
these wastes is likely to change with
time. In some cases, the exact
composition of the waste for a proposed
treatment or disposal facility may not be
known. However, for chemicals of known
composition, there is a need to know what
tests can be conducted to: (1) select a
chemically-resistant liner, and (2)
demonstrate (to the public and permit-
granting authorities) that long-term
resistance and compatibility are likely.
A variety of waste-liner
compatibility tests have been described in
the project report for Study II (Tratnyek
et a!., 1984). Some have been proposed by
regulatory agencies, while others have
been put forth by national
standards-setting organizations,
individual companies in the liner
technical community (resin manufacturers,
sheet manufacturers), research
organizations and foreign countries.
Among this group of available test
protocols there is not a universally
accepted compatibility test method.
Although details of the available
tests vary, all are laboratory tests in
which selected physical properties of the
FML are compared and evaluated after
contact with the chemical or waste for a
specified period of time and under
specified conditions (e.g., temperature).
All tests are tedious, time-consuming, and
expensive. Useful data for product
specification and application are derived
from these tests, but none adequately
addresses all issues and questions raised,
including chemical permeation of the liner
and liner life-time prediction.
For the benefit of regulators,
planners, and designers, the current test
methodology requires review and assessment
so that appropriate tests can be selected,
287
-------
the value of data from any particular test
assessed, and improvements made.
PURPOSE
The purpose of this Study was to
compile, review, and evaluate available
test methods for measuring and/or
estimating the chemical compatibility of
liquid wastes with polymeric flexible
membrane liners (FMLs) to be used at
hazardous waste treatment, storage and
disposal facilities.
The intention was not write a test
method for FML/waste-compatibility
testing, but to assess the current state
of affairs and direct the thinking of all
those concerned towards routes that might
lead to a satisfactory, comprehensive, and
reliable chemical compatibility
measurement technique.
APPROACH
In carrying out this task, the focus
has been placed on identifying and
describing not only tests designed to
assess FMLs, but also tests from related
technologies (e.g., for pipe, packaging,
and film) that are used to evaluate the
compatibility of chemicals with plastics
and other organic polymers. All types of
tests that were relevant to chemical
resistance and the transfer of liquids
through membranes were considered and
reviewed.
Levels of information sources are
categorized as follows:
I. International standards and
tests;
II. National standards and tests;
III. Industrial standards and tests;
IV. Project tests; and
V. Academic/literature tests.
The organizations, along with the
identification code used in the text, for
which tests are reviewed are listed below:
International Organization for
Standardization (ISO)
Geneva, Switzerland
Deutsches Institut fur Normung (DIN)
Berlin, Germany
British Standards Institution (BSI)
London, England
U.S. Environmental Protection Agency
(EPA) Washington, DC
National Sanitation Foundation (NSF)
Ann Arbor, MI
American Society for Testing
Materials (ASTM)
Philadelphia, PA
U.S. Military Standards (MIL)
Information from these organizations
and sources was obtained by a combination
of personal contacts and a literature
search using computerized bibliographic
data bases.
As part of the need to elucidate and
understand compatibility test methods and
FML requirements, a meeting with experts
in FML technology and applications was
held at Arthur D. Little, Inc. (Cambridge,
MA), in January 1984. The meeting was
attended by three FML manufacturers'
representatives, an independent FML
researcher, EPA representatives (Office of
Solid Waste and Office of Research and
Development), and members of the Arthur D.
Little, Inc. staff. This meeting provided
an opportunity to discuss the current test
methodology and recommendations for
improvements.
RESULTS
Information was obtained on 27
compatibility tests (Table 4). The
project report (Tratnyek jit_ al., 1984)
provides a written description and a
flowchart of each test. The study was
intended to document the parameters of the
chemical challenge to the FML and the
subsequent measurements and observations
made. The range of possibilities involved
in such compatibility tests is shown in
Figure 1.
Test methods are more or less similar
and the basic approach appears to be
patterned after previously developed test
methods used for rubber and plastic sheet.
However, a caveat is often given, warning
that the test may not correspond with the
service condition, and that the data are
only comparative in nature. Absolute
values or criteria for compatibility are
not provided by the tests.Thus,
interpretation of test results remains a
very subjective undertaking based upon
experience and an understanding of FML
properties.
The tests reviewed generally
exhibited the following features:
288
-------
TABLE 4.
LISTING OF COMPATABILITY TESTS REVIEWED
I. INTERNATIONAL STANDARDS AND TESTS
ISO 175 Plastics -- Effect of Liquid
Chemicals Including Water
ISO 1817 Vulcanized Rubbers -- Resistance
to Liquids
DIN 53 521 Rubber -- Resistance to
Liquids, Vapors and Gases
DIN 53 532 Permeability of Elastomers to
Liquid Fuel
BS 4618 Chemical Resistance of Plastics
to Liquids
BS 5173 Hoses -- Chemical Resistance
II. NATIONAL STANDARDS AND TESTS
EPA Method 9090
NSF Standard 54
ASTM D543 Resistance of Plastics to
Chemcial Agents
ASTM D814 Rubber -- Vapor Transmission
ASTM G20 Chemical Resistance of Pipeline
Coatings
ASTM D471 Rubber — Effect of Liquids
MIL-T-6396D Aircraft Tanks
III. INDUSTRIAL STANDARDS AND TESTS
Schlegel Lining Technology, Inc.
J.P. Stevens and Company, Inc.
Gundle Lining Systems, Inc.
IV. PROJECT TESTS
Simulation Test (Haxo)
Pouch Test (Haxo)
Tub Test (Haxo)
Immersion Test (Haxo)
NSF FML Project Short- and Long-term
Screening Tests
Harwell, England -- Accelerated Tests for
HOPE
V. SELECTED ACADEMIC/LITERATURE TESTS
Sequential chemical absorption (Carpenter
and Fisher)
Fluid Resistance of rubbers and
Elastomers (Minter and Meier)
Stress Cracking by Creep Rupture Test
(Modern Plastics Encyclopedia)
Chemical Stress Relaxation (Okuda)
Proposed — Environmental Stress Cracking
of Ethylene Plastics (Crissman)
• Immersion or one-sided exposure
of FML test specimen to waste;
• Exposure of test specimen
without stress, except for the
polyethylene-type plastics
(which are a special case);
• Exposure at ambient temperature
and some elevated temperature
(the latter to provide an
accelerated test);
• Exposure of varying time (short
to long), or to point of no
change in mechanical properties;
t Test in laboratory;
• Exposure to reagents or waste;
• Measurement of mechanical
properties and evaluation of
appearances as indicators of
compatibility.
At the present time, two specific
laboratory test methods for evaluating FML
performance in a chemical environment have
evolved: NSF Standard 54 (NSF, 1983) and
EPA's Method 9090 (EPA, 1984). The latter
was designed specifically as a
compatibility test for membrane liners in
the presence of hazardous wastes. It has
been proposed by EPA as the standard test
to be used for all FML-waste compatibility
data submitted in support of RCRA permit
applications for lined facilities. The
NSF method is a voluntary approach to
industry standardization. Both tests rely
on measurement of similar physical
property changes (e.g., weight,
dimensions, tensile strength, tear,
elongation), and both employ similar test
temperatures (23°C and 50°C), but neither
specifies criteria for failure.
No one test satisfies all needs for
FML-waste compatibility testing at RCRA
sites. Indeed, there appears to be a need
for at least three basic test protocols
that could be used at different stages in
a facility's life:
Test 1
I. Short-term tests
(up to 30 days'
exposure in
laboratory)
II. Intermediate-term
tests (up to 4
months exposure
in laboratory)
Purpose
Preliminary
Screening of
liner materials
Confirmatory tests
for selected
liner; data to
be submitted with
permit application
289
-------
FML MATE RIAL
Thermoplastic
Elastomer
Rubber
Reinforced vs. Unreinforced
MEASUREMENTS
& OBSERVATIONS. .
APPEARANCE
Color
Surface Character
Mass Features
Wear
DIMENSIONAL
Density/Weight
Linear Dimensions
Volume
STATIC PHYSICAL
Hardness
Tensile Stress/Strain
Compression Stress/Strain
Shear Stress/Strain
Flexural Stress/Strain
Tear
Puncture
DYNAMIC PHYSICAL
Resiliency
Creep
Stress Relaxation
Set
Fatigue
Stress Cracking
MASS TRANSFER
Permeability
Diffusion
I earhina
CHEMICAL CHANGE
Molecular Weight
Viscosity
Chemical Analysis
Spectral Analysis
MATE RIALS CONTACT
Field Coupon
Laboratory Specimen
Immersion
Contact One Side
Jar, Fixture, or Dish
Strained vs. Unstrained
INFORMATION
PROCESSING
& EVALUATION
COMPATIBILITY
TEST
DATA
LIQUID CHALLENGE
Waste Liquid
Reagent
Water
Circulated vs. Stagnant
PARAMETERS. .
TEMPERATURE
Ambient
Low
High
Cyclic
TIME
Hours
Days
Weeks
Years
CONCENTRATION
Absolute
Concentrated
Dilute
FML CONSTRUCTION
Thick-Thin
Reinforced
Unreinforced
Laminated
ENVIRONMENTAL
Light
Ozone (oxygen)
PH
Wet/Dry Cycling
Freeze-Thaw Cycling
SAMPLING & STATISTICS
Number of Samples
Sample Size
Form
F igure 1. The Com patibil ity Test Scheme
290
-------
III. Long-term tests Monitoring actual
(for life of durability of
facility) at liner after
actual site; use exposure to
removable coupons environment and
real wastes.
In addition to the development of
such protocols as described above,
research is required in other areas.
There is as yet no reliable way to predict
long-term performance of a FML from short-
term and/or intermediate term tests (up to
4 months). There are no generally
accepted rules for evaluating the test
data that are generated. And tests, in
general, must focus more on intended
product performance and less on product
specifications normally associated with
the purchase of plastic sheeting. The
whole scope of compatibility testing, and
its relationship to performance criteria
and their development, and to FML service
lifetime predictions are subjects that
require rethinking. New tests such as
stress-relaxation and dynamic mechanical
analysis may help in some cases.
III. STUDY III - CHEMICAL RESISTANCE
RATING ESTIMATION: SOLUBILITY PARAMETER
APPROACHES'
INTRODUCTION
The availability of FML-waste
compatibility tables summarizing data from
tests of materials immersed in the
chemicals/mixtures of concern or from
permeation testing would be useful in
selecting suitable liner materials for a
given waste. Unfortunately, the amount of
data available is extremely small relative
to the large variety of possible exposure
conditions. Furthermore, testing is not
always practical within the time and
budget constraints of many projects.
There is a need, therefore, to develop
analytical techniques that will predict
the chemical compatibility (i.e.,
resistance) of polymeric materials upon
their exposure to chemicals and chemical
mixtures.
APPROACH
In the initial stages of Study III
(Schwope e£ ^1_., 1985) a literature search
was undertaken to identify areas of proven
success, and areas of needed research, in
predicting chemical compatibility with
polymers. As a result of the literature
search, a decision was made to focus the
remainder of the program on several
predictive techniques based on solubility
parameter theory, and statistical analysis
and correlation of chemical properties
with chemical resistance.
All of the studies and analyses
carried out were based upon literature
data on solubility parameters and upon
manufacturers' ratings of chemical
resistance. No laboratory studies were
undertaken.
RESULTS
This program was completed in early
1985, and the draft final report (Schwope
e^ _§]_., 1985) is currently being reviewed
by the EPA. A summary of some of the more
important findings is presented below.
Solubility Parameter Theory
Solubility parameter theory has been
used as a very effective tool for
predicting chemical compatibility in
solute-solvent systems. According to
Hildebrand and Scott (1950), the overall
solubility parameter of a chemical, 6, is
its cohesive energy density:
6 = [(AHvap - RT)/V]*
(1)
where AHvap is the molar enthalpy of
vaporization, R is the universal gas
constant, T is temperature, and V is
liquid molar volume. 6 has units of
(pressure) . According to the theory,
substances with similar solubility
parameters (<$/Ax - 6/B\ < 3) will enter
readily into ^solution 'along the "likes
dissolve likes" axiom, while materials
with strongly differing solubility
parameters will be mutually insoluble.
As an expansion on this idea, Hansen
(1967) developed the three-component
solubility parameter. These three
components take into account the major
interactive forces of a chemical:
dispersive forces (D), polar forces (P),
and hydrogen bonding forces (H).
According to Hansen (1967):
2
6 P+
(2)
291
-------
For many applications, the
three-component approach has been shown to
be a better predictor of solute-solvent
behavior than the single-parameter
approach, and it has been widely
investigated and adopted.
Solubility Parameter Maps
One well-established graphical
approach for examining chemical
compatibility is the Solubility Parameter
Map. These "maps," which can be drawn in
two or three dimensions, allow estimation
of chemical compatibility visually, based
only on the three-component Hansen
parameter of a given chemical. They can
be created for virtually any resistant
material for which enough compatibility
data exist in order to confidently
establish regions of common chemical
resistance.
Figure 2 provides an example of a
three-dimensional solubility parameter map
for plasticized polyvinyl chloride. The
coordinates for each chemical were
calculated by the normalization of each of
the three solubility parameters components
to a scale of 0-100, i.e.:
x 100 (3)
p Recommended
V Not Recommended
O Recommended (cond.)
& Not Recommended (cond.)
Poor Resistance
Region
Good Resistance
Region
fd-0
Triangular Representation
Figure 2. Three-Dimensional Solubility Parameter Map for PVC
292
-------
Chemicals with different degrees of
compatibility, ranging from "excellent" to
"not recommended", are represented by
different symbols on the map. (Note that
the location of the point for any chemical
on such a map is independent of the
polymeric material; only the compatibility
rating may change.) In the case of Figure
2, two clear regions of different
compatibility show up, one of poor
resistance and one of good resistance;
there is also a large area for which no
information is available. To the extent
that reasonably clear areas of good and
poor resistance can be identified on such
maps, our ability to predict resistance for
new untested chemicals is greatly enhanced.
In a roughly similar manner,
two-dimensional solubility parameter maps
may be created by plots, for example, of 6
vs. 6H and 6 vs 6,,. Again chemicals with
different decrees of compatibility are
represented on the plot (cartesian
coordinates this time, not triangular) with
different symbols and the degree of
clustering evaluated.
Statistical Correlation
A statistical approach was undertaken
in an attempt to correlate solubility
parameters and other fundamental chemical
properties with material vendors' rating of
chemical resistance. Three resistant
materials were examined: CPE, Neoprene, and
Butyl Rubber. Using the vendors' ratings,
established from swelling data or similar
tests, we placed each chemical into one or
two categories; "Compatible" or "Not
Compatible" with each of the liner
materials. Using a computerized statistics
package, the categories were then
correlated with the three components of the
Hildebrand solubility parameter, as well as
such fundamental chemical properties as
molecular weight, molar volume, dipole
moment, boiling point, and molecular
connectivity.
The statistics program employs a
fundamental linear routine which
established coefficients for each of the
parameters involved. These coefficients
are then used to establish the probability
of a particular chemical falling into each
of the two rating groups. Based on these
probabilities, each chemical is assigned a
chemical resistance rating, and this new,
correlation-based rating is compared with
the original manufacturer's rating, and the
accuracy of the correlation assessed.
In one test of CPE resistance, for
example, a data set of 123 chemicals (86
considered "Compatible" and 37 "Not
Compatible") was evaluated using the three
components of the solubility parameters
(60, 6p and 6,,). The resulting correlation
is described By equations 4, 5, and 6:
-86.88 + 10.85
- 0.215 6p + 0.072 «H (4)
-109.69 + 12.88 6n - 0.291 6p - 0.089 6H (5)
Pl = exp(-.5 D^)/[£ exp (-.5 D/)] (6)
where D, and D~ are discriminant values for
chemicals in the "Compatible" and "Not
Compatible" classes, respectively, and P.,
is the probability (0 < P, < 1) of a
chemical being "Compatible. In a test of
this correlation, chemicals with Pn > 0.5
were considered "Compatible." In1 this
manner, the correlation was able to
correctly classify approximately 80% of the
chemicals in the test set to the correct
compatibility class. When molar volume and
liquid density were added as correlation
variables, no appreciable improvement in
correlation accuracy was observed.
Exchange Energy Density Parameter
The exchange energy density
parameters, A , is a second non-graphical
method for assessing chemica.1 compatibility
of resistant materials. A is a measure
of the change in energy density associated
with the process of mixing component "i"
with component "j," and it is calculated
with three-component solubility parameters.
The exchange energy density parameter
simply combines the difference between the
solubility parameter values of the
resistant material (polymer) and the
solubility parameter values of the solvent
for each of the three Hansen components:
A1
-6]
j2
-6HJ]2 (7)
where in this case the superscripts "i" and
"j" represent the polymer and the chemical,
respectively. Values of A ^ close to zero
are indicative of two materials that are
293
-------
likely to enter into solution ("Not
Compatible" in-the context of this study),
while higher A ^ values are indicative of
chemical differences that are likely to
exhibit themselves by mutual insolubility
("Compatible" in the context of this
write-up).
In using A1J to estimate a polymeric
material's performance for a particular
chemical, the three-component solubility
parameters for both the chemical and the
polymer must be known. Three-component
parameters for several hundred chemical
solvents are available from the literature
or can be estimated by group contribution
techniques. However, solubility parameters
for most polymeric materials are not
readily available. Furthermore, little is
published regarding the effects of
compounding ingredients on the solubility
parameter of commercial polymeric
materials. Thus it often becomes necessary
to estimate the solubility parameter of
polymer materials from swelling test or
other analytical methods. One relatively
easy method of rough estimation involves
the 3-dimensional solubility parameter maps
discussed earlier. From the maps, one can
estimate the point of "least compatibility"
as the center of the "Not Recommended"
region. This point will correspond to
maximum solubility (i.e., least chemical
compatibility), and thus will be indicative
of the solubility parameter components for
the polymeric material. The fractional
parameter values (f.) can be read off the
map, and 6 estimated by using the 6 value
of the chemical plotted closest to the
maximum solubility point. From this
information, 6D, 6p, and 6,,, for the
polymer can be calculated.
This study has evaluated the potential
of the exchange energy density parameter
approach for CPE and HOPE. Figure 3 shows,
for example, the results for CPE using data
for 113 chemicals. In this figure, the
percentage of chemicals rated as either
"Recommended" or "Not Recommended" within
eight different regions of A* values (0-2,
2-4, 4-6, etc.) is shown. The Figure shows
clearly that when a chemical has an A*
value above 8 (calculated for CPE), it is
almost certain to be compatible with the
liner material. Although the results for
HOPE are not as sharply defined as those
for CPE, they do show discernible trends.
ACKNOWLEDGMENTS
We would like to thank both Robert
Hartley (the EPA Task Manager) and Robert
Landreth for their technical guidance and
patience associated with the conduct of
these studies. The work described was
carried out as Task 16 under EPA Contract
No. 68-01-6160, and the assistance of
Michael Slimak as Project Officer is also
acknowledged.
The principal investigator for Study I
and Study III is Arthur Schwope. The
principal investigator for Study II is
Joseph Tratnyek.
REFERENCES
Environmental Protection Agency (EPA),
1984. "Hazardous Waste Management System;
Groundwater Testing and Monitoring
Activities," Federal Register. 49_(191):
38786 (October, 1984).
Hansen, C.M. , 1967. "Three Dimensional
Solubility Parameter - Key to Paint
Component Affinities: I. Solvents,
Plasticizers, Polymers and Resins," J.
Paint Techno!., .39(505): 104-117.
Hildebrand, J.H. and R.L. Scott, 1950.
Solubility of Noelectrolytes, 3rd ed.,
Reinhold, New York, NY.
National Sanitation Foundation (NSF), 1983.
NSF Standard Number 54: Flexible Membrane
Liners. National Sanitation Foundation,
Ann Arbor, MI.
Schwope, A.D., P.P. Costas and W.J. Lyman,
1984. Compatibility of Flexible Membrane
Liners with Chemicals and Wastes. Final
Report on Task 16 of EPA Contract No.
68-01-6160, prepared for U.S. Environmental
Protection Agency, Office of Research and
Development, Cincinnati, OH.
Schwope, A.D., P.P. Costas and W.J. Lyman,
1985. Prediction of Waste/Leachate
Resistance of Flexible Membrane Liners.
Draft Report on Task 16 of EPA Contract No.
68-01-6160, prepared for U.S. Environmental
Protection Agency, Office of Research and
Development, Cincinnati, OH.
Tratnyek, J.P., P.P. Costas and W.J. Lyman,
1984. Test Methods for Determining the
Chemical Waste Compatibility of Synthetic
Liners. Final Report on Task 16 of EPA
294
-------
Contract No. 68-01-6160, prepared for U.S.
Environmental Protection Agency, Office of
Research and Development, Cincinnati, OH.
100
90
80
or 70
LU
o
60
50
o
2 40
i
H
g
30
01
o
£ 20
Q_
10
RECOMMENDED
NOT
RECOMMENDED
0-2 2-4 4-6 6-8
A'/Z RANGES
8-10
10-12
Figure 3. Change in Chemical Compatibility Ratings for 113 Chemicals
with CPE as a Function of the Exchange Energy Density
Parameter, A
295
-------
RECOVERY AND TESTING OF A SYNTHETIC LINER FROM A
WASTE LAGOON AFTER LONG-TERM EXPOSURE
Nancy A. Nelson, Henry E. Haxo, Jr.
Matrecon, Inc.
Oakland, California 94608
Peter McGlew
U.S. EPA - Region 1
Boston, Massachusetts 02203
ABSTRACT
Samples of a 100-mil, high-density polyethylene (HDPE) liner were recovered from
a waste lagoon in the northeastern United States after 4.75 years of service. Samples
exposed to the waste were removed from different depths and tested to determine the
effects of the exposure on the physical properties of the liner. The recovery was
performed during closure of the lagoon in a Superfund Remedial Action. The impounded
waste liquid was predominantly aqueous and contained significant amounts of organics,
particularly chlorinated hydrocarbons, which increased in concentration with depth.
Overall, the liner appeared to be in satisfactory condition. No evidence in-
dicated that it had cracked or failed, but it did show considerable waviness and
distortion on the berm and slopes.
Liner samples were analyzed and tested. Samples of the HDPE liner from the bottom
of the lagoon showed an absorbed waste content of about 2%; they also showed a 10% loss
in tensile strength at yield and similar losses in S-100 and S-200 moduli and 30% loss
in modulus of elasticity. The samples taken from the slopes of the lagoon showed es-
sentially no changes in physical properties. Construction equipment was used in an
attempt to remove the waste without damaging the liner, but the liner on the bottom of
the lagoon was nonetheless destroyed during the cleanup operations.
INTRODUCTION
The recovery of samples from a mem-
brane liner at a waste lagoon in the
northeastern United States provided an
unusual opportunity to investigate a
high-density polyethylene (HDPE) liner
after 4.75 years of service in a rather
severe environment. The lagoon had been
used to impound wastes containing a
variety of organic chemicals. Investi-
gations of the performance of liners
that have been in service are important
to develop a data base on field experi-
ence, to verify laboratory studies on
liner/waste compatibility, and to estab-
the durability of liners under different
service environments.
Samples of the 100-mil HDPE liner
were obtained and tested when the lagoon,
a chemical waste storage and bulking
facility, was decommissioned and removed
as part of a Superfund Remedial Action.
This report describes the recovery
of HDPE liner samples from the lagoon and
presents results of the analyses and test-
ing of the samples. Also included is a
brief description of the history of the
lagoon and information on the wastes that
that were in contact with the liner.
SITE HISTORY
Construction of the lagoon by the
owners began with site preparation in
1978. The 100-mil HDPE liner was made in
Germany and installed by the manufacturer.
Installation began in the fall 1978, but
was delayed until the following spring
because of bad weather. The lagoon was
296
-------
completed in May 1979; it measured 117 by
122 ft and had a capacity of 700,000 gal.
Most of the lagoon was above the natural
grade. The depth ranged from 7 to 10 ft,
sloping down toward the sump in the north-
west corner. The sump consisted of a 2-
ft-diameter vertical HOPE pipe in the
lagoon bottom connected to an 8-in.-dia-
meter pipe that passed under the dike
wall to the outside of the lagoon and
connected to a 3-ft-diameter vertical
HOPE pipe.
According to the decommissioning
justification document, the U.S. Environ-
mental Protection Agency (EPA) became in-
volved in January 1981. The owner of
the facility had declared bankruptcy and
abandoned the site. An emergency was
declared, and clean up work began imme-
diately. The lagoon was sampled and drawn
down, as it was close to overflowing.
Drums of hazardous wastes on the site were
stabilized. In October 1982, a remedial
action master plan was completed for the
site, and a fast track feasibility study
was prepared for the lagoon.
Information from the decommissioning
justification document indicates that some
wastes were dumped into the lagoon before
the HOPE liner was installed. The docu-
ment also indicates that the lagoon over-
flowed during its use.
ANALYSIS OF THE WASTE
The decommissioning justification
document indicated that in July 1983, the
lagoon waste had been sampled from a small
boat. During that sampling, sediment
samples were collected from each of the
four corner areas (see Figure 1). Aqueous
waste samples were collected at the north-
west corner, the center, and the southeast
corner (see Figure 1). Different depths
were sampled at each point.
The sediment that was directly on the
lagoon liner underwent limited testing.
Data on its physical and chemical pro-
perties (Table 1) show the range in sedi-
ment composition at the four sampling
points. The lagoon sump is the deepest
point of the lagoon, and the southwest
corner is the shallowest. The low flash
point and high energy value for the sump
sediment indicate the presence of organic
solvents, which tend to be aggressive to
most lining materials. These samples were
extracted according to the Extraction
Procedure (EP), and the extract was ana-
lyzed for primary drinking water para-
meters.
9-
8-
7-
6-
5-
4-
3-
2-
1 —
AB C D E F GHI
I II II III
7 D6
10' deep -
©Sump
t S ^
: s,w :
! w ;
~- D, a2 -
\ ' S,W :
: 7' deep ;
~~ s i 1 1 1 1 1 1 1 1 1 1 1 1 r
"" ' "" ' °5
k
20 feet
Figure 1. Lagoon lay-out showing grid
pattern used in lagoon sampl-
ing, points of sample col-
lection, and location of sump
((o). Liner samples are indi-
cated by n with the liner
sample number, sediment
samples are indicated by S,
and aqueous waste samples are
indicated by W. The toe of
the slope is indicated by a
hashed line (+++++).
The aqueous waste samples were
analyzed for more parameters. Tests were
run for inorganic and organic priority
pollutants, priority pollutant pesticides,
PCB's, and other inorganics. Samples were
taken from three or four depths at each of
the three collection points. Concentra-
tions of most constituents increased with
depth, showing that the lagoon waste had
become stratified. Data are available for
all samples. Those from the greatest
depth at each collection point are of
greatest interest in terms of the liner,
297
-------
TABLE 1. PHYSICAL AND CHEMICAL PROPERTIES OF SEDIMENT SAMPLES
COLLECTED FROM THE LAGOON SITE ON JULY 25-26, 1983a
Point of sampling
Parameter
Lagoon sump
B-2
G-7
G-3
pH 5.4
Flash Point, °C 26
Corrosivity
Ignitability
Energy valueb, BTU/lb
Total residue, % 19.84
8.6
>60
Noncorrosive Noncorrosive
Ignitable Not ignitable
7880
56.83
10.3 10.3
>60 >60
Noncorrosive Noncorrosive
Not ignitable Not ignitable
7.89
13.19
aDate taken for the decommissioning justification document,
^Dry-weight basis.
since they represent the waste closest to
the liner on the lagoon bottom. These
samples are the most concentrated and
hence the most aggressive. Because in-
organic chemicals are not generally ag-
gressive to polymeric membrane liners, the
high concentrations (greater than 50 ppm)
of copper, nickel, zinc, and iron should
not affect the HDPE. On the other hand,
phenols and petroleum hydrocarbons (71 and
2100 ppm, respectively, in the sample from
the lagoon sump) could have some impact
on the liner. The total organic carbon
(TOC), which was greater than 6000 ppm in
all three of the deepest samples, could
also affect the liner. The analyses of
organic priority pollutants indicated the
presence of several aggressive chemicals
at concentrations greater than 50 ppm in
the sample from the sump (Table 2). Chlo-
rinated solvents are known to cause swell-
ing and deterioration of physical proper-
ties in HDPE liners. The analyses of
aqueous waste samples from the greatest
depth at the three sampling locations re-
sulted in the detection of only minute
amounts of the PCB Aroclor 1254 in two of
the samples.
TABLE 2. COMPONENTS OF THE AQUEOUS WASTE AND CHARACTERISTICS3
POTENTIALLY AGGRESSIVE TO THE LINER
Sampling point and depthb
G-3 E-5 Lagoon sump
Waste component or characteristic
Organic priority pollutant, mg/L:
Chloroform
Ethyl benzene
Methyl ene chloride
Tetrachloroethane
Trichloroethane
Toluene
1,1 ,1-Trichloroethane
1,2-Dichloro benzene
3.0 ft
2.65
3.30
102
1.7
28.6
10.4
0.31
0.96
4.0 ft
3.2
9.6
140
3.8
90.2
24.5
• • *
1.89
6.5 ft
56.6
1080
325
182
9100
522
50.0
151
Other:
Phenols, mg/L
Total organic carbon,
Petroleum hydrocarbons
PH
Flash point, °C
mg/L
, mg/L
20
6950
26.0
9.7
>60
32
8230
27.8
9.6
>60
71
17200
2100
9.8
>60
aData taken from the lagoon decommissioning document.
^Samples were taken from the greatest depths at three locations in
the lagoon (see Figure 1).
298
-------
LAGOON DISMANTLING
Remedial action on the lagoon began
in mid-December 1983. The liquid con-
tents of the lagoon were pumped into
tanker trucks for treatment and trans-
port to an approved disposal facility.
Approximately one month was needed to
remove the 100 truckloads of waste.
To solidify the remaining sludge
for removal, ground corn husk was dumped
at one corner of the lagoon and blended
into the sludge with a wheeled bobcat
(Figure 2). The wheeled vehicle and a
toothless bucket were used initially to
minimize damage to the lagoon liner.
However, when it became obvious that the
liner was being torn regardless of these
precautions, a tracked vehicle with better
traction was used. The combination of
driving over the liner with the vehicles
and snagging it with the bucket had se-
verely damaged the liner on the lagoon
bottom.
SAMPLING OF THE LINER FOR
ANALYSIS AND TESTING
The liner was inspected and samples
were collected for analysis and testing
on February 2, 1984, when the solidify-
ing operation was nearly complete. The
day was sunny and cool; about 1 ft of
snow covered the ground, the berm, and
some of the lagoon slopes, especially on
the south and west sides. Pieces of
liner from the lagoon bottom were mixed
in with the sludge and corn husks (Figure
3). The liner on the upper area of the
lagoon slopes was distorted and buckled
but in good condition. No apparent
cracking or splitting was observed in
that area, but in the lower area of the
lagoon slopes, the liner was scratched
and had many holes and tears (Figure 4).
In one place, a rock had penetrated the
liner. Samples of liner tears and
creases were collected from these areas
(Figure 5).
Two liner samples were removed from
the sludge and corn husk mixture on the
floor of the lagoon. An 18-in.-wide strip
of liner was cut from the northwest corner
of the lagoon; the strip ran from the
bottom of the lagoon to near the top. The
samples were wiped with hexane, wrapped in
aluminum foil, and double-bagged in poly-
ethylene bags. A sample of HOPE that had
not been exposed to waste was cut from a
roll of sheeting left over from the con-
struction and stored just northeast of the
lagoon (Figure 3). The outer, weathered
layer was peeled back, and the sample was
cut from inside the roll. The samples
collected are described in Table 3, and
Figure 2. Blending of ground corn husk into the sludge with a bobcat. The
husk had been dumped into the southeast corner of the lagoon.
299
-------
Figure 3. View of the north side of the lagoon. The sump is in the left (northwest)
corner and is still in the process of being pumped out in this photo. Pieces
of the liner from the lagoon bottom can be seen in the sludge/corn husk mix-
ture. An unused roll of HOPE sheeting was stored beyond the right (northeast)
corner between the two chainlink fences.
Figure 4. Extreme example of the holes and tears seen along the lower area
of the lagoon slopes.
their locations in the the lagoon are
shown in Figures 1 and 6.
The site was visited the following
day for further observations. Much of
the sludge and ground corn husk mixture
was frozen. The liquid level at the sump
had been drawn down enough to allow ex-
amination of the liner. Only a small
portion of the seam or connection between
the 2-ft-diameter polyethylene pipe and
the liner was visible and intact.
300
-------
Figure 5. A pleated sample (Sample 6) in situ before it was cut from
the lower slope of the liner in the center of the north side.
TABLE 3. SAMPLES COLLECTED FROM THE 100-MIL HOPE LAGOON LINER
Samp 1 e
number
ia
2
3a
4
5
Location in lagoon
Lagoon bottom, south-center
Lagoon bottom, southeast corner
Retained sample from roll
stored on site, northeast
of lagoon
Lower slope, southeast corner
Lower slope, south side
Size, in.
14x16
10x19
10x15
10x12
14x11
Featu re
• • •
• * •
• * •
Sample with
crack/tear
Sample with
6a
Lower slope, north side
7a Bottom to midway of slope,
northwest corner
8a Midway to top of slope,
northwest corner
tear next to
extrusion line
6x26 Pleated sample
with two
sharp creases
18x69 Strip cut from
bottom to top,
bottom half
18x114 Strip cut from
bottom to top,
top half
aAnalyzed and physically tested.
301
-------
See Northwest detail
Sump
0 20 feet
I i i
See Southwest detail
Northwest Detail
Northwest Edge
10'deep
Samples 4,5,6
Samples 1,2
Southwest Detail
Southwest Edge
7' deep
0 4 feet
Figure 6. Cross section of the lagoon from the northwest to southwest corners.
The approximate depths at which liner samples were collected are
shown in the details.
SAMPLE TESTING
Analytical and Physical Testing
The liner samples were photographed,
measured, and diagrammed upon receipt at
the laboratory. Descriptions of these
samples and their respective locations in
the lagoon are presented in Table 3. The
following samples were selected for full
testing:
- Sample 1, exposed liner from bot-
tom of lagoon.
- Sample 3, retained, unexposed mem-
brane left in roll on site.
- Sample 7, cut from the lower end of
the strip, i.e. at the slope bottom.
- Sample 8-mid, cut from the middle
section of the strip, midway up
slope.
- Sample 8-top, cut from the top of
the strip at top of slope.
These samples were tested as follows:
Property
Tensile strength
Modulus of elasticity
Tear strength
Puncture resistance
Hardness
Volatiles content
Extractables content
Test method3
ASTM D638
ASTM D882
ASTM D1004
FTMS 101C M2065
ASTM D2240
MTM 1
MTM 2
Results of the analytical and phys-
ical testing are presented in Tables 4
and 5. To assess the effect of depth on
the volatiles content of the liner, ad-
ditional testing was performed on speci-
mens cut at 2-ft intervals along the
strip, consisting of Samples 7 and 8
(Table 6).
Careful inspection of the sample
edges revealed that the sharp crease of
the pleated fold (Table 3, Sample 6) was
probably made by the blade of the bobcat
when it hit a wavy distortion of the
liner. Tensile tests of specimens cut
across the pleat yielded low tensile
strength values and breaks at yield in
the damaged area.
aASTM = American Society for Testing and Materials; FTMS = Federal Test Method Standard;
MTM = Matrecon Test Method (Matrecon, 1983).
302
-------
TABLE 4. PROPERTIES OF HOPE LAGOON LINER AFTER APPROXIMATELY 4.75 YEARS IN SERVICE
Properties
Analytical properties:
Volatiles, total loss, %
Over desiccant at 50°C
In oven for 2 h
at 105°C
Extractables, %
Physical properties:
Thickness, mil
Tensile at yield, psi
Elongation at yield, %
Tensile at break, psi
Elongation at break, %
Tensile set, %
S-100, psi
S-200, psi
Modulus of elasticity
(psi x 104)
Tear, ppi
Puncture resistance:
Gage, mil
Stress, Ib
Elongation, in.
Hardness, Durometer points:
Instant reading
5-second reading
Direction
of test
Machi ne
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Unexposed
sample of
similar
HOPE
0.06
...
...
103.0
2445
2440
20
15
4635
4445
1025
1010
900
890
1765
1710
1765
1720
7.86
7.87
839
850
98.5
131
0.73
55D
Sampling location
Retai ned
sample3,
No. 3
0.15
0.12
0.03
0.00
102.3
2705
2700
17
15
3530
4065
785
860
680
750
1920
1945
1930
1940
8.69
8.20
900
885
106
148
0.66
60D
57D
Slope
top, No.
8-top
0.22
0.15
0.07
1.26
93.4
2720
2835
18
15
4810
4355
965
875
840
760
1960
1965
1955
1963
8.98
8.68
920
910
98.5
139
0.72
58D
55D
in lagoon and sample number
Slope
midway up,
No. 8-mid
0.56
0.30
0.26
0.80
103.6
2725
2700
17
15
4510
4885
925
985
805
860
1925
1925
1920
1925
8.72
8.51
910
890
101
135
0.62
59D
560
Slope
bottom,
No. 7
1.00
0.64
0.36
1.18
100.6
2650
2665
15
18
3605
2735
760
640
655
550
1875
1965
1870
1955
8.37
7.54
855
875
100
134
0.61
59 D
56 D
Lagoon
bottom,
No. 1
2.26
1.90
0.36
0.80
92.4
2445
2440
15
17
3710
3885
810
845
700
730
1795
1790
1835
1790
5.97
6.05
830
830
93.1
118
0.59
58D
54D
aSample from unused roll of liner left onsite.
Three test dumbbells were cut across the
pleat to include the two creases. As the
pleat ran approximately in the machine
direction, this constituted a transverse
direction test. All three specimens broke
at the top crease, which was the sharper
of the two. Two of these specimens broke
at the yield point, which was at 20% elon-
gation. The average of this yield/break
value was 2280 psi for the two specimens
and was much lower than the value for the
retained sample, which yielded at 2700 psi
and broke at 4065 psi in the transverse
direction. The third specimen broke at
780% elongation, yielded at 2490 psi, and
broke at 3510 psi—still at a significant-
ly lower value than those of the retained
samples or of the samples taken at the
same depth on the north slope. Heating
the pleat at 105°C did not relieve the
creases; they appeared to be permanent.
Thermogravimetric Analysis, Differential
Scanning Calorimetry, and Specific Gravity
In addition to the volatiles and ex-
tractables analyses of the HOPE samples
removed from the lagoon, thermogravimetric
303
-------
TABLE 5. DATA ON THE RETAINED HOPE LINER SAMPLE FROM UNUSED ROLL AND
PERCENT RETENTION OF PHYSICAL PROPERTIES FOR EXPOSED SAMPLES
Sampling location and sample
Physical property
Tensile at yield
Elongation at yield
Tensile at break
Elongation at break
Tensile set
S-100
S-200
Modulus of elasticity
Tear
Hardness, Durometer points:
Instant reading
5-second reading
aSample had many scratches.
Direction
of test
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machine
Transverse
Machi ne
Transverse
Retained
sample,
No. 3
Initial
values
2705 psi
2700 psi
17%
15%
3530 psi
4065 psi
785%
860%
630%
750%
1930 psi
1945 psi
1930 psi
1940 psi
86900 psi
82000 psi
900 psi
885 psi
60D
57D
Slope
top, No.
8-top
101
105
106
100
136
107
123
102
123
101
101
101
101
101
103
106
102
103
-2D
-2D
Slope Slope
midway up, bottom,
No. 8-mid No. 7
Percent retention
101 98
100 99
100 88
100 120
128 102
120 67a
118 97
114 74a
118 96
114 73a
100 97
99 101
99 97
99 101
100 96
104 92
101 95
100 99
Change in points
-ID -ID
-ID -ID
number
Lagoon
bottom,
No. 1
90
90
88
113
105
96
103
98
103
97
93
92
95
92
67
74
92
94
-2D
-3D
TABLE 6. VOLATILES CONTENT OF SPECIMENS OF THE HOPE LINER TAKEN AT INCREASING DEPTH
Sampling location
Retai ned
sample,
Volatiles No. 3
Total volatiles, % 0.15
Over desiccant
at 50°C, % 0.12
In oven for
2 h at 105°C, % 0.03
Slope
top, Slope,
0 fta 4 fta
0.22 0.24
0.15 0.12
0.07 0.12
Slope, Slope
6 fta 8 fta
0.35 0.51
0.14 0.29
0.21 0.22
along slope
Slope,
10 ft6
0.56
0.30
0.26
from top to bottom
Slope, Slope, Slope,
12 ft5 14 ft5 16 ftB
1.08 1.26 1.00
0.73 0.82 0.64
0.35 0.44 0.36
Lagoon
bottom,
8 ftc
2.26
1.90
0.36
aCut from Sample 8 section of 18-in. wide strip taken from top to bottom of slope.
bCut from Sample 7 section of 18-in. wide strip taken from top to bottom of slope.
cCut from Sample 1 taken from the lagoon bottom which was about 8-ft deep at that point.
304
-------
analyses (TGA) and differential scanning
calorimetry (DSC) were performed. Speci-
fic gravity of the samples was also
determined. The results are presented in
Table 7 and are compared with those from
the data base in our laboratory on HOPE
sheeting from U.S. and German production.
The following observations can be made:
- The sample from the lagoon bottom
had a lower amount of crystallinity
than the retained sample from the
roll and the creased sample from the
side slope.
- The retained and the creased samples
had lower melting points and slightly
higher crystal!inity than the similar
German sample.
- Overall, the data-on the retained
liner sample and those in the data-
base on German membranes appeared
similar to each other and different
from those on the U.S.-produced
sample.
Gas Chromatographic Headspace
(GCHS) Analysis
As shown in Table 2, the impounded
waste liquid contained significant
amounts of organics that could be absorb-
ed by HOPE to some extent. Data on vol-
atiles and extractables confirm that a
small amount of absorption took place.
Also, the HOPE exposed at the bottom of
the lagoon softened—that is, it lost in
modulus and in hardness. Since some of
the organics are volatile, GCHS and GC
pyroprobe were explored to assess the
absorption by the HOPE, at least qual-
itatively. A large variety of organics
appeared to have been absorbed. Some
of the volatiles specifically listed in
the analyses in Table 2 were identified
by GCHS. Some nonvolatile plasticizers
were also identified.
DISCUSSION
The amount that the liner swelled
(as evidenced by the volatiles content)
TABLE 7. COMPARISONS OF THE TGA, DSC, AND SPECIFIC GRAVITY OF THREE LINERS
HOPE lagoon liner sample A similar
Analytical properties
Thermogravimetric
analysis:
Scan number
Volatiles, %
Polymer, %
Plasticizer, %
Carbon black, %
Ash, %
Bottom3, Top, Retained, Creasedb, German
No. 1 No. 8 No. 3 No. 6 Smooth
411
0
97.6
0
1.7
0.7
410
0
97.7
0
1.8
0.5
36
0
98.1
0
1.9
0
liner
Rough
37
0
98.6
0
1.4
0
U.S. liner
(domestic
materi al )
38
0
97.7
0
2.3
0
Differential scanning
calorimetery:
Scan number
TmC, °C
AHfd, cal/g
Cry stall inity6, %
Specific gravity
411
131
29.8
43.6
0.936 0.947
412
130
35.3
51.6
0.943
419
129
36.0
52.6
15
123
33.6
49.1
0.943
16
124
33.7
49.2
0.939
12
119
24.8
36.2
0.957
aSample had thoroughly dried out during shelf aging before being tested.
bSample taken at the apex of the crease in the pleat.
cMelting temperature of crystalline phase.
^Heat of fusion of crystalline phase.
ePercent crystallinity based on AHf value of 68.4 cal/g for 100% crystalline HOPE.
305
-------
increased with depth into the lagoon. The
bottom sample (No. 1) contained about 2.3%
volatiles, and the sample from the top of
the strip (No. 8) contained approximately
0.2% volatiles. The retained sample from
the roll (No. 3) also had about 0.2% vol-
atiles. All five of the tested areas (the
four exposed to waste and the retained
sample) showed good physical properties,
but the sample from the bottom of the
lagoon (No. 1) was the softest (lowest
durometer and modulus) and showed the low-
est values for several physical prop-
erties. The sample from the bottom end
of the slope strip (No. 7) had low values
in tensile break in the transverse direc-
tion. This sample was observed during
testing to be quite scratched, which pro-
bably explains the low values. Data for
other liner properties from this area are
much higher. The liner sample from the
top of the strip (No. 8) showed higher
tensile properties than the retained
sample (No. 3).
CONCLUSIONS
The exposed liner sample from the
lagoon bottom showed a small amount of
swelling (about 3%) and 10% loss in both
tensile at yield and in modulus. These
changes do not demonstrate incompatibi-
lity, but they do indicate that the waste
at the bottom of the lagoon had the great-
est impact on the liner.
Much buckling and distortion of the
liner was observed, especially on the
north slope of the lagoon. The day the
samples were taken was cold but sunny,
which should have minimized any thermal
expansion. On a hot day, this buckling
would have been greatly increased. Though
not in itself a significant problem, the
buckling places added stress on seams,
allows movement of the underlying earthwork
by creating cavities, and increases the
chance of mechanical damage by introducing
folds.
The liner showed no cracking such as
that encountered with environmental stress-
cracking; it lost in tensile strength at
yield and in ultimate tensile strength,
but it did not crack at folds or bends.
No evidence was apparent to indicate
that the HOPE liner lost its integrity.
All visible seams looked secure. The only
evidence of degradation was mechanical
damage in areas where the liner had pro-
bably been worked on with the bobcat. The
presence of rocks directly underneath the
liner aggravated this damage. The liner
in this impoundment was not able to with-
stand the clean-up operations.
DSC data indicate that the amount of
polyethylene crystallinity in the HOPE
liner changed during exposure, a factor
that should be checked in future testing.
Decommissioning of liner waste im-
poundments under Superfund Remedial
Actions affords an excellent opportunity
to collect information on the performance
and durability of lining materials.
Efforts should be made to incorporate
liner recovery and testing when such
sites become available.
REFERENCE
1. Matrecon, Inc., Lining of Waste Im-
poundments and Disposal Facilities,
SW-870, Revised (March 1983), U.S.
Environmental Protection Agency,
Washington, D.C. (GPO 055-00000231-2).
306
-------
STRENGTH AND DURABILITY OF FLEXIBLE MEMBRANE LINER SEAMS AFTER
SHORT-TERM EXPOSURE TO SELECTED CHEMICAL SOLUTIONS
Mary Ann Curran
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Ronald K. Frobel
U.S. Bureau of Reclamation
Denver, Colorado 80225
ABSTRACT
In March 1983, the United States Environmental Protection Agency (USEPA)
began research with the United States Bureau of Reclamation (USBR) to evaluate
flexible membrane liner (FML) seams exposed to selected environmental
conditions over short periods. Thirty-seven combinations of reinforced and
nonreinforced FML materials joined by various seaming methods have been
subjected to six chemical solutions, brine, tapwater, and accelerated outdoor
aging by ultraviolet light. Effects of these types of environmental
conditions were evaluated using shear and peel strength tests before and after
exposure. The tests were performed under dynamic load at room temperature,
and static dead load at 50°C.
To date, tests have been completed on unexposed seams and seams that have
been exposed for 3 months at room temperature. Exposures for 6 and 12 months
are currently underway. The results and evaluation of the mechanical tests
completed on 3-month chemical exposures are presented in this paper.
INTRODUCTION
Laboratory tests and pilot-scale
studies have been conducted by various
Government and non-Government group to
assess the effects of chemical waste
products on the integrity of FML's;
however, little has been done to assess
the performance of the various factory
and field seams used in joining the
manufactured roll goods in the factory
and the panels seamed in the field.
The objective of this study is to
evaluate bonded seams of commonly used
FML materials and the bonding methods
employed. Destructive tests are being
conducted in an attempt to detect
physical changes in these seams when
exposed to selected chemicals and
simulated environmental conditions.
This paper describes the liner
samples which were obtained for study,
field seam fabrication, the immersion
test setup, and UV (ultraviolet) light
exposure testing. Some preliminary
results are presented on seam strength
tests after short-term (3-month)
chemical immersion at room temperature.
APPROACH
Materials Used in the Study
The polymeric materials evaluated
for this study were obtained from 12
industry representatives who manufacture
and/or fabricate FML's which are
commonly used in pollution and seepage
control. Manufacturers and fabricators
were contacted and asked to participate
in the study in an attempt to obtain a
307
-------
broad selection of recommended membrane
types (including reinforced and
nonreinforced samples), scrim, and
seaming techniques.
Polymer sheet materials submitted by
industry for this project include the
following:
High Density Polyethylene (HOPE)
30-mi1 nonre i nforced
High Density Polyethylene (HOPE)
80-mi1 nonre i nforced
Linear Low Density Polyethylene (LLDPE)
30-mi1 nonreinforced
Ethylene Propylene Diene Rubber (EPDM)
30-mi1 reinforced
Polyyinyl Chloride (PVC)
30-mi1 nonreinforced
Chlorinated Polyethylene (CPE)
30-mi1 nonreinforced
Chlorinated Polyethylene (CPER)
36-mil reinforced
Chlorosulfonated Polyethylene (CSPER)
36-mil reinforced
Ethylene Interpolymer Alloy (EIA)
38-mil reinforced
Some sources submitted more than one
sample and/or seaming method of PVC,
HOPE, CPE, and CSPE. One composite seam
of PVC/CPE was also submitted. Various
scrim types, thickness, and seaming
methods were represented in the
submittals.
Field Seam Fabrication
In addition to the polymer roll
goods and factory seams submitted by
industry, field seams were constructed
at the Bureau of Reclamation's
Engineering and Research Center under
representative field conditions.
The field seams were fabricated by a
third party commercial installer in
accordance with manufacturer's
instructions and utilizing manufacturer-
supplied equipment and products where
required. Some field seams which could
not be constructed at the Bureau's
facility were witnessed and collected at
job sites by Bureau personnel.
Factory and field seam types
submitted or field-fabricated for
evaluation in the project were as
follows:
Thermal-hot air
Thermal - dielectric
Thermal - hot wedge
Extrusion - fillet weld
Extrusion - lap weld
Bodied solvent adhesive
Gumtape/cement
Vulcanized
Thirty-seven combinations of sheet
materials and seaming methods are being
evaluated in the chemical and water
immersions and the accelerated outdoor
aging.
Chemical and Aqueous Solution Immersions
All nonreinforced and reinforced-FML
seams were cut into large samples for
environmental exposure. The
nonreinforced seams were cut into 300-mm
(12-in) wide samples in order to later
yield twelve 25-mm (1-in) wide specimens
for mechanical testing and the
reinforced samples were cut into 450-mm
(18-in) wide samples to provide four
50-rrm (2-in) wide and eight 25-mm (1-in)
wide specimens.
To allow for possible edge effects
due to scrim exposure, 25-mm (1-in) of
each side of the scrim-reinforced
samples were cut off and discarded. All
samples were attached with polypropylene
ties to fiberglass rods for exposure
testing.
The six chemical solutions, chosen
to represent a wide range of chemical
groups, are:
10 percent phenol (organic acid),
10 percent hydrochloric acid
(inorganic acid),
10 percent sodium hydroxide
(inorganic base),
10 percent methyl ethyl ketone (ketone),
5 percent furfural (aldehyde),
308
-------
100 percent methylene chloride
(halogenated hydrocarbon).
Methylene chloride is not soluble in
water; therefore, pure solvent was used
to avoid the problem of phase
separation. Pure chemicals or aqueous
chemical solutions were selected for
testing rather than simulated or actual
wastes from waste sites to simplify
verification of testing procedures. The
use of representative groups of
chemicals also allows for reasonable
interpretation of the data.
Polyproplyene and polyethylene
tanks, 45-gallon (170 L) capacity, were
filled separately with each of the above
chemicals, the seam samples were
attached to the fiberglass rods, and the
samples were then suspended into the
chemicals.
In parallel with the tank
immersions, smaller coupons of the liner
materials are immersed in small clear
glass jars for periodic weighings and
thickness measurements. The smaller
coupons allow for easier inspection of
the polymeric sheet materials for
obvious excessive degradation, swelling,
or change in color or surface texture.
If any accelerated response is observed
in the coupons, the seam samples are
removed from the larger tanks before
they are destroyed completely. The
ratio of the volume of liquid to the
surface area for each coupon is 6.2
ml/cm2 (40 ml/in2).
Three tanks of room temperature
tapwater and six tanks of saturated
sodium chloride brine solution (three
tanks at room temperature and three at
50°C (122°F) were also set up for
immersing seam samples. Double-sided
exposure of all samples was used to
accommodate the large number of samples
in minimum space. The accelerated
effect of double-sided exposure compared
to one-sided exposure reduces the time
it would take to see effects the liquids
may have on the samples.
Accelerated Outdoor Exposure
In addition to the above exposure
conditions, the Desert Sunshine Exposure
Test Laboratories in Phoenix, Arizona,
have performed accelerated outdoor
sunlight exposure testing on
representative seam samples. The
exposures were accomplished on the
EMMAQUAvR) accelerated weathering test
machine located at the test facility.
The EMMAQUA'R' machines are capable of
tracking the sun and focusing the sun's
rays on the 5-inch-wide seam specimens
for optimum UV exposure. One year of
concentrated sunlight exposure is
approximately equal to 5 to 7 years of
actual sunlight exposure. After recent
completion of exposure, the samples were
visually inspected, photographed, and
returned to the Bureau for testing to
determine any change in peel strength.
This testing had not been completed at
the time of preparation of this paper.
Physical/Mechanical Testing
At the end of 3-month, 6-month, and
12-month chemical, brine, and water
exposures and after completion of the
accelerated outdoor exposure, the
following physical and mechanical tests
are being performed to determine changes
in physical properties:
1. Volume (thickness) and weight -
Coupon samples were measured for volume
and thickness before immersion. Exposed
coupon samples have been measured every
week for the first month, monthly for
the next 3 months, and quarterly
thereafter.
2. Dynamic tensile shear testing -
50-mm (2-in) wide reinforced or 25-mm
(1-in) wide nonreinforced seam specimens
are tested in tensile shear using
dynamic loading at a low strain rate of
0.8 mm/s (2 in/min). The reinforced
samples are wider for tensile testing in
an effort to eliminate edge effects due
to the scrim reinforcement. Results are
in N/m (lbf/in) of seam width.
3. Dynamic peel testing - 25-mm (1-in)
wide specimens are subjected to 180°
peel at a strain rate of 0.8 mm/s (2
in/min). Results are in N/m (lbf/in)
of seam width.
4. Static dead load peel testing at
elevated temperature - 22-mm (1-in) wide
specimens are subjected to a 180° peel
dead load equivalent to less than 5
percent of the material's breaking
strength per inch of seam width at a
309
-------
temperature of 50°C (122°F).
Results will be expressed by pass/fail,
time to failure, and mode of failure.
A summary of seam and peel strength
under dynamic load, and failure mode for
both original dry samples and samples
subjected to 3-month chemical immersion
is presented in Tables 1 and 2.
Discussion of Preliminary Results
There is considerable variation
between seam types in the 3-month test
results, as seen in Tables 1 and 2.
Early in the exposure period, most of
the seam samples immersed in methylene
chloride (MCI) were destroyed and
removed from the exposure tanks. Thus,
the data in Tables 1 and 2 for MCI
include only seven entries.
No attempt has been made to
quantitatively analyze the results after
3-month exposure. After all data for
all exposure periods have been stored on
a computer file, a thorough analysis
will be completed. However,
qualitatively, it can be seen in these
preliminary data that many of the
failure modes were unchanged in samples
before and after 3-month exposure to all
the chemicals. In addition to MCI, it
is evident that phenol had a deleterious
effect on most samples after 3 months -
both in shear and peel.
Some samples, particularly the PVC,
show increased shear strength values due
to the migration of plasticizer and
subsequent stiffening after immersion in
some of the chemicals. Samples that do
well in shear, however, may not do so
well in peel as evidenced by even these
3-month test results. The peel test is
showing deterioration of bond after
chemical immersion; whereas, the shear
test may not readily indicate potential
deterioration of bond.
After all test results are compiled
and analyzed for the 3, 6, and 12 months
of exposures, it is anticipated that
definite aging trends will surface for
the combinations of materials and
bonding techniques examined in this
research program.
310
-------
Table 1. - Seam shear strength after 3 months exposure to six chemical solutions
Immersion
Material
36 ml 1 CPER
36 ml 1 CSPER
36 ml
36 ml
36 ml
36 ml
36 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
80 ml
30 ml
80 ml
30 ml
80 ml
30 ml
30 ml
30 ml
30 ml
36 ml
30 ml
36 ml
30 ml
36 ml
30 ml
36 ml
36 ml
36 ml
36 ml
30 ml
38 ml
CSPER
CPER
CSPER
CSPER
CSPER
CSPER
CSPER
CSPER
LLDPE
LLDPE
EPDM
HOPE
EPDM
HOPE
PVC
HOPE
PVC
HOPE
CPE
CPE
CPE
CPE
CSPER
PVC
CSPER
PVC
CSPER
PVC/CPE
CSPER
CPER
CPER
CSPER
PVC
EIA
38 mil EIA
Seam,
type'
A
A
F
F
A
F
A
F
A
F
C
C
1
D
H
0
6
E
G
C
G
G
A
G
B
A
F
G
F
A
F
A
F
F
G
A
A
Original data
Shear Fall
(Ibf/ln) mode2
78.4
90.4
90.0
94.3
31.0
31.2
63.8
65.0
79.9
65.0
65.5
62.5
70.2
73.7
91.4
203.0
57.2
242.0
47.7
224.0
42.0
33.8
33.1
33.6
73.5
52.9
91.8
53.2
1 18.7
38.2
138.4
123.4
124.5
97.4
59.6
353.8
356.6
7
6
7
7
7
7
7
7
7
7
9
9
7
9
7
9
2
9
2
9
2
2
2
2
7
2
7
2
7
2-CPE
7
7
7
7
2
6
6
Phenol
-' lot
HCl -
Shear Fall Shear
(Ibf/ln) mode'' (Ibf/ln)
7.4
16.4
18.8
12.3
24.7
25.2
24.9
25.6
32.8
54.7
61.2
66.0
58.2
91.0
61.3
243.0
65.8
267
60.5
247.0
3.4
9.5
11.3
4.3
28.0
71.8
50.4
64.5
29.8
12.7
64.0
61.0
48.8
38.8
66.4
120.1
9
5
5
5
5
7
5
5
4
5
5
6
6
5
6
5
6
2
6
2
6
3
2
6
3
5
6
5
6
5
3
5
5
5
5
2
3
9
36.5
40.5
64.7
58.9
31.1
32.5
50.5
53.5
139.0
89.5
61.6
66.6
88.0
88.6
58.9
219.0
48.6
265.0
52.8
220.0
35.6
31.6
30.4
38.4
70.9
53.2
119.7
58.8
86.3
26.0
105.8
150.0
111.5
124.9
62.6
9
9
16*
NaOH -
Fall Shear
mode2 (Ibf/ln)
7
7
7
7
7
7
7
7
7
7
6
6
7
6
7
6
2
6
2
6
2
2
2
2
7
2
7
2
7
2-CPE
7
1
7
7
2
9
9
22.5
31.7
44.8
27.3
24.8
24.6
35.0
41.2
42.0
54.3
65.6
67.8
67.6
78.4
44.2
256.6
78.4
256.0
81.4
231.0
44.1
39.0
33.2
39.3
79.2
123.5
80.3
78.8
98.0
33.3
99.9
143.5
60.0
80.0
76.0
9
9
Media
TOT
Fa 1 1
mode
7
7
7
5
5
5
7
7
7
7
6
2
7
6
5
6
2
2
2
6
2
2
2
2
7
6
5
6
7
2-CPE
7
1
3
5
2
9
9
M£k - tOf MCI - 100J
Shear
(Ibf/ln)
31.4
45.0
38.1
36.7
32.0
32.8
45.3
46.5
57.2
51.5
58.0
61.0
82.0
77.5
70.9
229.0
52.4
247.0
49.4
223.5
30.4
26.1
29.0
31.8
61.3
50.1
138.5
59.4
80.5
18.8
145.5
140.5
115.9
122.5
67.5
9
9
Fall Shear ^11
mode2 (lbt/ln> mode2
7
7
7
7
7
7
7
7
7
7
6 66.0 6
6 67.0 6
7 111.4 7
6 85.0 6
7 -
6 268.5 6
2 — —
6 253.3 6
2 -
6 225.0 6
2
2
2
2
7
2
7
2
7
2-CPE
7
7
7
7
2
9
9
Furfural
Shear
(Ibf/ln)
26.8
43.5
73.0
63.0
34.8
34.7
46.8
38.8
71.5
64.5
58.0
59.0
92.0
83.7
69.0
205.5
51.8
263.0
66.0
224.0
16.7
14.1
15.2
18.3
100.0
53.8
121.5
56.7
97.0
15.9
148.0
124.5
99.0
144.8
66.0
9
9
- 10J
mode2
7
5
7
7
7
7
7
5
5
6
6
6
7
6
7
2
2
6
2
6
2
2
2
2
7
2
7
2
7
2-CPE
7.
5
6
7
2
9
9
1 A: Thermal - hot air; B: Thermal - dielectric; C: Thermal - hot wedge; D: Extrusion - fillet weld; E: Extrusion - lap
weld; F: Bodied solvent adhesive; (i: Solvent adhesive; H: Gumtape/cement; I: Vulcanized.
2 1: Parent material failure at clamp edge; 2: Parent material failure at seam edge; 3: Within the adhesive bonding system;
4: Planar failure between adhesive and liner surface or between two thermal-bonded surfaces; 5: Scrim delamination and
and pullout; 6: Failure in parent material (other than 1, 2, or 5); 7: Scrim break; 8: Initial peak with subsequent ply
adhesion; 9: No failure.
-------
Table 2. - Seam peel strength after 3 months exposure to six chemical solutions
Material
36 ml
36 ml
36 ml
36 ml
36 ml
36 ml
36 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
30 ml
80 ml
30 ml
80 ml
30 ml
80 ml
30 ml
30 ml
30 ml
30 ml
36 ml
30 ml
36 ml
30 ml
36 ml
30 ml
36 ml
36 ml
36 ml
36 ml
30 ml
38 ml
38 ml
CPER
CSPER
CSPER
CPER
CSPER
CSPER
CSPER
CSPER
CSPER
CSPER
LLDPE
LLDPE
EPDM
HOPE
EPDM
HOPE
PVC
HOPE
PVC
HOPE
CPE
CPE
CPE
CPE
CSPER
PVC
CSPER
PVC
CSPER
PVC/CPE
CSPER
CPER
CPER
CSPER
PVC
EIA
EIA
Seam
type
A
A
F
F
A
F
A
F
A
F
C
C
1
D
H
D
6
E
G
C
G
G
A
G
B
A
F
G
F
A
F
A
F
F
6
A
A
Original
Peel
(Ibf/ln)
59.8
48.3
53.4
71.8
29.6
31.7/22.5
61.1
31.0
33.6/15.1
32.8/17.7
64.4
60.6
36.2/21.2
72.3
8.6
123.0
14.8
175.0
15.7
120.8
21.3
17
29.6
12.3
30.9/11.4
38.1
34.4/17.7
19.2
3o!o/14.8
29.8
31.7/15.1
16.2
23.4
29.5/16.2
24.8
43.8/20.5
39.3/16
data
(-all
mode*
7
7
6
6
2
8
7
3
8
8
9
9
8
2
3
2
4
2
3
4
2
3
2
3
8
2
8
3
8
2-CPE
8
4
3
8
3
8
8
Phenol
Peel
(Ibf/ln)
8.2
24.0
30.4
21.5
24.0
23.1
12.8
2.9
35.4/12.8
38.2/21.8
64.2
70.1
43.1/21.5
87.0
5.9
143.3
23.3
161.3
26.9
150.3
<0.3
0.45
8.4
<0.3
24.4/9.24
43.1
25.5/17.8
13.7
33.0/20.2
4.1
21.8/12.8
7.5
3.4
29.6/14.9
19.0
2.3
19.3
IQI 11*1 — m»
Fal 1
mode
4
4
4
5
5
5
4
4
8
8
6
6
8
6
4
2
3
4
3
4
3
3
4
3
8
2
8
3
8
3
8
3
3
8
3
3
4
Peel Fall
* (Ibf/ln) mode*
36.0
41.9
63.6
65.7
29.9
30.5
51.7
12.2
36.4
35.1
61.7
64.8
35.8
80.6
5.4
77.5
15.4
191.4
12.3
169.8
20.1
13.2
21.8
14.2
31.9/13.4
39.5
25.3
21.5
32.3/19.3
19.4
30.2/15.4
20.2
19.5
37.3/21.8
30.5
41.2/11.5
31.3/17.1
5
7
7
7
5
5
4
3
4
4
6
6
4
2
4
2
3
2
3
2
3
3
2
3
8
2
7
3
8
3
8
3
4
8
4
8
8
MnftH -
reoun
Peel
(Ibf/ln)
35.8
39.1
54.2
45.7
27.0
26.1
36.4
1 1.6
31.6/11.8
27.7/17.5
65.7
65.1
40.9
69.0
6.4
97.2
25.4
144.3
34.7
152.7
22.7
15.1
25.0
12.7
30.5/15.4
23.9
29.3/16.3
27.6
25.5/18.4
20.8
26.8/16.4
26.6
-
25.1/16.4
31.9
25.7/23.7
31.7/17.9
~nj|~~
UFV -
Fa 1 1 Pee 1
mode2 (Ibf/ln)
7
7
7
7
5
5
4
3
8
8
2
2
4
2
3
2
3
2
3
2
3
3
4
3
8
2
8
3
B
2
8
4
-
8
3
3
8
33.8
48.9
67.8
56.1
24.9
30.9
47.7
3.8
27.0
24.9
58.8
59.7
34.7
65.4
5.5
158.7
15.4
148.2
10.1
170.4
19.1
14.4
20.0
11.7
30.8/14.5
35.4
25.7/15.3
18.5
28.6/16.3
15.6
22.3/14.0
12.4
17.3
30.8/18.4
29.5
31.3/12.5
34.2/17.9
TO?
Fall
mode^
7
5
6
7
4
5
7
3
4
4
6
6
4
6
3
2
3
2
3
2
6
3
2
4
8
2
e
3
8
2
4
4
4
8
4
8
8
Peel Pall Peel
(Ibf/ln) mode^ (Ibf/ln)
30.8
48.5
74.3
66.8
31.1
14.7
20.6
7 4
29.9/15.5
27.2/15.5
2.70 6 61.7
3.10 6 64.2
14.70/0.55 8 38.4/19.4
7.80 2 86.3
6.9
41.00 2 149.3
4.5
21.50 2 213.0
13.3
9.50 2 169.2
3.8
4.9
9.8
4.1
30.9/12.8
37.1
28.9/17.0
8.8
30.6/18.6
1 1.2
18.0
25.8/10.5
17.4/24.5
20.8
19.7
5.7
31.4/18.9
Fal U
mode*
7
7
7
7
5
3
4
T
J
8
8
6
6
8
2
3
2
3
2
3
2
3
3
3
3
8
2
8
3
8
4
3
8
8
3
3
3
8
1 A: Thermal - hot air; B: Thermal - dielectric; C: Thermal - hot wedge; D: Extrusion - fillet weld; E: Extrusion - lap
weld; F: Bodied solvent adhesive; U: Solvent adhesive; H: Uumtape/cement; I: Vulcanized.
2 I; Parent material failure at clamp edge; 2: Parent material failure at seam edge; 3: Within the adhesive bonding system;
4: Planar failure between adhesive and liner surface or between two thermal-bonded surfaces; 5: Scrim delamination and
and pullout; 6: Failure in parent material (other than 1, 2, or 5); 7: Scrim break; 8: Initial peak with subsequent ply
adhesion; 9: No failure.
-------
CHEMICAL MASS TRANSPORT MEASUREMENTS TO DETERMINE
FLEXIBLE MEMBRANE LINER (FML) LIFETIME
Arthur E. Lord, Jr., Robert M. Koerner and Eric C. Lindhult
Drexel University
Philadelphia, PA 19104
ABSTRACT
The prediction of the service life of flexible membrane liners (FML) when exposed to
chemicals has usually been by way of testing for physical or mechanical property changes
after periodic exposure times. This paper presents an alternate approach by evaluating
the chemical properties via four different diffusion related measurements of the exposed
FML. These tests are water vapor transmission (WVT), radioactive tracer transmission
(RT), water absorption (WA) and water vapor absorption (WVA).
This approach was tested for up to 6-month exposure times. The FMLs used were PVC,
EPDM and CPE. The chemicals used were 10% NaOH (in water), 10% H2S04 (in water), 10%
phenol (in water) and 100% xylene. The WVT and RT were found to be quite reliable test
methods, whereas the WA and WVA techniques experienced serious problems. The various
transport coefficients showed all the expected types of behavior with chemical exposure:
. constancy with exposure (acids and bases on all FMLs)
decrease with exposure (plasticizers leaching from FMLs)
. increase with exposure (Phenol-treated CPE and EPDM in H2S04)
The WVT and RT results were generally-comparable in all but one situation, namely
the CPE-phenol case. Here WVT showed about a fourfold increase in permeability, whereas
RT showed essentially no change in diffusivity.
The work reported here, even though it is preliminary, lends credence to the use of
mass transport measurements to determine structural change in FMLs and hence a prediction
of their lifetime.
INTRODUCTION
Before any meaningful design can be-
gin on a FML-based liner or cover, its
chemical compatibility with the contained
materials must first be assured. Further-
more, this compatibility must be guaran-
teed for the lifetime of the facility. The
challenge becomes quite formidable, how-
ever, when one deals with a complex waste
stream from a landfill, i.e., leachate, or
with waste streams from proposed facilities
still in the design stage. The approach
generally taken in such cases is to place
the candidate FMLs in the actual or synthe-
sized waste stream, expose them for a
period of time, and then test them for
changes in physical or mechanical proper-
ties. Depending on the magnitude of these
changes, each particular geomembrane is
then accepted or rejected. Problems in-
variably come up in these decisions, how-
ever, as to what are tolerable or intolera-
ble changes in properties. Obviously, the
ideal is no change. Yet, most FMLs do
respond (even to water) with changes in
thickness, volume, stiffness, modulus,
elongation and strength. Often these
changes are reasonably uniform with time.
313
-------
PURPOSE
The present approach still uses an ex-
posure of the FMLs to the proposed li-
quid, but now assesses the material's per-
formance on the basis of its chemical pro-
perties - namely diffusion and permeabil-
ity, instead of using mechanical and phy-
sical properties as the indicators.
The basic hypothesis is that if the
FML has been degraded by the liquid to
which it has been exposed, its internal
structure will have been altered. If so,
some form of diffusion properties (actu-
ally its mass transport characteristics)
would probably be among the most sensitive
parameters to evaluate. The trend in mass
transport coefficient as a function of
chemical exposure time could lead to a
predictive method for FML lifetime deter-
mination. For example, it is well known
in polymer science, that if the glass tem-
perature is approached from the high tem-
perature side (say from loss of plastici-
zer), the mass transport properties of the
polymer can decrease significantly.
APPROACH
The following tests were evaluated
and are discussed:
. water vapor transmission (WVT)
. radioactive tracer transmission (RT)
. water absorption (WA)
. water vapor absorption (WVA)
What is anticipated in these diffusion mea-
surements is that abrupt or long term
changes in the measured diffusion parameter
will indicate unacceptable performance in
a more clear-cut and unequivocal manner
than physical or mechanical property test-
ing. Stated differently, the structural
changes, as measured by diffusion tests,
might be more (or at least as) sensitive
to chemical degradation of the FML as the
structural changes, as measured by physi-
cal or mechanical tests.
Graphically, the anticipated behavior
is as shown in the three alternates of
Figure 1. Curve (a) shows an unwanted be-
havioral trend by having a breakthrough
occur and a rapid increase in the measured
diffusion property. Curve (b) illustrates
the case where chemical degradation has
occurred and the products of the reaction
have "blocked" the microstructure result-
ing in an abrupt decrease in the measured
diffusion property. (This could also be
interpreted as a "stiffening" of the
matrix.) Finally, in curve (c) the mea-
sured diffusion property remains constant
showing an unaffected geomembrane diffu-
sion to the liquid it was exposed to.
Curve (c) is the preferred response.
Conditioning of Geomembranes
Conditioning of FMLs (exposed to
waste) is vital to assessing their ulti-
mate in-situ performance. The procedure
selected should conform to the following
requirements.
. representative of field conditions
. reproducible from one sample to
another
. realistic to perform on a routine
basis
Tub exposure is used here where the
FML is folded into an empty tub which is
used for its support and then the chemical
liquid is placed within this FML lining.(8)
Folds are indeed present, at 90° and 180°,
as well as flat surfaces along the bottom
and partially wet and dry surfaces along
the sides. Most results given here will
be for the flat samples. This procedure
seems to closely simulate most of the FML
lined field containment sites.
Three tub exposure setups have been
constructed out of 4' x 8' plywood sheets
(3/4" thick, exterior grade), each con-
taining 12 tubs with FMLs in each. Three
FML types (PVC, CPE and EPDM) were exposed
to four chemicals (10% phenol, 10% hydro-
chloric acid and 10% sodium hydroxide -
all in water solution, and 100% xylene).
They were removed after short, medium and
long exposure times (4 weeks, 12 weeks and
6 months) and when combined with the as-
received data, give four data points to
evaluate the aging process. The exposed
FMLs were water-rinsed and air-dried after
removal from the tubs. (Some were also
vacuum-dessicated after this process to
try to "dry" the samples completely).
Next the individual specimens for diffusion
testing were taken from the flat areas (no
bend), the side portion (90° bend) and the
corners (180° bend), although the majority
of the work presented here is for the flat
area samples. Test specimens were then
stored in sealed plastic bags until testing
commenced.
314
-------
Methods of Measurement
a) Water Vapor Transmission
Water vapor transmission (WVT) tests
were conducted according to the ASTM E-96
standard tests(1), with mechanical clamps
being used for sealing these relatively
thick (30 mil) FMLs. The test cups (con-
taining water) were placed in an environ-
mental chamber maintained at constant hu-
midity and temperature. The cups were
withdrawn at periodic intervals and
weighed on a sensitive analytic balance.
Calculation of WVT resulted from weight
versus time according to the ASTM
Standard(l).
b) Radioactive Tracer Transmission
The diffusion coefficient of a parti-
cular molecule can be determined if part
of the molecule can be made radioactive.
The diffusion coefficient is determined by
monitoring the radioactive disintegrations
versus time at some convenient point. We
have used an energetic e-emitter, carbon-
14 in benzene for the diffusing molecule.
Figure 2 shows a schematic diagram of the
test device. The diffusion coefficient is
extracted from the counts versus time at
the Seiger-Mueller tube face, using
standard diffusion theory(5>°). The theory
assumes one dimensional diffusion (i.e.,
a wide, thin sample).
c) Water Absorption
The diffusion coefficient for water
in a particular FML can be determined by
immersing the sample in water and monitor-
ing the weight gain (i.e., its absorption)
as a function of time. Diffusion theory
is again used to obtain the diffusion co-
efficient^,9). in most cases, however a
final weight was not achieved making the
analysis impossible.(10)
dj Water Vapor Absorption
In precisely the same manner as the
-water absorption test just described, the
use of water vapor as an absorbed gas can
be used to obtain the diffusion coef-
ficient. The theory is exactly the same,
and only a slight modification in the ex-
perimental setup is necessary. Now, in-
stead of directly suspending an FML in a
controlled temperature bath, it is en-
capsulated in a separate glass cylinder
with water in the bottom which in turn is
suspended in the water bath. Thus 100%
humidity conditions are present, which al-
lows for water vapor to enter the FML as a
function of time. The increase in weight
is plotted against time and diffusion
theory used to extract the diffusion co-
efficient. However, the results we found
are very suspect, since in most cases the
initial part of the curve does not follow
the t* law as required from diffusion
theory.(10)
The details of all four test proce-
dures (with actual data) can be found in
Lord and Koerneru).
PROBLEMS ENCOUNTERED
The major problem encountered during
this study was the inability of the ab-
sorption methods (WA and WVA) to give con-
sistently meaningful values for the diffu-
sion coefficients. Running duplicate (and
in many cases triplicate) samples does not
mitigate this problem. This precludes the
use of these methods to reliably monitor
diffusion changes due to chemical exposure,
even though very general (order of magni-
tude) values for the diffusion coef-
ficients can be obtained, if enough
samples are run.
In the water absorption method, for a
great many cases, a final weight was not
achieved even after 200 days of immersion.
Without a final, equlibrium weight, the
diffusion coefficient cannot be obtained.
In the water vapor absorption method,
for a great many cases, the initial time
dependence for absorption is nowhere near
ti, even though a constant weight is ac-
hieved in most cases. Thus a reliable
diffusion coefficient cannot be obtained,
because simple diffusion is not taking
place.
Other problems encountered were minor
compared to these.
RESULTS
The results of each test method on
each FML will be discussed separately.
Table 1 gives a summary of all results.
In Table 1 is given the percentage change
after six month exposure, plus the trend
occurring at the six month time (increas-
ing, decreasing or steady). The criteria
315
-------
for effect of the chemical on the FML is
twofold:
. the absolute change in mass trans-
port property
. the trend (time variation) in the
mass transport property.
Both criteria must be fulfilled for a dele-
terious effect to have occurred, namely a
large absolute increase and a significant
rate of increase (with exposure time) at
6 month's exposure.
a) WVT Results
The results of chemical exposure are
shown in Figures 3, 4 and 5. The results
are the average of two runs, which never
differed from each other by more than 5%.
In Figure 3, it is seen that the NaOH and
H2S04 solutions (10%) have very little ef-
fect on the WVT of PVC, even for periods
of one half year. On the other hand, the
chemicals which "leach" plasticizer from
PVC, namely phenol and xylene, produce a
very definite lowering of the WVT. PVC is
noticeably stiffer after exposure to these
two chemicals. It is very well known in
the polymer field that if a materials be-
comes stiffer, the transport properties
usually decrease in magnitude. This re-
sult lends much credence to the use of
simple transport measurements to monitor
the structure of the FMLs.
Figure 4 shows that the WVT in EPDM
is affected very little by 6 months' ex-
posure to the four chemicals.
Figure 5 shows that the WVT in CPE is
affected very little by acids and bases,
but significantly increases upon exposure
to phenol. CPE is completely destroyed by
xylene in about a day, so no WVT results
are shown for this combination.
Thus it is seen that the WVT can be
significantly reduced (Fig. 3), remain the
same (Fig. 4) or increase significantly
(Fig. 5) upon chemical exposure. The most
interesting result would seem to be the
effect of phenol on CPE, for if the WVT
trend continues upward, a time will be
reached very soon where the WVT would
reach an unacceptably high value for con-
tainment of phenol-type wastes.
b) Radioactive Tracer Measurements
Figures 6, 7 and 8 show the results of
the diffusion coefficient for benzene as a
function of chemical exposure. Duplicate
samples were run in many cases to check the
reproducibility; it was always better than
5%. The tracer results for PVC (Fig. 6)
are in general agreement with those found
in the WVT work in PVC (Fig. 3). The ex-
posure to acid and base do little to the
diffusion coefficient, whereas the leach-
ing agents, phenol and xylene lower the
diffusion rate as the plasticizer is
leached-out and the PVC becomes stiffer.
The results for EPDM (Fig. 7) are in
exact agreement (with respect to relative
change) with those given by WVT (Fig. 4).
This agreement between the two methods for
PVC and EPDM is very encouraging, and
again gives confidence in the general mass
transport measurement approach to monitor
structural changes in the FMLs.
The CPE diffusion results, as a func-
tion of chemical exposure, are shown in
Figure 8. Agreement with the WVT results
for CPE (Fig. 5) is good for the acid and
base, but is very poor for phenol. Where-
as the WVT results showed CPE to become
approximately four times more permeable
after 6 months exposure to phenol, the
benzene diffusion rate shows essentially
no increase after 6 month's exposure to
phenol. All the phenol results on Figure
8 were duplicated within agreement of 5%.
This is quite an important (and possibly
discouraging) result, because the CPE/
phenol situation appeared from the WVT to
be a prime candidate for FML lifetime pre-
diction via transport measurements.
c) Water Absorption and Waste Vapor Absorp-
tion
As mentioned in the previous section,
there are significant problems in deter-
mining the diffusion coefficients in FMLs
via water absorption or water vapor ab-
sorption. The main problem with liquid
water absorption is that in many cases the
FML does not achieve a constant (equili-
brium) weight during a very long immersion
time. Many samples were monitored for 200
days with no equilibrium weight achieved.
The diffusion coefficient cannot be deter-
mined without this final weight. In spite
of this serious problem, many samples
showed very nearly the expected (Fickian)
316
-------
t^ power law in the initial stages of ab-
sorption.
In the case of water vapor absorption,
the main problem is the lack of an initial
t^ time dependence, even though a constant
weight is achieved in a reasonable time.
The diffusion is not simple Fickian and
hence calculation of a diffusion value
from the data is extremely suspect.
There are other problems in this mea-
surement technique:
. PVC absorbs a very small amount of
water, making analysis next to im-
possible.
. Some samples actually lose weight
while in contact with water, in-
dicating some of the chemical
gained during exposure is still
coming out of the FML (this even
happened in some vacuum-dessicated
samples).
. In many cases the data is scattered
and shows no definite trends.
The conclusion is that water and
water vapor absorption are not reliable
methods with which to determine the diffu-
sion coefficient of water in commercial
FMLs. Some others'4,2,3) who has worked
in this area of absorption measurement have
obtained the same type of results (and pro-
blems) as those found here, however their-
conclusions were not as negative, with re-
spect to the use of the absorption method,
as the present authors feel is warranted.
On the positive side, when the method
does work, it shows diffusion coefficients
consistently in the 10~9 cm2/sec range for
both the thermoplastic (CPE and PVC) and
thermoset (EPDM) types. However, it is not
possible to systematically follow the de-
gradation of FMLs via these absorption
type measurements. There are simply too
many problems with this technique as ap-
plied to these commercial materials.
CONCLUSIONS
The first goal of the present work
was to ascertain if mass transport mea-
surements could be used reliably to deter-
mine structural changes in FMLs. If this
was the case, then the second goal was to
determine if systematic trends developed
in the mass transport values as a function
of exposure to various liquid wastes.
If present, these trends would be used (the
final goal) to determine, probably by long
term extrapolation of the data, the life-
time of the FML in a particular waste.
Using a matrix of 3 FMLs and 4 chemi-
cals, the first two goals were definitely
realized. Two of the four mass transport
techniques used (WVT and RT) gave very re-
producible and apparently quite reliable
indication of changes observed from the
trends in the data. Furthermore in all but
one case the two methods agreed completely
as to the relative changes. The well known
resistance of most FMLs to acids and bases
has been amply demonstrated. The mechanical
stiffening of PVC due to loss of plasticizer
has been demonstrated via transport mea-
surements. In metallic and polymeric sys-
tems mass transport properties have been
found to be at least as sensitive (and in
some cases more sensitive) to structural
change as mechanical and physical proper-
ties.
The most interesting situation as far
as lifetime prediction is concerned is the
CPE-phenol case. Here a very significant
(about 4-fold) increase in the WVT takes
place with 6 month's exposure. If this
significant increase continues with expo-
sure time, this would be an ideal case in
which to determine a limiting upper value
of WVT, above which the FML would become un-
useable for containment. At this juncture
the RT results do not corroborate the WVT
for the CPE-phenol case, there being es-
sentially no change in the diffusion co-
efficient after 6 month's exposure.
Thus the indicated significant in-
crease in WVT for CPE exposed to phenol
must be investigated in greater detail, to
ascertain if it is a true effect, or if
the lack of an effect, as indicated in the
RT measurements, is closer to the true
structural effect. This situation should
be run again for complete verification of
the data, and possibly other experimental
techniques (e.g., microstructural observa-
tions, glass temperature measurements,
etc.) should be brought to bear on this
particular FML/chemical situation. Other
systems with similar chemistry should also
be investigated.
It is felt that the goals of the
study were achieved and a very definite
possibility exists for determining the
lifetime of FMLs through mass transport
317
-------
measurements on chemically-exposed samples.
ACKNOWLEDGEMENTS
The authors would like to thank the
U. S. Environmental Protection Agency's
Land Pollution Control Division, Hazardous
Waste Engineering Research Laboratory,
Cincinnati, Ohio for supporting this work.
Our project officer, Mr. Paul dePercin has
offered a great deal of encouragement, ad-
vice and help.
REFERENCES
1. ASTM Standard E-96-66, "Standard
Methods of Test for Water Vapor Trans-
mission of Materials in Sheet Form,"
American Society for Testing Materials
(ASTM), Philadelphia, PA.
2. Baker, G. R. and Thompson, C. M., "The
Effect of Seawater on Polymers," NRL
Memorandum Report 4097, Transducer
Branch, Underwater Sound Reference De-
tachment, Orlando, Florida, November,
1979.
3. Brzozowski, K. J. and Kumins, C. A.,
"Barrier Properties of Earth Lining
for Pollution Control," Proc. of Sym-
posium on Permeability of Plastic Films
and Coatings to Gases, Vapor and
Liquids, Los Angeles, April 1974,
Plenum Press, NY, pp. 321-329.
4. Carpenter, C. N. and Fisher, A. 0.,
"Sequential Chemical Absorption Tech-
niques for Evaluating Elastomers,"
Materials Performance, Vol. 20, No. 1,
1981, pp. 40-45.
5. Crank, J. and Park, G. S., Editors,
Diffusion in Polymers, Academic Press,
New York, 1968.
6. Joos, W., Diffusion in Solids, Liquids
and Gases, Academic Press, New York,
1960, p. 25.
7. Lord, A. E. and Koerner, R. M. , "Funda-
mental Aspects of Chemical Degradation
of Membranes," Proc. Intl. Conf. on
Geomembranes, Denver, Colorado, June
1984, pp. 293-298, Published by IFAI,
St. Paul, Minnesota.
8. Tub type exposures have been used in
the work by: Matrecon, Inc., "Lining
of Waste Impoundment of Disposal
Facilities," Report to U. S. EPA
Municipal Environmental Research Labo-
ratory, Solid and Hazardous Waste Re-
search Div., R. Landreth, Project Of-
ficer, Cincinnati, Ohio, 1982.
9. Southern, E. and Thomas, A. G., "Diffu-
sion of Liquids in Crosslinked Rub-
bers," Rubber Chemistry and Tech-
nology, 42, pp. 495-504 (1969). They
mention the theory was developed by J.
Crank, The Mathematics of Diffusion,
Oxford University Press, 1956.
10. Wood, J. J., "Diffusion Characteris-
tics of Flexible Membrane Liners,"
Master's Thesis, Drexel Univ., 1984.
TABLE 1. SUMMARY OF THE WVT AND RT RESULTS
FOR 6 MONTH EXPOSURE
PVC
H2S04
NaOH
Phenol
Xylene
EPDM
H2S04
NaOH
Phenol
Xylene
CPE
H2S04
NaOH
Phenol
Xyl ene
6 Month's Exposure
WVT
+ 12% - steady
+ 10% - steady
- 30% - steady
- 25% - steady
+138% - rising
+100% - steady
+ 70% - steady
-
+ 54% - steady
+ 46% - steady
+260% - rising
-
RT
+ 4% - steady
-11% - steady
-50% - steady
-85% - steady
+37% - rising
+20% - steady
+17% - steady
-
+15% - rising
-11% - steady
- 9% - steady
-
318
-------
to) Increasing Diffusion
(it., breakthrough)
(c) Constant Diffusion
/(ii., acceptable)
(b) Decreasing Diffusion
d« , blocking) '
2 3 A 5
EXPOSURE TIME (months)
TIME
FIG. 1. Various Possible Trends in Diffu-
sion Properties as a Result of
FML Exposure Time
FIG. 3. WVT Results for PVC as a Function
of Exposure Time to Various
Chemicals
GLASS-
GROUND— JzJpn
GI ACC r -E
JOINTS-^* kdJfL
t
*
t
£
t
?
t
-'•'RADIOACTIVE -
'.'.• LIQUID.- :•
rrmti ..,.===
GM
TUBE
1
t
t
*
&
(?
f>
2.0
1.5
"I ,n
qP1 t'°
^-
| 0.5
0
(
NoOH — v. H2SQ — v s*
^— a — .'"'
s*" ___^ "" ~ ~.~ .— ---^ ••
i/ xylene — ' phenol-^
|EPDM|
1123456
EXPOSURE TIME (months)
counter
FIG. 2. Schematic Diagram for Determining FIG' 4' KJctlSS^f Expo™ ?ime to
Diffusion Constant Using Radio- Various Chemicals
active Tracer*
319
-------
1 2 3 A 5 6
EXPOSURE TIME (months)
5 15
Ul
8
sl.O
NaOH
phenol-
—©
2 3 A 5
EXPOSURE TIME (months)
FIG. 5. WVT Results for CPE as a Function
of Exposure Time to Various
Chemicals
FIG. 7. Diffusion Coefficient of Radio-
actively Tagged Benzene (l^C) in
EPDM as a Function of Exposure
Time to Various Chemicals
\ xylene-v
Xj. » _ _»
V W
2 3 A 5 6
EXPOSURE TIME (months)
I8
|6
\ ^^.
\
H2S04-
NaOH-
2345
EXPOSURE TIME (months)
FIG. 6. Diffusion Coefficient of Radio-
actively Tagged Benzene ( C) in
PVC as a Function of Exposure
Time to Various Chemicals
FIG. 8. Diffusion Coefficient of Radio-
actively Tagged Benzene ("C) in
CPE as a Function of Exposure
Time to Various Chemicals
320
-------
DEVELOPMENT OF A TECHNIQUE FOR RETROFITTING
IMPOUNDMENTS WITH GEOMEMBRANES
David W.R. Shultz, Daren L. Laine, and John W. Cooper
Southwest Research Institute
San Antonio, Texas 78284
ABSTRACT
A technique has been tested for retrofitting in-service liquid impoundments with a
flexible membrane liner. Tests of the pull-through technique under simulated field condi-
tions were implemented at a 0.4 hectare (ha) (1 acre) impoundment test facility.
The liner material selected for these tests was 0.91 millimeter (mm) chlorosulfonated
polyethylene, which was constructed of two 0.4 mm layers with reinforcing scrim between
them. A specially-designed steel bar, a self-adjusting rope bridle, and a tow vehicle
were used to pull the sheets of liner material along the contour of the water-filled
impoundment.
These tests demonstrated the feasibility of using the technique to install large
sheets of liner material. The size and shape of the impoundment, liner material, and
weather conditions are factors that influence the successful application of the tech-
nique. The test set-up and results are described in this paper.
INTRODUCTION
This paper presents the results of
a project to develop liner retrofit tech-
niques and to demonstrate one or more
methods under large-scale conditions.
Three in-service retrofit techniques were
evaluated during the project. These tech-
niques were termed pull-through, pump-over,
and float-and-sink. The results of the
conceptual and model studies have been
previously reported (1), and will not be
included here* The pull-through technique
was selected for large-scale tests.
RESEARCH APPROACH AND FIELD ACTIVITIES
Large-Scale Study
The test impoundment could accommodate
liquid to a depth of 2 meters (m). This
facility was previously used as a test
impoundment that required the conventional
installation of a geomembrane. Tests of
the retrofit technique were conducted with
this previously installed liner in place.
Approach for Single Sheet Tests
Single sheet tests of the pull-through
technique were performed using one sheet of
liner material approximately 12.2 m by 64 m
in size. The material selected was 0.91 mm
chlorosulfonated polyethylene, which is
constructed of two 0.4-mm layers with rein-
forcing scrim between them. This material
was selected because of its high tensile
strength, availability, and its prevalent
use in the liner industry.
Figure 1 shows a conceptual drawing
of the single sheet test setup. Two steel
plates, eyebolts, a self-adjusting rope
bridle, and a tow vehicle were used to pull
the sheet lengthwise down the side slopes,
across the bottom, and up the opposite
slope. The leading edge of the sheet was
positioned at the impoundment berm. Two
6.1-m sections of flat steel plate hinged
together in the center were attached to the
edge of the sheet by wrapping the sheet
around the bar. The bar was equipped with
two nuts welded along the top of the bar at
321
-------
WEIGHTED
LEADING EDGE
FLAT STEEL
PLATE)
OJ
K5
IN LINE
DYNAMOMETER
TOW
VEHICLE
Figure 1. Conceptual drawing of the single sheet pull-through test setup.
-------
approximately 2-m intervals. After wrap-
ping the liner material around the bar,
holes were cut through the liner at each
attachment location and eyebolts were then
screwed into the nuts. A self-adjusting
bridle made of 12-mm polyethylene rope was
attached to the eyebolts. The rope bridle
was comprised of short segments of rope
connected by shackles. The shackles
allowed the segments of rope to adjust
position under tension. This design was
chosen to distribute the pulling load as
evenly as possible along the bar. The
other end of the bridle was attached to a
tow vehicle located on the opposite side of
the impoundment. A dynamometer was placed
in-line between the bridle and the vehicle
to measure the tension in the line.
Four tests were conducted using the
single sheet and two-section steel bar. In
three of the tests, the sheet was pulled
down the side slopes and across the bottom
of the impoundment. The objectives of
these tests were to determine: (1) the
forces required to pull the sheet and steel
plates across the impoundment; (2) how
well the steel bar distributed the pulling
forces along the sheet; and (3) whether or
not the weighted leading edge would keep
the membrane on the bottom of the pond.
The fourth trial involved pulling
the panel along the sloping side of the
impoundment, keeping the center of the bar
over the toe of the side slope. In this
configuration, half of the panel was lying
at at angle on the side slope, partially
out of the water. The other half of the
panel was lying flat on the bottom of the
pond. The objectives of this test were to
determine: (1) whether or not the panel
could be pulled along the berm without
tearing; and (2) whether or not the bar
would bend at the hinge and therefore
conform to the change in slope angle at
the toe.
To accomplish the pull along the toe
of the berm, the sheet was moved into posi-
tion so that its center and hinge of the
bar would be in line with the toe. The bar
was then moved into position. The bar was
attached to the sheet in the same manner as
before. The sheet was then pulled along
the side slope to the opposite side of the
impoundment.
Single Sheet Test Results
The maximum pulling forces for each
test as measured with the dynamometer were
as follows:
(1) Test 1 - 8,900 newton (N) (2000
Ib-force)
(2) Test 2 - 7,600 N (1700 Ib-force)
(3) Test 3 - 9,800 N (2200 Ib-force)
(4) Test 4 - 6,200 N (1400 Ib-force)
These values occurred when the bar was
pulled up the side slope. The force
required to pull the sheet along the berm
(Test 4) is the smallest of the four
results because approximately one half of
the sheet was sliding over the dry surface
of the existing liner where the friction
was less than that under water. For an
earthen impoundment, the friction along the
dry soil berm would likely be higher than
the wet subgrade below water level. Tests
of the tensile strength of the sheet under
field conditions were not performed, since
the minimum strength of the sheet at simi-
lar temperatures was available from the
manufacturer.
Approach for Double Sheet Test
A second large-scale test was per-
formed to determine the feasibility of
pulling wider sheets of liner material.
For this experiment, the liner material
previously used for the single sheet tests
was repositioned to be pulled broadside
into the test impoundment. A longer bar
similar to that used previously was
attached to the long edge of the sheet.
The sheet was then pulled down the slope
until the trailing edge reached the top of
the berm. A second sheet of the same size
was then seamed to the first sheet and
both sheets were pulled into the water
until the second sheet covered the side
slope. Figure 2 shows a conceptual draw-
ing of this experimental setup. The first
sheet is shown pulled into the water and
along the bottom of the impoundment. The
second sheet is shown bonded to the first
sheet. The entire double sheet is pulled
along the contour of the impoundment by
the weighted leading edge. Three ropes
were used to attach the bar and bridle to
the tow vehicle.
323
-------
-------
The bar used for this experiment con-
sisted of three rigid steel bar sections
joined with two hinged joints. The details
of the bar are illustrated in Figure 3.
The two flexible joints were designed to
allow the bar to follow the contour of the
impoundment. The flat bar cross-section
was selected to provide minimum rotational
stress on the liner material as it was
pulled along the bottom of the impoundment.
The bridles used for this test were
similar in design to that used for the sin-
gle sheet tests. However, they contained
many more rope segments due to the addi-
tional width of the leading edge and the
increased weight. One large bridle was
attached to the eyebolts along the 45.7-m
rigid center section of the bar. Two
smaller bridles were attached to the 9.14-m
end sections of the bar.so that the centers
of the bridles were in line with the flexi-
ble joints. Shackles were used to attach
the bridles to the bar and to hold the bri-
dle rope segments together in a triangular
pattern. Tow lines made of the same rope
were then attached to each of the three
bridles. The rope was rated at a breaking
strength of 22,300 N (5000 Ib-force). Each
line was run across the surface of the test
impoundment. This setup was shown previ-
ously in Figure 2. The two outside lines
were rigged through snatch blocks to the
tow vehicle. The center line was run
directly across the impoundment to the
tow vehicle.
Double Sheet Test Results
To begin the test, the slack was taken
out of the bridle and the bar was pulled
forward to straighten out the liner. Pull-
ing was stopped to inspect the sheet. A
tear was discovered just behind one of the
flexible joints. This failure occurred
just behind a factory seam. Possible
causes were: (1) unequal loading on the
sheet from the bar; (2) a defect in the
sheet as a result of manufacturing; or
(3) a small puncture or tear introduced in
previous tests using the same material.
There were no other failures anywhere along
the sheet. To remedy this failure, the
bridle was removed from the bar and the
sheet was then wrapped around the bar past
the failure point. The bridle was recon-
nected and the test pull was restarted.
The first sheet was pulled over the
berm of the existing liner and down the
side slope. The first sheet was pulled
into the water until the trailing edge
reached the top of the berm. At that posi-
tion, the tow vehicle was stopped to allow
the second sheet to be attached. The sec-
ond sheet, measuring 12.2 by 64 m, was
fan-folded and positioned along the first
sheet, maintaining a 0.15-m overlap with
the first sheet. The sheets were seamed
together using a bodied solvent adhesive
recommended by the manufacturer. The seam
was allowed to cure for two hours before
being pulled. According to one manufac-
turer of chlorosulfonated polyethylene, the
seam should have cured to approximately 60
percent of the final cure strength in two
hours. Since no failures occurred at this
seam, the bond strength was strong enough
to withstand the tensile loading of the
second sheet.
With the two sheets bonded together,
there were approximately 1,560 m^ of sheet
weighing 1,907 kilograms (kg). The bar
weighed approximately 7.55 kg per linear
meter, or 483 kg. The total mass of the
bar and sheet was approximately 2,400 kg
including hardware and the plates used at
each fixed joint.
After the pull was completed, the
sheet was examined for damage. One tear
was found, located approximately 0.3 m
behind the flexible joint. The tear was
similar to the first tear, occurring at
a factory seam. It was larger than the
first tear, approximately 0.3 m in length.
The maximum force required to pull
both sheets was approximately 22,200 N
(4990 Ib-force), which occurred at the
start of the test. After the starting
friction was overcome, the force dropped
to approximately 13,300 N (2990 Ib-force).
Stress Loads on Liner Material
The tensile load on the liner mate-
rial during the single and double sheet
tests can be estimated to provide a basis
for evaluation of the technique and its
application. During the single and double
sheet tests, loadings on the liner material
were not measured directly. Direct mea-
surement would have required the use of
strain gauges attached to the sheet, or
some similar technique. This effort was
beyond the scope of the project. However,
these loadings can be estimated by assuming
that the loads measured by the dynamometer
325
-------
WEIGHTED LEADING EDGE
HEX-NUTS WELDED TO
TOP OF PLATE
Figure 3. Details of the weighted leading edge bar used during the double sheet test.
-------
TYPICAL FLEXIBLE JOINT DETAIL
TOP VIEW
HEX-NUT
. 15 MX. 02 M THREADED
STEEL EYEBOLT
SIDE VIEW
STEEL HINGES
STEEL PLATE SECTION
7
Figure 3. Details of the weighted leading edge bar used during the double sheet test (Cont'd),
-------
TYPICAL FIXED JOINT DETAIL
TOP VIEW
N3
00
3 M OR 6 M x .1 M x .01 M FLAT STEEL
PLATE SECTIONS
/ 1 '
X
0 O
o
o q
/ \
i
o ° ' \/
o i X
GO /A
/
\ .08 M x .02 M
N TYPICAL HEX-
HEAD BOLTS
SIDE VIEW
BUTT JOINT
BETWEEN SECTIONS
,25 M x .1 M x .01 M FLAT STEEL
PLATE
-------
were uniformly applied entirely to the
liner material. This represents a worst
case analysis, because some percentage of
the dynamometer reading is attributable to
the weight of the bar and frictional forces
in the rope bridle system. These forces
are not imposed on the liner. For the
worst case, the loadings measured by the
dynamometer can be assumed to approximate
the average forces delivered to the liner
immediately behind the steel bar attached
to the material during each test pull.
For the single sheet tests, the
assumptions and data needed to estimate
liner tension loadings are as follows:
in tension. The manufacturer's data for
tensile strength of the material indicate
that the scrim by itself can withstand 131
N/cm tensile loading at approximately 49°C.
The test method used to measure this value
specifies a rate of pull, or length of time
the material is in tension.
Calculation of Theoretical Sheet Size
Given the calculations from the field
test data and the manufacturer's data, it
is possible to evaluate the loadings expe-
rienced in the field and to calculate the
theoretical sheet size that could be pulled
using the technique.
(1) All forces are equally distrib-
uted along the length of the
sheet immediately behind the bar;
(2) The length of the sheet is
12.2 m; and
(3) The maximum force measured during
the four separate single sheet
tests is 9,800 N.
The tension per linear centimeter of
liner material is calculated as follows:
= 9.800 N 1 m
1 12.2 m X 100 cm
= 8.03 N/cm
For the double sheet tests, the assumption"
of equal loading behind the bar is not com-
pletely valid, since a small tear occurred
at one point. However, to derive a load
comparable with the single sheet load cal-
culated above, the tension per linear cen-
timeter for the double sheet test is:
= 22,200 N 1 m
2 64 m X 100 cm
= 3.47 N/cm
A representative of the manufacturer of
the liner material indicated that there
are two factors which will influence the
loading that the material can withstand
(A. Peterson, J.P. Stevens Co., personal
communication, 1984). These are: (1) the
temperature of the sheet; and (2) the
length of time the material is in tension.
The increased temperature softens the
elastomer, reducing the tensile strength.
The longer the material is in tension, the
more the material will exhibit slow elonga-
tion. This implies that a relationship
exists between the strength of the liner
material and the rate that it is loaded
During the single sheet tests, no
liner failures occurred. The calculated
maximum uniform tensile load of 8.03 N/cm
is well below the expected maximum strength
of 131 N/cm. During this test, the bar
was rigid and flat across its length. The
force on the bar was probably distributed
equally to the liner. The sheet tempera-
ture when the maximum loading occurred was
approximately 32°C. Thus, the strength of
the sheet was above 131 N/cm, corresponding
to a sheet temperature of approximately
49°C. The bar was pulled at a constant
rate across the pond and stopped at the
opposite berm. The maximum dynamometer
reading occurred when the bar was pulled
up the opposite berm. The highest tension
in the sheet occurred just before it began
to move up the slope.
The successful application of the
pull-through technique for retrofitting
in-service impoundments will depend upon
many factors. The most important factors
are as follows:
(1) The tensile strength of the liner
material under the prevailing
site conditions;
(2) The variation in liner material
strength due to sheet temperature;
(3) The uniform application of pull-
ing forces along the entire
length of the bar;
(4) The geometry of the impoundment;
(5) The conditions of the bottom and
side slopes of the impoundment;
and
329
-------
(6) Wind and rain conditions during
installation of the liner.
The tensile strength of the liner
material is the most critical technical
factor associated with successfully apply-
ing the pull-through technique to any size
impoundment facility. The material must
withstand the mechanical stresses from
the bar as well as any loading due to the
weight of solids or liquids in the impound-
ment. The mechanical aspects of pulling
the sheet will not limit the size of facil-
ity that can be retrofitted. Any number of
winches can be operated together to move
the bar and liner at a constant rate.
The slope of the embankments will
influence the application of the technique.
The steeper the slope, the more force will
be required to pull the bar up the slope
and over the berm. The conditions of the
bottom and side slopes of the impoundment
are also important factors. Debris on the
bottom of the impoundment may tear the
liner or prevent the bar from moving prop-
erly along the bottom. Unstable, soft sub-
grade will not adequately support the liner
material after retrofit. All of these fac-
tors must be considered before initiating
a retrofit project using the pull-through
technique. As a minimum, the bar should
be initially pulled across the impoundment
without the liner to help identify any for-
eign objects that may damage the liner or
prevent the retrofit from being completed.
The proper liner material must also be
selected, based upon the required mechani-
cal properties for successful installation.
The theoretical limits of the pull-
through technique with respect to impound-
ment size can be computed assuming: (1)
equal loading on the sheet; (2) pulling
forces equal the weight of the sheet; and
(3) constant pulling rate along the bar.
The following example calculation will
illustrate the computation for a rectangu-
lar impoundment:
Given: (1) Rectangular impound-
ment with ends 76.2 m
in width;
(2) 0.91-mm chlorosulfo-
nated polyethylene
liner material with
minimum tensile
strength of 131 N/cra;
Find:
Solution:
Force
(3) Installation tempera-
ture of 21°C;
(4) Side slopes no steeper
than 3 to 1; and
(5) Mass of liner 1.2
kg/m2.
(1) The length and
surface area of the
impoundment that can
be theoretically
retrofitted using
the pull-through
technique.
(1) Compute the total ten-
sile force that can be
sustained by 76.2 m of
liner material:
131 N/cm x 76.2 m x 100 cm/m
998,220 N
(2) Compute the area of
liner material equal
to the load from (1)
above:
Area =
998,220 N
1.2 kg/m2
84,883 m2
(3) Compute the length and
surface area:
84,883 m2
76.2 m
1,114 m
84,883 m2
Surface Area
10,000 m2/ha
= 8.49 ha
This calculation suggests that, given
the assumptions listed above, a sheet of
liner material measuring 76.2 m wide by p
1,114 m long would be strong enough to ;
withstand pulling from one end. A piece of
material this size would be sufficient to
retrofit an impoundment having a surface
area of nearly 8.5 ha.
330
-------
This calculation takes into considera-
tion only the specified maximum tensile
strength of the material at 49°C, and does
not include a factor of safety. Other fac-
tors, such as the effect of sheet tempera-
ture on strength, the uniform application
of pulling forces, the impoundment geome-
try, and the condition of the bottom and
side slopes of the impoundment, must be
considered to estimate the limits of this
retrofit technique.
SUMMARY
The bar and bridle performed as
intended for the single-sheet tests. The
bar conformed to the contour of the
impoundment, keeping the sheet below the
liquid level. Bending of the bar occurred
because of variations in stretch of the
polypropylene rope and uneven slack in the
lines. This problem would be remedied by
attaching additional lines to the bar to
distribute the pulling forces more evenly
along the leading edge.
The large-scale tests demonstrated the
feasibility of the pull-through technique
for moving large sheets of liner material
along the contour of the test impoundment.
The double sheet, measuring 64 m wide by
24.4 m deep, was pulled a distance of
approximately 24.4 m. The force required
to initially move the sheet and bar was
approximately equal to the force of gravity
on the sheet and bar. The force required
to keep the sheet moving was approximately
13,300 N.
Weather conditions during the pull-
through test were unavoidably hot (ambient
temperature of approximately 38°C) and
sunny. These conditions caused heating of
the sheet, reducing its tensile strength.
The coefficient of friction between the
sheet and the ground was large because at
elevated temperature the sheet conforms
more closely to the irregularities of the
ground. These factors probably contributed
to the two sheet failures during the double
sheet test. Materials similar to chloro-
sulfonated polyethylene, such as polyvinyl
chloride and chlorinated polyethylene, will
also be softened under similar conditions.
Whenever these materials are used, the
application of this technique should be
limited to cooler temperature conditions.
The method used to attach the liner
material to the bar was satisfactory for
this demonstration. The sheet did not
unwrap from the bar at any point.
Estimation of the stress loading on
the liner material during the tests approx-
imate the weight of the material itself.
The stress loading per unit length of bar
thus depends on the length of the bar. For
example, the stress loading for the single
sheet test was 8.03 N/cm of bar width,
whereas it was 3.47 N/cm of bar width for
the double sheet test. A retrofit applica-
tion project can be designed to take advan-
tage of this fact so as not to exceed the
strength of the liner material chosen.
REFERENCES
1. Cooper, John W., and D. W. Shultz.
Development of Liner Retrofit Concepts
for Surface Impoundments. In: Pro-
ceedings of the Ninth Annual Research
Symposium on Land Disposal of Hazardous
Wastes. U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1983, p203.
331
-------
THE ROLE OF IMPOUNDMENT SIZE IN CONTROLLING
ENVIRONMENTAL RISK AT HAZARDOUS WASTE SURFACE IMPOUNDMENTS
Gordon B. Evans, Jr., Jey K. Jeyapalan, Joe Sai, Ronald Shiver, and David C. Anderson
K. W. Brown and Associates, Inc.
College Station, Texas 77840
ABSTRACT
The environmental damage due to a hazardous waste surface impoundment failure
is greatly affected by the impoundment size (volume). This research determined if
multiple-small impoundments, designed to contain the same waste volume as one large
impoundment, would present less environmental risk than the large impoundment.
Three surface impoundment design scenarios (large, small-attached, small-unattached)
were compared based on the environmental damage that would result from two failure
situations: surface (catastrophic) and subsurface (long-term). The reduced risk
from using smaller impoundments was then compared to the increase in construction
cost due to constructing more than one impoundment. It was concluded that there
are no real advantages to limiting the size of surface impoundments. Several design
features are suggested which should greatly reduce the risks associated with surface
impoundment failures.
INTRODUCTION
Basis for optimizing the design of
surface impoundments (SI) should invlude
consideration of the following: cost,
geotechnical and hydrogeologic factors, land
use, engineering design, environmental
performance and total waste volume. Since
many of these concerns are interrelated,
optimization SI design can be a complex
task. To help simplify the task of SI de-
sign optimization, this paper addresses the
following more general question:
"For a given desired SI volume,
what are the costs and environ-
mental risks associated with using
one large SI instead of multiple
small Si's."
An extensive literature review and inquiries
directed to SI experts indicated there had
been little previous study of this question.
Efforts to address the question focused on
the following aspects of surface
impoundments:
1) Design Scenarios;
2) Construction Costs; and
3) Surface and Subsurface SI Failure.
Mechanisms of failure are examined along
with the risk associated with each type of
surface failure for several impoundment
configurations (Figure 1).
IMPOUNDMENT DESIGN SCENARIO SELECTED FOR
STUDY
An EPA study (1983) found that almost
85% of industrial Si's are under 4 acres
(1.6 hectares) in surface area (Figure 2).
Consequently, the 1.6 hectare was selected
to represent a single large SI. To facili-
tate comparison, four impoundments were
selected to represent multiple small Si's.
Because multiple small Si's may be con-
structed with or without common dikes
(Figure 1) and because shared dikes may
increase the risk of multiple SI failures
(discussed later in this paper), both of
these were considered in this study. In
each case small Si's were assumed to contain
1/4 the volume of the one large SI, and all
Si's were assumed to have been constructed
by the excavate and fill method (ie., par-
tially above and partially below grade).
For an unbiased comparison of the three
configurations (one large, multiple small
332
-------
( a )
( b )
( c)
FIGURE I THE THREE SI CONFIGURATIONS
CONSIDERED
(a)LARGE Sl(b)4 SMALL INDEPENDENT SI
(c)4 SMALL SI JOINED BY COMMON DIKES
connected, and multiple small independent
Si's), the fill efficiency ratio concept was
used to establish dimensions. Since opti-
mizing construction cost would be one goal
of the operator, the most cost effective
dimensions for a given SI volume are those
with the highest ratio of storage volume to
fill volume. Therefore, the hypothetical
Si's were modeled and optimized for highest
fill efficiency ratio.
Survey work by Geraghty and Miller
(1980) indicated that the majority Si's are
sited near stream valleys, tidal estuaries,
or bays. The study also indicated that most
Si's are located in unconsolidated sediments
over near-surface groundwater that dis-
charges directly to surface water bodies.
Consequently, failure of a liner system
could cause rapid and extensive environmen-
tal contamination. Increased awareness of
the water quality threat posed by Si's is
reflected in the Hazardous and Solid Waste
Amendments of 1984 (HWSA). A new disposal
SI must now have a double liner system and
groundwater monitoring while most existing
Si's must be closed or upgraded to current
liner standards within four years of en-
actment of HWSA. The systems chosen for the
scenarios in this study are three with
double liners and one with a triple liner
(i.e., double flexible membrane liner -
clay) (Figure 3).
COMPARATIVE CONSTRUCTION COSTS
Replacing one large SI with multiple
small Si's has a major impact on the cost of
SI construction. The optimum dimensions of
the Si's in the scenarios evaluated in this
study were used to estimate generic engi-
neering and construction costs; unit cost
data were taken from current estimates by
the EPA (Craig, 1984). Table 1 lists the
costs for double flexible membrane lined
(FML) impoundments. Cost of four small Si's
would be nearly double that of one large SI.
The relative costs of each scenario holds
for the other three liner scenarios shown in
Figure 3.
SUBSURFACE FAILURES
All SI configurations examined in this
study had either double or triple liner
systems. Consequently, subsurface failures
are considered due to multiple liner
failures. While there are numerous types
and brands of liners, only generic failure
mechanisms for FML's and clay liners were
considered. Both liner types can be made to
leak excessively if they are incompatible
with the waste, improperly designed, im-
properly installed or misused. Excessive
lateral leakage can reduce the structural
stability of earthen dikes (see following
section on surface failures). Excessive
vertical leakage is primarily a risk due to
its potential for contaminating groundwater
and downgradient surface water.
In the case of Si's lined with FML's,
the smaller the SI, the greater will be the
length of seams per unit SI volume. For the
large SI used in this study, there would be
a total seam length of 2050 meters. Each
smaller SI would have a total seam length of
333
-------
100
AGRICULTURAL
MUNICIPAL
« INDUSTRY
* OIL a GAS
01 02
10 2.0 3.O 40 SO
SURFACE AREA (oc)
10.0
600
1000
FIGURE 2 SURFACE IMPOUNDMENT SIZES
Table 1. Cost Estimate Parameters for Double FML Liner System
Parameters
One Small
Large SI SI
Total Excavation Volume (M3) 12,286 4,365
Four Small
Si's
17,460
30 mil FML (MM
Geo textile (M2)
Layer 2
Sand (MJ)
Layer 3
30 mil (tf)
Layer 4
Geotextile (MM
Sand (M3)
Number of leachate Branches
Leachate Laterals (M)
Leachate Main (M)
16 ,384
11,621
4,839
16.384
16 ,384
2,420
8
862
128
6,053
3,969
1,771
6,053
6,053
886
5
315
78
25,079
15,876
7,078
25,079
Zb,07b
3,539
20
1,260
311
of leaks from the small Si's due to their
greater seam length if leakage is propor-
tional to total seam length. The result of
these two factors should be a net 28%
DOUBLE FML
FML - CLAY
-FML
-SAND
-FML
-SAND
-FML
-SAND
-CLAY
913 meters. Consequently, the four small
Si's would require 178% of the seam length
required for the one larger SI. Presuming
leaks are more likely through seams due to
inadequate quality control in the field
seaming than through the middle of an FML
sheet, there should be significantly fewer
leaks per unit volume of waste with the
larger SI. On the other hand, hydraulic
head (total potential liquid depth) in the
large and small Si's would be 320 cm and 230
cm respectively. Consequently, while there
is a 39% greater hydraulic head in the large
SI, there could be a 78% increase in number
DOUBLE CLAY
DOUBLE FML-CLAY
-SAND
-CLAY
-FML
-SAND
-FML
-CLAY
FIGURE 3. LINER SYSTEMS APPLICABLE TO
SURFACE IMPOUNDMENTS.
334
-------
greater leakage rate from the smaller Si's
that are lined with FML's.
Assuming that both size Si's are lined
with 60 cm of clay, the hydraulic gradients
in the large and small SI would be 6.3 and
4.8 respectively. .Considering only hy-
draulic gradient, the large SI should leak
at a rate 31% greater per unit area than the
small SI. The four small Si's, however,
have a total bottom surface area 37% greater
than that of the single large SI. Due to
this additional bottom area, the multiple
small clay lined Si's should have a total
leakage volume of 5% more than the single
large unit lined with clay.
SURFACE FAILURES
Performance of one large versus multi-
ple small Si's was evaluated assuming occur-
rence of the various surface failure mech-
anisms, broadly categorized as either cat-
astrophic dike failure or overtopping. An
implicit dynamic wave flood routing model
was used to compare resultant damages from
the different failure types for the three SI
design scenarios.
Catastrophic Dike Failures
Because of their dire environmental
consequences, catastrophic dike failures are
the major concerns for the performance of an
SI. Such catastrophic failures can be
caused by stability problems with the dike
or by overtopping (discussed in greater
detail below). Because little data exists
to substantiate the overall performance and
failure rate of Si's, this evaluation em-
phasized what is known about dike stability
analysis and the factors influencing stabi-
lity.
Stability of dikes is primarily a func-
tion of engineering properties of the
material and quality control exercised in
design and construction rather than cross-
sectional area. Stability analysis is based
upon soil cohesion, the angle of internal
friction, pore pressures, density, and geo-
metry of the dike material (Smith, 1982).
When pore pressure is low, the remaining
factors dictate the stability analysis
(Whitlow, 1983). If double or triple liner
systems are performing as designed, an in-
significant amount of liquids should pene-
trate the liners and enter the dike
materials. Therefore, when liners perform
properly, increases in pore pressure due to
liner leakage should be of little conse-
quence to the stability of a given cross-
sectional area of dike for small or large
Si's. In the case of four small independent
Si's there would be an apparent increased
risk of dike failure as compared to the
single large SI simply because of the addi-
tional external dike area.
When seepage from the liner system
occurs, the stability of the dike will be
reduced and may be threatened. With sub-
stantial leakage, material saturation and
the resultant increase in pore pressure can
cause a decrease in dike stability (Whitlow,
1983). Piezometers placed within the dike
can be used to monitor fluctuations in pore
pressure that may indicate a failure in the
liner system. Liner leaks also present a
problem with erosion of the dike beneath the
point of leakage and can increase the risk
of dike failure. In the case of dike
failure, environmental damages tend to in-
crease with the SI size. In any situation
where risk of dike failure is heightened
(e.g., earthquake zones or storage of waste
that may damage a liner), there are advanr
tages to using multiple smaller Si's.
In Si's with common dikes (Figure 1),
the risk of failure is also increased by the
rapid drawdown of liquid in an adjacent SI
(Smith, 1982). A sudden change in loading
of the adjoining dikes may place them in a
less stable condition. The presence of
double or triple liner systems should reduce
this likelihood of multiple SI failure.
However, if the first SI failure was due to
liner-waste incompatibility and the other
Si's have the same liners and waste, multi-
ple SI failures could occur.
Overtopping
Overtopping of an SI can occur as a
result of wind induced wave action or
failure of level control systems. Compari-
son of these two causes for overtopping
reveals that a level control device failure
is a much greater hazard because of the
rapidity with which it may result in cata-
strophic dike failure. Maximum liquid level
should be controlled by an outfall mechanism
that is relatively insensitive to inflow
(e.g., a weir or a spillway). Freeboard
(vertical distance from the top of the SI
embankment to the maximum liquid surface)
should be considered not as reserve storage
volume, but as wave protection.
If losses of liquid due to wave action
were considered alone, possible damages
335
-------
ensuing from such a discharge would be rela-
tively minor. The resultant dike softening
and wave cutting action may, however, cause
catastrophic failure. To account for the
risk of such major failures, current stan-
dards require that at least 2 feet of free-
board to be maintained (EPA, 1982). All
models developed thus far for determining
freeboard requirement are based on large
water bodies with fetches (i.e., un-
obstructed diagonal lengths) of greater than
1 mile. Zabcik and Kissock (1983) reviewed
these methods, and concluded that even the
largest SI considered in this study would
not be large enough to overtopping from wave
action alone where two feet of freeboard is
maintained.
As indicated previously, the use of
some level control mechanism is necessary to
maintain adequate freeboard in an SI. Since
level control is a mechanical feature of
Si's, the probability of failure of one of
the four small Si's is four times greater
than for one large SI. The degree of
failure by overtopping is a function of the
method employed for level control as follow:
1) automated or manually operated
controls; and
2) automated or manually operated
controls backed by a passive
control such as a spillway.
Failure of controls in the former case would
result in the loss of freeboard and in
potential overtopping or levee softening
that could lead to catastrophic dike
failure. In such an event, the potential
damage assessment would follow the same
approach as for other catastrophic failure
mechanisms (see the following section). If
failure occurred in the latter case, the
passive control system would allow a
limited discharge of contaminants but would
avert a catastrophic dike failure and mas-
sive waste release.
Resultant Damages
Failures can take two forms, catastro-
phic loss of all the above grade contents of
the SI or limited losses from overtopping.
Conditions leading to catastrophic failure
include slope failure of dikes, overtopping
and erosive cutting of dikes caused by level
control failure, and rapid drawdown slope
failure of one small impoundment in response
to failure of another SI that shares a com-
mon dike (Figure 1).
An implicit dynamic wave flood routing
model was developed to map potential inunda-
tion areas and to evaluate the relative
effects of large versus small SI failures.
This model transforms the partial differen-
tial equations of unsteady flow (Saint-
Venant, 1871) into a set of nonlinear alge-
braic equations which are solved simul-
taneously by the Newton-Raphson iterative
scheme. As a result of a breach in a dike,
a discharge hydrograph is produced and used
as the upstream boundary condition for solu-
tion of the model equations. The failure of
dikes is usually neither complete nor in-
stantaneous, with breach formation occurring
in a matter of minutes to a few hours and
with an average width of 1 to 3 times the
dam height. Breach shape is modeled as
rectangular, triangular, or trapezoidal.
The downstream boundary condition is deter-
mined according to a rating curve of flow
rates with corresponding depths of flow at
downstream nodes. Initial conditions con-
sisted of the base flow at the downstream
nodes. Downstream geometry must also be
specified to allow a solution of the un-
steady flow equations.
Parametric studies were run with the
model to determine the influence of the
various parameters on waste flow. It was
found that because breaches occur relatively
quickly in dikes, the resultant peak flows
occur within 0.5 to 3 hours. Rapid failure
is significant in that, regardless of SI
size, failures are likely to occur before
the operator can take mitigative actions.
Peak flows are roughly proportional to im-
pounded volume and peak discharge increases
with decreasing time of breach formation
(Figure 4). The model is, therefore, useful
for predicting peak flows to be used in the
design of secondary containment structures
to withstand the impact of the released
waste materials. The model presents a
strong case for the need for secondary con-
tainment structures around Si's which would
be analogous to containment berms around
tanks, since the effects of a failure of
smi5"1!] though1 Create?-1 Vea^o^"-
materials would undoubtedly occur in the
event of a large SI failure, any failure
without secondary containment would result
in widespread environmental damages.
The influence of impoundment size on
damages from overtopping is a function of
whether or not a spillway is a part of the
design. Where no spillway exists, over-
topping can lead to the kind of catastrophic
336
-------
500
400
300
200
100
—I r
BREACH TIME,
1*1.0 hour
WIDTH OF BREACH, BB = 25 ft
SLOPE = 5 ft/mile
4.0 hour
IMPOUNDMENT VOLUME, SA
: 36 acre-ft
9 acre-ft
DAM HEIGHT, H
10.5 ft
7.5 ft
I
0.0
0.5
1.0
1.5
2.0
TIME (hours)
2.5
3.0
3.5
FIGURE 4. VARIATION OF DISCHARGE FROM BREACH WITH
RATE OF BREACH FORMATION.
failure discussed in the previous sections.
For designs that include spillways, impound-
ment size has an effect only if the overflow
is from precipitation. In this event, the
volume of material released is a function of
SI surface area since area controls the
volume of precipitation caught within the
SI. If the only source of the flow is
incoming waste or wastewater, volume lost
over the spillway is independent of impound-
ment size.
CONCLUSIONS
The relative risks of one large versus
multiple small Si's are summarized as fol-
lows:
1) For functional liner systems,
there is no significant dif-
ference in failure probability
of a given unit length of dike
since, in such cases, failure
is primarily due to engineering
properties of the materials,
design and construction and not
dike size. There may be a
greater overall risk of failure
in the multiple SI case due to
the greater external dike
length, however, the extent of
resultant damage from
either SI size would be wide-
spread;
2) If major liner leakage does
occur, the depth of liquid im-
pounded increases the pro-
bability of dike failure.
Again, however, the resultant
damage from either case would
be substantial;
3) Only when overtopping caused by
precipitation occurs with a
spillway present does larger SI
size increase the quantity of
material lost; however, the
risk of failure is greater for
multiple Si's since a greater
number of level control
mechanisms are in place.
4) Subsurface leakage would pro-
bably be greater from multiple
small Si's than from a single
large SI due to differences in
total bottom area and, in the
case of FML's, seam length (per
unit area) if each is con-
structed with the highest fll-1
efficiency ratio. This is be-
cause optimal SI depth (and
hence hydraulic gradient) in-
creases with SI size.
337
-------
With the following design features
included, the risks of large or small SI can
be greatly reduced.
1) spillways for controlling da-
mages from uncontrolled re-
leases;
2) piezometers to monitor pore
pressures in the dikes;
3) secondary containment struc-
tures to limit contaminant
spread in the event of SI
failure;
4) for multiple small Si's separ-
ate independent structures
rather than common dikes.
Where these features are included in the SI
design, no real advantage is gained from the
use of multiple impoundments rather than one
large structure.
ACKNOWLEDGMENTS
This study has been supported by the
Land Pollution Control Division, Hazardous
Waste Engineering Research Laboratory,
Office of Research and Development, U. S.
EPA, Cincinnati, Ohio (Contract No. 68-03-
1816). The authors wish to thank Mr. Paul
dePercin, Project Officer, for his support
and guidance throughout this study.
REFERENCES
1. Craig, Jim, 1984. Personal Communcia-
tion.
2. EPA, 1982. Standards for Owners and
Operators of Hazardous Waste Treatment,
Storage, and Disposal Facilities.
Federal Register, 47:143, pp32274-32388.
3. EPA, 1983. Subsurface Impoundment
Assessment National Report. Office of
Drinking Water, U. S. Environmental
Protection Agency, Washington, D.C. EPA
570/9-84-002. 73pp.
4. Saint-Venant, B. De, 1871. Theory of
Unsteady Water Flow, With Application to
River Floods and to Progation of Tides
in River Channels. Acad. Sci. 73,
pp!48-240. (Translated into English by
U. S. Corps of Engineers, 1949).
5. Smith, G. N, 1982. Elements of Soil
Mechanics for Civil and Mining Engi-
neers. Granada Publishing Limited. New
York, NY. pp!42-162.
6. Whitlow, R, 1983. Basic Soil Mechanics.
Construction Press. pp282-299.
338
-------
UPDATE: STORAGE OF HAZARDOUS WASTE IN MINED SPACE
(Results of work conducted under U.S. Environmental
Protection Agency Contract #68-03-3191)
Ronald B. Stone
FENIX & SCISSON, INC.
1401 South Boulder
Tulsa, Oklahoma 74119
ABSTRACT
This report presents the results of an assessment of the current status of using
mined space for long-term retention of non-radioactive hazardous waste. The report is an
update of previous studies conducted from 1974-1977, and examines: recent literature
published on the subject, recent project activities, involvement of government agencies,
review of regulatory and permitting requirements, selection of existing mines for a
demonstration project and presents a new approach proposed by industry which is the use
of solution-mined space for retention of hazardous waste in salt.
INTRODUCTION
This paper presents highlights of the
study conducted under U.S. Environmental
Protection Agency Contract #68-03-3191
entitled, "Using Mined Space for Long-Term
Retention of Non-radioactive Hazardous
Waste."
With enactment of the 1976 Resource
Conservation and Recovery Act (RCRA), and
the hazardous waste regulations promul-
gated by the EPA, a national hazardous
waste management program exists.
The presently accepted methods of
disposal of hazardous waste are deep well
disposal, engineered landfills, land
treatment of hazardous waste and
incineration.
After all present treatment and stor-
age methods have been tried and found to
be lacking in one aspect or another, there
remains the need for an ultimate disposal
technique or method which can handle the
residues from the other methods.
Use of selected, existing mined space
appears to be, technically and economical-
ly, a feasible method of permanent storage
of untreatable wastes or the toxic end
products of hazardous waste treatment.
Among the advantages of mined space
for hazardous waste storage are:
• In deep mined space the waste would
be below drinking water aquifers.
• Isolation from the public and the
surface ecology.
t If required, waste can be isolated
from hydrological environment by
encapsulation or containerization.
• Security can be readily maintained.
• In a sealed mine, no continuing
maintenance will be required.
• If retrievability is desired, the
mine could be used as a long-term
underground warehouse.
339
-------
PURPOSE
The purpose of this project was to
update the state of the art of using mined
space for hazardous waste storage in the
U.S. A secondary purpose was to determine
the feasibility of a limited demonstration
project.
APPROACH
The approach included review of
activities conducted during the past ten
years to describe changes in the state-of-
the-art related to the use of mined space
for retention of hazardous waste. This
review was based on a literature search to
identify past and present research, gov-
ernment activities, activities associated
with the development of commercial facili-
ties, and foreign activities involved with
the use of underground space for storing/
disposing of wastes.
The approach also included a review
of federal and state regulatory require-
ments which would be involved in obtaining
a permit for an underground hazardous
waste disposal facility.
The selection of candidate mines was
based on a review of past government and
private client studies and reports. This
assessment indicated a number of conven-
tional mined spaces in limestone, salt,
potash and gypsum mines which might have
potential for hazardous waste storage.
The literature search indicated that
solution-mined storage caverns in salt
offer additional storage possibilities.
Literature Search: U.S. and Foreign
The first step in reviewing the
literature as to the status of non-radio-
active hazardous waste storage in mined
space was a data base literature search.
Publication citations were retrieved for
literature published since 1974.
Recent Activities: Foreign
Germany: The Herfa-Neurode facility
continues to operate satisfactorily 12
years after its start-up in 1972. Approx-
imately 270,000 tons of hazardous waste
were placed in the mine in ten years, and
the present annual volume is running
between 35,000 and 40,000 tons. Approxi-
mately 25% of this tonnage originates out-
side of Germany. The reuse of stored
waste is possible, and over 1,000 tons has
been retrieved by a waste producer and
returned to the originator for further
use.
India: Studies for storage of hazar-
dous waste in mined space have been con-
ducted for the central region of India as
reported by K.D. Ghosh of the Geological
Survey of India, in Rockstore, 1980.
Recent Commercial Activities - United
States
Louisiana: Empak, Inc. of Houston,
Texas applied to the Louisiana Department
of Conservation for a permit to build a
hazardous waste facility on the Vinton
Salt Dome in southwest Louisiana. This
project envisioned using solution-mined
caverns in the salt dome as final storage
for hazardous waste. After this project
was announced, a state law was passed
(1983) forbidding emplacement of hazardous
waste in salt domes for a period of two
years. This was intended to allow the
state time to evaluate the proposed use
prior to issuing a permit.
Texas: United Resource Recovery,
Inc. of Houston submitted an application
to the Texas Department of Wate |