&EPA
         United States
         Environmental Protection
         Agency
GEOSYNTEC
CONSULTANTS
ILLINOIS
                                             UNIVERSITY
EPA/600/R-02/099
December 2002
         Assessment and Recommendations
         for Improving the Performance of
         Waste Containment Systems
 3000

 2500

f: 2000 -

O 1500 -

 1000 -

  500 -

   0
                                 CO
                                           Active Period
                                           - of Operation -
                                       Initial Period
                                       of Operation
               IL
                                                    MSW Landfill
                                                    (Pennsylvania)
                                         OOOOOOOOOOOO
                                                      (0 -s ra
Rudolph Bonaparte, Ph.D., P.E.
GeoSyntec Consultants
Atlanta, GA 30342
         by

 David E. Daniel, Ph.D., P.E.
 University of Illinois
 Urbana, IL 61801
          Robert M. Koerner, Ph.D., P.E.
          Drexel University
          Philadelphia, PA 19104
                         performed under

                  EPA Cooperative Agreement Number
                         CR-821448-01-0
                          Project Officer

                        Mr. David A. Carson
               United States Environmental Protection Agency
                   Office of Research and Development
               National Risk Management Research Laboratory
                       Cincinnati, OH 45268

-------
                             DISCLAIMER

This  publication  was  developed  under  Cooperative  Agreement  Number
CR-821448-01-0 awarded by the United States Environmental Protection Agency
(EPA).  EPA made comments and  suggestions on the document intended to
improve the  scientific  analysis  and  technical accuracy of  the  document.
However,  the  views  expressed  in  this  document are those of GeoSyntec
Consultants, the University of Illinois,  and  Drexel University.  EPA does not
endorse any products or commercial services mentioned in this publication.

-------
                                  FOREWORD
The United States Environmental Protection Agency (EPA) is charged by Congress with
protecting the  Nation's land, air, and water resources.  Under a  mandate of national
environmental  laws, the Agency strives to formulate and implement actions leading to a
compatible balance between human activities and the ability of natural systems to support
and nurture life.  To meet this mandate, EPA's research program is providing data and
technical support for  solving environmental problems today  and building  a science
knowledge base necessary to manage our ecological resources wisely,  understand how
pollutants affect our health, and prevent or reduce environmental risks in the future.

The  National  Risk Management Research Laboratory is the Agency's center  for
investigation of technological and management approaches for preventing and reducing
risks from pollution that threatens human health and the environment. The focus of the
Laboratory's research program is on methods and their cost-effectiveness for prevention
and control of pollution to air, land, water, and subsurface resources; protection of water
quality in public water systems; remediation of contaminated sites, sediments and ground
water; prevention and control of indoor air pollution; and restoration of ecosystems. NRMRL
collaborates with both public and private sector partners to foster technologies that reduce
the cost of compliance and to anticipate emerging problems. NRMRL's research provides
solutions to environmental problems  by: developing and  promoting technologies that
protect and improve the environment; advancing scientific and engineering information to
support regulatory and policy decisions; and providing the technical support and information
transfer to  ensure  implementation  of environmental regulations and strategies at the
national, state, and community levels.

This publication has been produced as part of the Laboratory's strategic long-term research
plan. It is published and made available by EPA's Office of Research and Development to
assist the user community and to link researchers with their clients.
                                     E. Timothy Oppelt, Director
                                     National Risk Management Research Laboratory

-------
                                  ABSTRACT

This broad-based study addressed three categories of issues related to the design,
construction, and performance of waste containment systems used at landfills, surface
impoundments, and waste piles, and in the remediation of contaminated sites. The
categories of issues, the locations in this report where each category is addressed, and
the principal investigator for the study of each category are as follows:

   •  geosynthetic tasks are described in Chapter 2 and Appendices A and B; the
      principal investigator for these tasks was Professor Robert M. Koerner, P.E.;
   •  natural soil tasks are described in  Chapters 3 and 4 and Appendices C and D;
      the principal investigator for these tasks was Professor David E.  Daniel, P.E.;
      and
   •  field  performance tasks are described in Chapter 5  and Appendices E  and F;  the
      principal investigator for these tasks was Dr. Rudolph Bonaparte, P.E.

Each portion of the report was authored by the identified principal investigator, and
individuals working with the principal investigator.  However, each principal investigator
provided input and recommendations to the entire study and peer-reviewed and
contributed to the entire report.

Geosynthetic  materials (e.g., geomembranes (GMs), geotextiles (GTs), geonets (GNs),
and plastic  pipe) have been used as essential components of waste containment
systems since at least the early 1980's. Five separate laboratory and/or analytical tasks
were undertaken to address technical issues related to the use of these materials in
waste containment systems. The technical issues related to geosynthetics are: (1)
protection of GMs from puncture using needlepunched nonwoven GTs; (2) behavior of
waves in high density polyethylene (HOPE) GMs when subjected to overburden stress;
(3) plastic pipe stress-deformation behavior under high overburden stress; and (4)
service life  prediction of GTs and GMs. Conclusions are:  (1) needlepunched nonwoven
GTs can provide adequate protection of GMs against puncture by adjacent granular
soils; a design methodology for GM puncture protection was developed from  the results
of laboratory tests and  is presented; (2) temperature-induced waves (wrinkles) in GMs
do not disappear when the GM is subjected to overburden stress (i.e., when the GM is
covered with soil),  rather the wave height decreases somewhat, the width of the wave
decreases even more,  and the void space beneath the wave becomes smaller; (3)
waves may induce significant residual stresses in  GMs, which may reduce the GM's
service life; residual stresses induced in HOPE GMs by waves may be on the order of 1
to 22% of the  GM's short-term yield strength; (4) if GM waves after backfilling are to be
avoided, light-colored GMs can be used,  GMs can be deployed and seamed without
intentional slack, GMs can be covered with an overlying light colored temporary GT until
backfilling occurs, and backfilling can be performed only in the coolest part of the day or
even at night;  (5) based on finite element modeling results, use of the Iowa State
                                       IV

-------
formula for predicting plastic pipe deflection under high overburden stress is
reasonable; (6) polypropylene GTs are slightly more susceptible to ultraviolet (UV) light
degradation than polyester GTs, and lighter weight GTs degrade faster than heavier
GTs; (7) GTs that are partially degraded by UV light do not continue to degrade when
covered with soil, i.e., the degradation process is not auto-catalytic; (8) buried HOPE
GMs have an estimated service life that is measured in terms of at least hundreds of
years; the three stages of degradation and approximate associated durations for each
as obtained from the laboratory testing program  described in this report are: (i)
antioxidant depletion (« 200 years), (ii) induction (« 20 years), and (iii) half-life (50%
degradation) of an engineering property (« 750 years); these durations were obtained
from the extrapolation of a number of laboratory tests performed under a limited range
of conditions; it is recommended that additional testing be performed under a broader
range of conditions to develop additional insight into the ultimate service life of HOPE
GMs, and other types of  GMs as well.

Geosynthetic clay liners (GCLs) are a relatively new type of liner material, having first
been used in a landfill in  1986.  One of the key issues with respect to field performance
of GCLs is their stability on permanent slopes, such as found on landfill final cover
systems.  Fourteen test plots, designed to replicate typical final cover systems for solid
waste landfills, were constructed to evaluate the internal and interface shear strength of
GCLs under full-scale field conditions on 2H:1V and 3H:1V slopes. Five different types
of GCLs were evaluated, and performance was observed for over  four years. All test
plots were initially stable, but over time, as the bentonite in the GCLs became hydrated,
three slides (all on 2H:1V slopes) that involved the GCLs  have occurred.  One slide
involved an unreinforced GCL in which bentonite that was encased between two GMs
unexpectedly became hydrated.  The other two slides occurred at  the interface between
the woven GTs of the GCLs and the overlying textured HOPE GM. Conclusions are:
(1) at the low normal stresses associated with landfill final cover systems, the interface
shear strength is generally lower than the internal shear strength of internally-reinforced
GCLs;  (2) interfaces between a woven GT component of the GCL  and  the adjacent
material should always be evaluated for stability; these interfaces may  often be critical;
(3) significantly higher interface shear strengths were observed when the GT
component of a GCL in contact with a textured HOPE GM was a nonwoven GT, rather
than a woven GT; (4) if bentonite sandwiched between two GMs has access to water
(e.g., via penetrations or at exposed edges), water may spread laterally through waves
or wrinkles in the GM and hydrate the bentonite over a large area;  (5) if the bentonite
sandwiched between two GMs does not have access to water, it was found that the
bentonite did not hydrate over a large area; (6) current engineering procedures for
evaluating the stability of GCLs on slopes (based on laboratory direct shear tests and
limit-equilibrium methods of slope stability analysis) correctly predicted which test plots
would remain stable and which would undergo sliding, thus validating current design
practices; and (7) based  on the experiences of this study, landfill final cover systems
with 21-1:1 V sideslopes may be too steep to be stable with the desired factor of safety

-------
due to limitations with respect to the interface shear strengths of the currently available
geosynthetic products.

To evaluate the field performance of compacted clay liners (CCLs), a database of 89
large-scale field hydraulic conductivity tests was assembled and analyzed.  A separate
database for 12 soil-bentonite admixed CCLs was also assembled and analyzed.  In
addition, case histories on the field performance of CCLs in final cover test sections
were collected and evaluated. Conclusions are: (1) 25% of the 89 natural soil CCLs
failed to achieve the desired large-scale hydraulic conductivity of 1 x 10~7 cm/s or less;
(2) all of the 12 soil-bentonite admixed CCLs achieved a large-scale hydraulic
conductivity of less than 1 x 10~7 cm/s; however, all of these CCLs contained a relatively
large amount (more than 6%) of bentonite; soil-bentonite admixed CCLs will not be
discussed further; (3) the single most common problem in achieving the desired low
level of hydraulic conductivity in CCLs was failure to compact the soil in the zone of
moisture and dry density that will yield low hydraulic conductivity; (4) the most
significant control parameter of CCLs was found to be a parameter denoted "P0", which
represents the percentage of field-measured water content-density points that lie on or
above the line of optimums; when P0 was high (80% to 100%) nearly all the CCLs
achieved the desired field hydraulic conductivity, but when P0 was low (0 to 40%), fewer
than half the CCLs achieved the desired field hydraulic conductivity; (5) practically no
correlation was found between field hydraulic conductivity and frequently measured soil
characterization parameters, such as plasticity index and percentage of clay,  indicating
that CCLs can be successfully constructed with a relatively broad range of soil
materials;  (6) hydraulic conductivity decreased with increasing CCL thickness, up to a
thickness of about 1 m; and (7) analysis of CCLs constructed in the final cover test
sections generally showed that CCLs placed without a GM overlain by soil tended to
desiccate and lose their low hydraulic conductivity within a few years.

Liquids management data were evaluated for 187 double-lined cells at 54 landfills to
better understand the field performance of landfill primary liners, leachate generation
rates, and leachate chemistry. Conclusions are: (1) average monthly active-period leak
detection system (LDS) flow rates for cells with HOPE GM primary liners constructed
with construction quality assurance (CQA) (but without ponding tests or electrical leak
location surveys) will often be less than 50 Iphd, but occasionally in excess of 200 Iphd;
these flows are attributable primarily to liner leakage and, for cells with sand LDSs,
possibly construction water; (2) average monthly active-period LDS flow rates
attributable to leakage through GM/GCL primary liners constructed with CQA will often
be less than 2  Iphd, but occasionally in excess of 10 Iphd; (3) available data suggest
that average monthly active-period LDS flow rates attributable to  leakage through
GM/CCL and GM/GCL/CCL primary liners constructed with CQA are probably similar to
those for GM/GCL primary liners constructed with CQA; (4) GM liners can achieve true
hydraulic efficiencies in the 90 to 99% range, with higher efficiencies occasionally being
achievable; (5) GM/GCL, GM/CCL, and GM/GCL/CCL composite liners can achieve
                                       VI

-------
true hydraulic efficiencies of 99% to more than 99.9%; (6) GMs should not be used
alone in applications where a hydraulic efficiency above 90% must be reliably achieved,
even if a thorough CQA program is employed, except perhaps in situations where
electrical leak location surveys or ponding tests are used to identify GM defects and the
defects are repaired; (7) GM/CCL and GM/GCL/CCL composite liners are capable of
substantially preventing leachate migration over the entire period of significant leachate
generation for typical landfill operations scenarios without leachate recirculation or
disposal or liquid wastes of sludges; (8) leachate collection and removal system (LCRS)
flow rates were highest  at the beginning of cell operations and decreased as waste
thickness increased and daily and intermediate covers were applied to the waste;
leachate generation rates decreased on average by a factor of four within one year after
closure and by one order of magnitude two to four years after closure; within nine years
of closure, leachate generation rates were negligible for the landfill cells evaluated in
this study; (9) municipal solid waste (MSW) cells produced,  on average, less leachate
than industrial solid waste (ISW) and hazardous waste (HW) cells; for cells of a given
waste type, rainfall fractions were highest in the northeast and lowest in the west; the
differences in leachate generation rates are a function of type of waste, geographic
location, and operational practices; (10) in general, HW landfills produced the strongest
leachates and coal ash  landfills produced the weakest leachates; MSW ash leachate
was more mineralized than MSW leachate and the other ISW leachates;  (11) the solid
waste regulations of the 1980s and 1990s have resulted in the improved  quality of MSW
and HW landfill leachates; and (12) the EPA  Hydrologic Evaluation of Landfill
Performance (HELP) computer model, when applied using an appropriate simulation
methodology and an appropriate level of conservatism, provides a reasonable basis for
designing LCRSs and sizing leachate management system  components; due to the
complexity and variability of landfill systems,  however, the model will generally not be
adequate for use in a predictive or simulation mode, unless  calibration is  performed
using site-specific measured (not default) material properties and actual leachate
generation data.

Waste containment system problems were identified at 74 modern landfill and surface
impoundment facilities located throughout the U.S. The purpose of this aspect of the
project was to better understand the identified problems and to develop
recommendations to reduce the future occurrence of problems.  Conclusions are: (1)
the number of facilities with identified problems is relatively small in comparison to the
total  number of modern  facilities nationwide;  however, the search for problems was by
no means exhaustive; (2) the investigation focused on landfill facilities: 94% of the
identified problems described herein occurred at landfills; (3) among the landfill
problems, 70% were liner system related  and 30% were cover system related; however,
the ratio of liner system  problems to cover system problems is probably exaggerated by
the fact that a number of the facilities surveyed were active  and did not have a cover
system; (4) based on a waste containment system component or attribute criterion, the
identified problems can  be grouped into the following general categories: (i) slope
                                       VII

-------
instability of liner systems or cover systems or excessive deformation of these systems
(44%); (ii) defectively constructed liners, leachate collection and removal systems
(LCRSs) or LDSs, or cover systems (29%); (iii) degraded liners, LCRSs or LDSs, or
cover systems (18%); and (iv) malfunction of LCRSs or LDSs or operational problems
with these systems (9%);  (5) considering a principal human factor contributing to the
problem criterion, the identified problems are classified as follows:  (i) design (48%); (ii)
construction (38%); and (iii) operation (14%); (6) the main impacts of the problems
were: (i) interruption of facility construction and operation; (ii) increased maintenance;
and (iii) increased costs; (7) problems detected at facilities were typically remedied
before adverse environmental impacts occurred; (8) impact to groundwater or surface
water was only identified at one facility, where landfill gas migrated beyond the edge of
the liner system and to groundwater; (9) all of the identified problems can be prevented
using available design approaches, construction materials and procedures, and
operation practices;  (10) although the environmental impact of problems has generally
been negligible thus far, the landfill industry should do more to avoid future problems in
order to:  (i) reduce the potential risk of future environmental impact; (ii) reduce the
potential health and  safety risk to facility workers, visitors, and  neighbors; (iii) increase
public confidence in  the performance of waste containment systems; (iv) decrease
potential impacts to construction, operation, and maintenance; and (v) reduce costs
associated with the investigation and repair of problems.
                                       VIM

-------
                     ACRONYMS AND ABBREVIATIONS
ALCD       Alternative Landfill Cover Demonstration
ALR        action leakage rate
AOS        apparent opening size (of geotextile)
ARAR       applicable or relevant and appropriate requirements
ASTM       American Society for Testing and Materials
AZ          acceptable zone
BAT        commercial term for a type of porous probe
BNA        base neutral extractable
BOD        biological oxygen demand
BTEX       benzene, toluene, ethylbenzene, and xylenes
BuRec      U.S. Bureau of Reclamation
C&DW      construction and demolition waste
CAT        Caterpillar construction equipment
CCL        compacted clay liner
CERCLA    Comprehensive Environmental Response, Compensation, and
            Liability Act (aka Superfund Act)
CFR        U.S. Code of Federal Regulations
CH          soil classification symbol for a high plasticity clay soil
CL          soil classification symbol for a low plasticity clay soil
COD        chemical oxygen demand
CQA        construction quality assurance
CQC        construction quality control
CSPE       chlorosulfonated polyethylene
DSC        differential scanning calorimeter
EPA        U.S. Environmental Protection Agency
EPDM       ethylene propylene diene monomer
ET          evapotranspiration
FDEP       Florida Department of Environmental Protection
FEM        finite element model
fPP         flexible polypropylene
                                     IX

-------
FOS        filtration opening size (of geotextile)
FS          factor of safety
GC         geocomposite
GCL        geosynthetic clay liner
GDL        geocomposite drainage layer
GEC        geosynthetic erosion control (material)
GM         GM
GN         geonet
GT         geotextile
HOPE       high density polyethylene
HELP       Hydrologic Evaluation of Landfill Performance (computer program)
HLR        high level radioactive (waste)
HP-OIT     high-pressure oxidative induction time
HSWA      Hazardous and Solid Waste Amendments
H/W        height/width ratio (of GM waves)
HW         hazardous waste
ISW        industrial solid waste
k           hydraulic conductivity
kfieid         hydraulic conductivity measured in the field
kiab         hydraulic conductivity measured in the laboratory
LCRS       leachate collection and removal system
LDLPE      low density linear polyethylene
LDR        land disposal restrictions
LDS        leak detection system
LL          liquid limit
LLDPE      linear low density polyethylene
LLRM       low level radioactive mixed (waste)
LLR        low level radioactive (waste)
LMDPE     linear medium density polyethylene
Iphd        liters/hectare/day (1.0 Iphd = 9.35 gallon/acre/day (gpad))
LYS        lysimeter

-------
MCL        maximum containment level
MF         modification factor
MP         modified Proctor (compaction test)
MSW       municipal solid waste
NCP        National Contingency Plan
NE         northeast
NW         nonwoven (geotextile)
OD         outside diameter
OH         original height (of GM waves)
OIT         oxidative induction time
OWC       optimum water content
PCB        polychlorinated biphenyl
PCDD       polychlorinated dibenzo-p-dioxins
PCDF       polychlorinated dibenzo-furans
PE         polyethylene
PET        polyester
PI          plasticity index
PP         polypropylene
PPL        priority pollutant list
PVC        polyvinyl chloride
QA         quality assurance
RC         relative compaction
RCRA       Resource Conservation and Recovery Act
RF         reduction factor
RP         reduced Proctor (compaction test)
SARA       Superfund Amendments and Reauthorization Act
SC         soil classification symbol for a sandy clay
SDR        standard dimension ratio (of pipe)
SDRI       sealed double ring infiltrometer
SE         southeast
SMCL       secondary maximum containment level
                                     XI

-------
SP          standard Proctor (compaction test)
Std-OIT     standard oxidative induction time
SVOC       semivolatile organic compound
TCLP       toxicity characteristics leaching procedure
IDS        total dissolved solids
TOC        total organic carbon
TSB        two-stage borehole test
TSCA       Toxic Substances Control Act
TSDF       treatment, storage and disposal facility
TSS        total suspended solids
UMTRCA    Uranium Mill Tailings Radiation Control Act
UV          ultraviolet
VFPE       very flexible polyethylene (includes LLDPE, LDLPE and VLDPE)
VLDPE      very low density polyethylene
VOC        volatile organic compound
W          west
                                     XII

-------
                            ACKNOWLEDGEMENTS

The authors express appreciation to the U.S. Environmental Protection Agency for the
funding that enabled this project to be performed. In particular, the authors recognize
Robert E. Landreth (retired) for initiating the project and David A. Carson for providing
the continuity throughout the activity.

The project involved tasks divided into three broad categories: (i) geosynthetics; (ii)
natural soils; and (iii) field performance. Professor Robert M. Koerner, P.E. was the
principal investigator (PI) for the geosynthetics tasks.  Professor David E. Daniel, P.E.
was the PI for the natural soils tasks. Dr.  Rudolph Bonaparte, P.E. was the PI for the
field performance tasks.  The Pis and the members of each Pi's project team authored
their respective sections of the report. However, each PI provided input to the entire
study and each provided peer-review and contributions to the entire project.

With respect to the geosynthetics tasks, the financial contributions made by the
Geosynthetic Research Institute (GRI) and its membership is recognized and
appreciated.  Several of the geosynthetic tasks described in this report are ongoing
under the continuing financial support of the GRI. Y. (Grace) Hsuan, George R.
Koerner, and  Te-Yang Soong were involved in many of the  individual geosynthetic
tasks, including preparation of Appendices A and B.  Marilyn Ashley typed Chapters 1
and 2, and Appendices A and B. She also capably assembled all of the information
during the entire project.  Her work is sincerely appreciated  by the entire team of
principal investigators.

With respect to the natural soils tasks, numerous organizations assisted with the
construction of the geosynthetic clay liner (GCL) field test plots in Cincinnati.
Supplemental financial support was provided by CETCO, Claymax Corp. (now CETCO),
Gundle (now GSE) Lining Systems, and National Seal Co.  (now Serrot International).
Fluid Systems, Inc. (now Serrot International), provided the  geonet/GT drainage
geocomposites material,  Akzo (now Colbond),  Synthetic Industries, and Tensar
provided erosion control materials. Waste Management of North America and the staff
of the Elda  Landfill in Cincinnati, Ohio provided space at the Elda Landfill for the test
plots and provided personnel and equipment to assist with construction and
maintenance.  James Anderson, John Bowders, David Bower, Richard Carriker, Mark
Cadwallader,  Ted Dzierzbicki, Richard Erickson, John Fuller, George Koerner, Larry
Lydick, Majdi  Othman, Heather Scranton, John Stark, Fred Struve, and Robert Trauger
made major contributions to the program.

A database on performance of CCLs was assembled from published literature and from
unpublished data obtained from the files of Craig H.  Benson and Gordon P. Boutwell.
Stephen J. Trautwein provided information on where in-situ  hydraulic conductivity tests
had been performed in CCLs.  John J. Bowders assisted with the development of the
                                       XIII

-------
database and analysis of data. Owen Michaelis also helped with assembling data and
preliminary analysis of the database.

With respect to the field performance tasks, the task manager for Appendix E was Majdi
A. Othman. The authors of Appendix E were Majdi Othman, Rudolph Bonaparte, Beth
A. Gross, and Dave Warren, all of GeoSyntec Consultants. The authors would like to
acknowledge the following  landfill owners and operators and state regulatory agencies
for providing the data and information presented in Appendix E on the performance of
waste containment facilities:  Broward County (FL) Office of Environmental Services;
Browning-Ferris Industries; Camp Dresses & McKee, Inc.; Cape May County (NJ)
Municipal Utilities Authority; Chemical Waste Management; Chester County (PA) Solid
Waste Authority; Cumberland County (NJ) Improvement Authority; Indiana Department
of Environmental Management; Laidlaw Waste  Systems; Michigan Department of
Natural Resources;  New Jersey Department of  Environmental Protection and Energy;
New York State Department of Environmental Conservation; New York State Electric
and Gas Corporation;  Phillips Petroleum Company; Rollins Environmental Services, Inc;
USA Waste Services,  Inc.;  Utah Department of Environmental Quality; Waste
Management, Inc.; WM J. Huff Engineering; and others.

The task manager for Appendix F was Beth A. Gross. The authors of Appendix F are
Beth A. Gross, J. P. Giroud, and Rudolph Bonaparte, all of GeoSyntec Consultants.
The authors would like to acknowledge the following individuals and organizations for
providing information on some of the waste containment system problems described
herein: California Regional Water Quality Control Board (Greg Vaughn); Florida
Department of Environmental Protection (Kathy Anderson,  Robert Butera, Joe Lurix,
Jack McMelty, Susan Pelz, Richard Tedder);  New York State Department of
Environmental Conservation (Papa Chan Daniel, Matthew Eapen, Robert Phaneuf,
Thomas Reynolds, Melissa Treers); Ohio Environmental Protection Agency (Doug
Evans, Chuck Hull, Virginia Wilson); Metropolitan Dade County Department of Solid
Waste (Lee Casey); Yolo County, California (Ramin Yazdani); GeoSyntec Consultants
(Jeff Dunn, Scott Luettich, Neven Matasovic,  Bert Palmer, Jeffrey Palutis, Dennis
Vander Linde); Geosynthetic Research Institute (George Koerner, Robert Koerner);
Colder Construction Services, Inc. (Frank Adams); Hazen and Sawyer (John Bove); I-
CORP International, Inc. (Ian Peggs); Strata Systems (John Paulson); and others.
                                     XIV

-------
                                  CONTENTS

Disclaimer                                                                 ii

Foreword                                                                  iii

Abstract                                                                   iv

Acronyms and Abbreviations                                                ix

Acknowledgements                                                         xiii

Contents                                                                  xv

Chapter 1  Introduction

1.1  Goals of Waste Containment                                            1-1
1.2  Regulations                                                            1-4
1.3  Waste Containment System Components                                 1-9
1.4  Liner System and Final Cover System Components                        1-10

    1.4.1  Liner/Barrier Materials                                            1-10

          1.4.1.1   Compacted Clay Liners                                  1-12
          1.4.1.2   Geomembranes                                         1-13
          1.4.1.3   Geosynthetic Clay Liners                                 1-14
          1.4.1.4   Composite Liners                                       1-16

    1.4.2  Drainage Materials                                               1-18

          1.4.2.1   Granular Soils                                           1-19
          1.4.2.2   Geosynthetics                                           1-21

    1.4.3  Filtration Materials                                                1-23

          1.4.3.1   Granular Soils                                           1-24
          1.4.3.2   Geotextiles                                             1-24

    1.4.4  Ancillary Materials and Components                                1-26

          1.4.4.1   Plastic Pipe (aka Geopipe)                               1 -26
          1.4.4.2   GM Protection                                           1-27
          1.4.4.3   Erosion Control                                         1-27

1.5  Issues Evaluated in This Study                                           1-27
                                       XV

-------
    1.5.1  Geosynthetic Materials Tasks                                      1-28

          1.5.1.1   Puncture Protection of GMs                              1 -28
          1.5.1.2   Wave Behavior in HOPE GMs                            1-28
          1.5.1.3   Plastic Pipe Behavior Under High Overburden Stresses      1 -29
          1.5.1.4   Prediction of GT Service Life                             1-29
          1.5.1.5   Prediction of GM Service Life                             1-30

    1.5.2  Natural Materials Tasks                                           1-30

          1.5.2.1   GCL Test Plots in Cincinnati, Ohio                        1 -30
          1.5.2.2   CCL Test Pad Analysis                                  1 -31
          1.5.2.3   Admixed Liners                                         1-31
          1.5.2.4   CCLs in Final Covers                                    1-31

    1.5.3  Field Performance Tasks                                          1-32

          1.5.3.1   Review of Published Information                          1-32
          1.5.3.2   Data Collection and Analysis                             1-32
          1.5.3.3   Assessment of Problem Facilities                         1-32
          1.5.3.4   Comparison of Actual and HELP Model Predicted
                   LCRS Flow Rates                                       1-33

1.6 References                                                            1-33

Chapter 2 Geosynthetic Tasks

2.1 Puncture Protection of GMs                                             2-1

    2.1.1  Overview                                                        2-1
    2.1.2  Theoretical Aspects of GM Puncture                                2-2
    2.1.3  Experimental Aspects of GM Puncture                              2-4
    2.1.4  Puncture Protection Design Methodology                           2-8
    2.1.5  Examples                                                       2-8

2.2 Wave Behavior in GMs                                                 2-8
    2.2.1  Large-Scale  Experiments                                          2-9
    2.2.2  Small Scale Experiments and Results                              2-9
    2.2.3  Data Extrapolation and Analysis                                    2-12
    2.2.4  Discussion                                                      2-13

2.3 Plastic Pipe Behavior Under High Vertical Stresses                        2-13

    2.3.1  Leachate Removal Configurations                                  2-14
    2.3.2  Characteristics of Plastic Pipe                                      2-15
    2.3.3  Design by the Iowa State Formula                                  2-17
                                      XVI

-------
    2.3.4  Design by Finite Element Model                                   2-19
    2.3.5  Comparison of Design Methods                                   2-22

2.4 Prediction of GT Service Lifetime                                         2-24

    2.4.1  Behavior of Partially Ultraviolet Degraded GTs (PP and PET)         2-25
    2.4.2  Oxidative Degradation of PP GT Yarns and PE Geogrid Ribs         2-28
    2.4.3  Hydrolytic Degradation of PET GT Yarns                           2-28

2.5 Prediction of GM Service Lifetime                                        2-31

    2.5.1  Degradation of HOPE  GMs                                        2-31
    2.5.2  Simulated Applications                                           2-34
    2.5.3  Antioxidant Depletion Time                                        2-35
    2.5.4  Induction Time                                                   2-38
    2.5.5  Halflife of Engineering Properties                                  2-39
    2.5.6  Summary of Lifetime Prediction                                    2-41

2.6 References                                                            2-42

Chapter 3 Slope Stability of Full-Scale Test Plots Containing Geosynthetic Clay
          Liners to Simulate Final Cover Systems

3.1 Introduction                                                            3-1
3.2 Background on GCLs                                                   3-3

    3.2.1  Introduction                                                      3-3
    3.2.2  Advantages  and Disadvantages of GCLs                           3-5
    3.2.3  Shear Strength of GCLs                                          3-6

          3.2.3.1   Magnitude of Normal Stress                              3-6
          3.2.3.2   Water Content                                         3-8
          3.2.3.3   Type of Hydrating Liquid                                 3-8
          3.2.3.4   Rate of Loading                                         3-8
          3.2.3.5   Reinforcement                                         3-8
          3.2.3.6   Amount of Deformation                                  3-9
          3.2.3.7   Seismic Loading                                        3-10

    3.2.4  Interface Shear Strength                                          3-10

3.3 Field Test Plots                                                        3-10

    3.3.1  Rationale for 2H: 1V and 3H: 1V Slopes                             3-13
    3.3.2  GCLs                                                           3-13
    3.3.3  Other Materials                                                  3-15
    3.3.4  Construction                                                    3-15
                                       XVII

-------
3.4 Instrumentation                                                        3-21

    3.4.1   Moisture Sensors                                                3-21
    3.4.2   Displacement Gauges                                            3-22

3.5 Laboratory Direct Shear Tests                                            3-23
3.6 Performance of Test Plots                                               3-26

    3.6.1   Construction Displacements                                      3-26
    3.6.2   Post-Construction Performance of 3H:1V Slopes                    3-28

           3.6.2.1   Test Plot A (Bentonite Between Two GMs)                 3-28
           3.6.2.2   Test Plots B, C, and D (GT-Encased GCLs)                3-28
           3.6.2.3   Test Plot E (Unreinforced GCL)                           3-30

    3.6.3   Post-Construction Performance of 2H: 1V Plots                      3-31

           3.6.3.1   Test Plots G and H                                      3-32
           3.6.3.2   Test Plots F and P (Bentonite Encased Between Two
                   GMs)                                                  3-34
           3.6.3.3   Plots I and N with Nonwoven GT Component
                   Facing Upward                                         3-36
           3.6.3.4   Plots J, K, and L with No GM                             3-37

    3.6.4   Comments on Adequacy of Current Engineering Practice             3-37

3.7 Erosion Control Materials                                                3-39
3.8 Summary and Conclusions                                              3-40
3.9 References                                                            3-41

Chapter 4 Summary of Natural  Materials Tasks

4.1 CCLs Constructed from Natural Soil Liner Material                         4-1

    4.1.1   Introduction                                                     4-1
    4.1.2   Database                                                       4-2

           4.1.2.1   Source of  Data                                         4-2
           4.1.2.2   The Database                                          4-3
           4.1.2.3   Field Hydraulic Conductivity                              4-14

    4.1.3   Hydraulic Conductivity Results                                    4-15

           4.1.3.1   Field Hydraulic Conductivity                              4-15
           4.1.3.2   Hydraulic Conductivity from Small-Diameter Samples       4-18
                                      XVIII

-------
    4.1.4   Soil Characteristics                                               4-20

           4.1.4.1  Liquid Limit (LL)                                         4-20
           4.1.4.2  Plasticity Index (PI)                                      4-21
           4.1.4.3  Percent Fines                                           4-23
           4.1.4.4  Clay Fraction                                            4-25

    4.1.5   Compaction Conditions                                           4-26
    4.1.6   Construction Parameters                                         4-33
    4.1.7   Thickness of Liner                                                4-39
    4.1.8   Field Hydraulic Conductivity Testing Method                        4-41
    4.1.9   Case Histories                                                   4-42

           4.1.9.1  Test Pads at Sites 26 and 27                             4-43
           4.1.9.2  Test Pad at Site 21                                      4-44
           4.1.9.3  Test Pads at Sites 55-63                                 4-45
           4.1.9.4  Test Pads at Sites 64 and 65                             4-45
           4.1.9.5  Test Pads at Sites 43 and 44                             4-46

    4.1.10  Practical Findings from Database                                  4-46

4.2 Soil Bentonite Mixtures                                                  4-47

    4.2.1   Database                                                        4-47
    4.2.2   Hydraulic Conductivity Results                                    4-49
    4.2.3   Conclusions                                                     4-52

4.3 Compacted Clays in Final Cover Systems                                 4-52

    4.3.1   Omega Hills Final Cover Test Plots                                4-53
    4.3.2   Test Plots in Kettleman City, California                             4-57
    4.3.3   Test Plots in Hamburg, Germany                                  4-59
    4.3.4   Final Covers in Maine                                             4-62

           4.3.4.1  Cumberland Site                                        4-62
           4.3.4.2  Vassalboro Site                                         4-63
           4.3.4.3  Yarmough  Site                                           4-63
           4.3.4.4  Waldoboro Site                                         4-63
           4.3.4.5  Discussion                                              4-64

    4.3.5   Alternative Cover Demonstration at Sandia National Laboratory       4-64
    4.3.6   Test Covers in East  Wenatchee, Washington                       4-66
    4.3.7   Test Covers at Los Alamos National Laboratory                     4-67
    4.3.8   Other Studies                                                    4-68
                                       XIX

-------
4.4 Summary and Conclusions                                              4-68
4.5 References                                                            4-70

Chapter 5 Detailed Summary of Field Performance Tasks

5.1 Introduction                                                            5-1

    5.1.1   Scope of Work                                                  5-1
    5.1.2   Terminology                                                    5-2
    5.1.3   Data Collection Methodology                                      5-4

5.2 Evaluation of Liquids Management Data for Double-Lined Landfills           5-4

    5.2.1   Scope of Work                                                  5-4
    5.2.2   Description of Database                                          5-5
    5.2.3   Data Interpretation                                               5-6

           5.2.3.1   Landfill Development Stages                             5-6
           5.2.3.2   Primary Liner Leakage Rates and Hydraulic Efficiencies     5-7
           5.2.3.3   Leachate Generation Rates                              5-10
           5.2.3.4   Leachate Chemistry                                     5-11

    5.2.4   Evaluation Results                                               5-11

           5.2.4.1   Primary Liner Leakage Rates and Hydraulic Efficiencies     5-11
           5.2.4.2   Leachate Generation Rates                              5-32
           5.2.4.3   Leachate Chemistry                                     5-38

5.3 Lessons Learned from Waste Containment System Problems at Landfills     5-44

    5.3.1   Scope of Work                                                  5-44
    5.3.2   Description of Database                                          5-45
    5.3.3   Study Findings                                                  5-50
    5.3.4   Recommendations                                               5-56

5.4 Assessment of EPA HELP Model Using Leachate Generation Data          5-68

    5.4.1   Introduction                                                     5-68
    5.4.2   Description of HELP Model                                       5-69
    5.4.3   Literature Review                                                5-72
    5.4.4   Evaluation of HELP Model                                        5-74
    5.4.5   Study Findings                                                  5-76

5.5 References                                                            5-86
                                       xx

-------
Chapter 6 Summary and Recommendations

6.1  Rationale and Scope of Chapter                                         6-1

    6.1.1   Geosynthetics                                                  6-1
    6.1.2   Natural Soils                                                   6-3
    6.1.3   Field Performance                                              6-4

6.2  Liner Systems                                                         6-6

    6.2.1   Construction Quality Assurance                                  6-7
    6.2.2   Liner System Stability                                           6-7
    6.2.3   Waste Stability                                                 6-8
    6.2.4   Performance of Composite Liner                                 6-9
    6.2.5   Single vs. Double Liner System                                   6-10
    6.2.6   Fate of Liner Systems                                           6-12

6.3  Liquids Management                                                   6-13

    6.3.1   Construction Quality Assurance                                  6-13
    6.3.2   Potential for Clogging and Reduction of Flow Capacity              6-14
    6.3.3   Perched Leachate                                              6-15
    6.3.4   Fate of Liquids Management Systems                             6-16

6.4  Final Cover Systems                                                   6-16

    6.4.1   Construction Quality Assurance                                  6-17
    6.4.2   Compacted Clay Barriers                                        6-18
    6.4.3   Final Cover System Stability                                     6-18
    6.4.4   Cover Soil Erosion                                              6-19
    6.4.5   Fate of Final Cover Systems                                     6-20

6.5  Gas Management                                                     6-21

    6.5.1   Construction Quality Assurance                                  6-23
    6.5.2   Gas Uplift                                                      6-23
    6.5.3   Landfill Settlement                                              6-24
    6.5.4   Landfill Fires                                                   6-25
    6.5.5   Fate of Gas Management Systems                               6-26

6.6  Long-Term Landfill Management                                        6-26

6.7  References                                                           6-27
                                      XXI

-------
Appendix A - Behavior of Waves in High Density Polyethylene Geomembranes

A-1  Overview and Focus                                                 A-1
A-2  Experimental Setup and Monitoring                                    A-7
A-3  Experimental Results - 1,000 Hour Tests                               A-15
A-4  Experimental Results - 10,000 Hour Tests                              A-23
A-5  Analysis of Test Results                                              A-23
A-6  Summary and Conclusions                                            A-40
A-7  Recommendations for the Field Placement of GMs                       A-45
A-8  References                                                         A-47

Appendix B - Antioxidant Depletion Time in High Density Polyethylene Geomembranes

B-1  Introduction                                                         B-1
B-2  Formulation, Compounding and Fabrication of HOPE GMs                B-2
B-3  Stages of Degradation in HOPE GMs                                   B-5

     B-3.1  Depletion of Antioxidants                                        B-6
     B.3.2  Induction Time                                                 B-6
     B.3.3  Material Property Degradation                                   B-8

B-4  Major Influences on Oxidation Behavior                                 B-10

     B-4.1  Internal Material Effects                                         B-10
     B-4.2  External Environmental Effects                                  B-12
     B-4.3  Commentary on Various Influences                              B-13

B-5  Overview of Antioxidants                                              B-13

     B-5.1  Function of Antioxidants                                         B-14
     B-5.2  Types and Characteristics of Antioxidants                         B-15
     B-5.3  Antioxidant Depletion Mechanisms                               B-16

B-6  Experimental Design                                                 B-18
B-7  Evaluation Tests on Incubated Samples                                 B-20

     B-7.1  Standard Oxidative Induction Time (Std-OIT) Test                  B-21
     B-7.2  High Pressure Oxidative Induction Time (HP-OIT) Test             B-21
     B-7.3  Commentary on Different OIT Tests                              B-22

B-8  Data Extrapolation Method                                            B-23
B-9  Results and Data Analysis on Antioxidant Depletion                      B-24

     B-9.1  Preparation of OIT Test Specimens                              B-24
     B-9.2  Results and Data Analysis of Incubation Series I                   B-25
                                      XXII

-------
     B-9.3  Results and Data Analysis of Incubation Series III                   B-31
     B-9.4  Status of Incubation Series II and IV                               B-36

B-10 Summary                                                            B-36
B-11 Conclusion                                                           B-37
B-12 References                                                           B-38
B-13 Acknowledgements                                                    B-39

Appendix C - Field Performance Data for Compacted Clay Liners

C-1  Introduction                                                           C-1
C-2  Data for Natural Soil Liner Materials                                      C-1
C-3  Data for Soil-Bentonite Admixed Liners                                   C-2

Appendix D - Cincinnati Geosynthetic Clay Liner Test Site

D-1  Introduction                                                           D-1
D-2  Test Plots                                                            D-1

     D-2.1    Expectations at the Beginning of the Project                      D-2
     D-2.2    Layout of the Test Plots                                        D-2
     D-2.3    Plot Compositions                                             D-2
     D-2.4    Anchor Trenches                                              D-5
     D-2.5    Toe Detail                                                    D-6
     D-2.6    Instrumentation                                               D-6

             D-2.6.1  Moisture Sensors                                     D-6
             D-2.6.2  Displacement Gauges                                 D-17

     D-2.7    Construction                                                  D-18
     D-2.8    Cutting of the Geosynthetics                                    D-22
     D-2.9    Supplemental Analyses of Subsoil Characteristics                 D-24
     D-2.10  Results of Water Absorption Tests                               D-25

D-3  Laboratory Shear Tests                                                 D-29

     D-3.1    Testing Method                                               D-29
     D-3.2    Results                                                       D-29

D-4  Performance of Test Plots                                              D-33

     D-4.1    Construction Displacement                                     D-33
     D-4.2    Post-Construction Displacement of 3H :1 V Slopes                 D-36

             D-4.2.1  Test Plot A (Bentonite Between Two GMs)               D-36
             D-4.2.2  Test Plots B, C, and D (GT-Encased GCLs)              D-36
                                      XXIII

-------
             D-4.2.3  Test Plot E (Unreinforced GCL)                        D-36

     D-4.3    Post-Construction Performance of 2H: 1V Plots                   D-38

             D-4.3.1  Test Plots F and P (Bentonite Encased Between
                     Two GMs)                                          D-38
             D-4.3.2  Test Plots G and H                                   D-39
             D-4.3.3  Plots I and N with Nonwoven GT Component
                     Facing Upward                                      D-40
             D-4.3.4  Plots J, K, and L with No GM                          D-41

     D-4.4    Moisture Gage Readings                                      D-41

D-5  Tests to Study Lateral Spreading of Water in Bentonite                    D-42
D-6  Erosion Control Materials                                             D-44
D-7  Additional Laboratory Direct Shear Testing  on an Unreinforced GCL        D-45

     D-7.1    Materials Tested                                             D-45
     D-7.2    Direct Shear Tests                                           D-46

             D-7.2.1  Testing Equipment                                   D-46
             D-7.2.2  Testing Variables                                    D-46
             D-7.2.3  Hydration Time                                      D-46
             D-7.2.4  Shear Rate and Normal  Stress                        D-47

     D-7.3    Specimen Description and Preparation                          D-48
     D-7.4    Corrections for Shear Box Friction                             D-48
     D-7.5    Test Results                                                D-49

             D-7.5.1  Effect of Hydration Time                              D-49
             D-7.5.2  Effect of Shear Rate and Normal Stress                D-50

     D-7.6    One-Dimensional Consolidation Test                           D-51

D-8  References                                                         D-57

Attachment 1 - Results of Laboratory Direct Shear Tests on GCL Interfaces      D-58
Attachment 2 - Plots of Total Down-slope Displacements of GCLs Versus Time   D-66
Attachment 3 - Plots of Differential Displacement Between Upper and
              Lower Surfaces of GCLs Versus Time                         D-80
Attachment 4 - Plots of Moisture Sensor Readings Versus Time                D-94
Attachment 5 - Results of Laboratory Direct Shear Tests Performed on
              64-mm-Wide Specimens in University of Texas Laboratories     D-109
                                     XXIV

-------
Appendix E - Evaluation of Liquids Management Data for Double-Lined Landfills

E-1  Introduction                                                           E-1

     E-1.1      Purpose and Scope of Appendix                               E-1
     E-1.2      Organization of Appendix                                     E-1
     E-1.3      Definitions                                                  E-2

               E-1.3.1   Landfills                                            E-2
               E-1.3.2  Liner,  Liner System, and Double-Liner System          E-2
               E-1.3.3  Double-Liner System Components and Groups         E-2
               E-1.3.4  Cover, Daily Cover, Intermediate Cover, and Final
                       Cover System                                       E-4
               E-1.3.5  Waste Types in Landfills                              E-6
               E-1.3.6  Regions of the United States                          E-6
               E-1.3.7  LCRS Operational Stages                            E-6
               E-1.3.8  LDS Operational Stages                              E-9

E-2  Literature Review                                                      E-10

     E-2.1      Field Performance of Primary Liners                           E-10

               E-2.1.1   Overview                                           E-10
               E-2.1.2  GM Primary Liners                                  E-11

                       E-2.1.2.1  Bonaparte and Gross (1990, 1993)          E-11
                       E-2.1.2.2  Maule,  et al. (1993)                        E-11
                       E-2.1.2.3  Tedder (1997)                             E-11
                       E-2.1.2.4  Conclusions from Previous Studies          E-12

               E-2.1.3  Composite Primary Liners                            E-12

                       E-2.1.3.1  Bonaparte and Gross (1990, 1993)          E-12
                       E-2.1.3.2  Feeney and Maxson (1993)                 E-13
                       E-2.1.3.3  Workman (1993)                          E-13
                       E-2.1.3.4  Bergstrom et al. (1993)                     E-14
                       E-2.1.3.5  Bonaparte etal. (1996)                     E-15
                       E-2.1.3.6  Conclusions from Previous Studies          E-18

     E-2.2      Leachate Generation Rates                                   E-19
               E-2.2.1   Overview                                           E-19
               E-2.2.2  Feeney and Maxson (1993)                           E-19
               E-2.2.3  Maule etal. (1993)                                  E-19
               E-2.2.4  Haikola etal. (1995)                                  E-20
               E-2.2.5  Bonaparte etal. (1996)                               E-20
               E-2.2.6  Tedder (1997)                                       E-20
                                      XXV

-------
               E-2.2.7  Conclusions from Previous Studies                    E-20

     E-2.3      Leachate Chemistry                                          E-21

               E-2.3.1  Overview                                           E-21
               E-2.3.2  MSW                                              E-23

                       E-2.3.2.1  Introduction                               E-23
                       E-2.3.2.2  NUS(1988)                               E-23
                       E-2.3.2.3  Gibbons et al. (1992)                      E-28
                       E-2.3.2.4  Tedder (1992)                            E-29
                       E-2.3.2.5  Rowe(1995)                              E-30
                       E-2.3.2.6  Hunt and Dollins (1996)                    E-30
                       E-2.3.2.7  Conclusions from Previous Studies          E-31

               E-2.3.3  HW                                                E-32

                       E-2.3.3.1  Introduction                               E-32
                       E-2.3.3.2  Bramlett et al. (1987)                      E-36
                       E-2.3.3.3  NUS(1988)                               E-36
                       E-2.3.3.4  Gibbons et al. (1992)                      E-36
                       E-2.3.3.5  Pavelkaetal. (1994)                      E-37
                       E-2.3.3.6  Conclusions from Previous Studies          E-37

               E-2.3.4  ISW                                                E-38

                       E-2.3.4.1  Introduction                               E-38
                       E-2.3.4.2  MSW Ash                                E-38
                       E-2.3.4.3  Coal Ash                                 E-39
                       E-2.3.4.4  C&DW                                   E-40
                       E-2.3.4.5  Conclusions from Previous Studies          E-40

E-3  Data Collection and Reduction                                          E-41

     E-3.1      Overview                                                   E-41
     E-3.2      General Description of Cells                                  E-41
     E-3.3      LCRS and  LDS Flow Rate Data                               E-115
     E-3.4      Landfill Chemistry Data                                       E-115

E-4  Leakage Rates Through Primary Liners                                  E-117

     E-4.1      Overview                                                   E-117
     E-4.2      Leakage Rates Through GM Primary Liners                    E-118

               E-4.2.1  Description of Data                                  E-118
               E-4.2.2  Analysis of Data                                    E-120
                                      XXVI

-------
                       E-4.2.2.1   Interpretation of Data                       E-120
                       E-4.2.2.2   Summary of Flow Rate Data                E-120
                       E-4.2.2.3   Effects of LDS Material and CQA on LDS
                                 Flow Rates                               E-120
                       E-4.2.2.4   GM Primary Liner Efficiencies               E-128

              E-4.2.3  Implications for Landfill Performance                   E-130

     E-4.3     Leakage Rates Through Composite Primary Liners              E-131

              E-4.3.1  Description of Data                                  E-131
              E-4.3.2  GM/GCL Composite Primary Liners                    E-140

                       E-4.3.2.1   Interpretation of Data                       E-140
                       E-4.3.2.2   Summary of Flow Rate Data                E-140
                       E-4.3.2.3   Liner Efficiencies                           E-141

              E-4.3.3  GM/CCL and GM/GCL/CCL Composite Primary
                       Liners                                              E-141

                       E-4.3.3.1   Interpretation of Data                       E-141
                       E-4.3.3.2   Analysis of Flow Rate Data                 E-143
                       E-4.3.3.3   Analysis of Chemical Data                  E-144

              E-4.3.4  Comparison to Liner Leakage Calculation Results       E-163
              E-4.3.5  Implications for Landfill Performance                   E-165

E-5  Leachate Generation Rates                                             E-165

     E-5.1     Overview                                                   E-165
     E-5.2     Description of Data                                           E-166
     E-5.3     Analysis of Data                                             E-166
     E-5.4     Implications for Landfill Performance                           E-180

E-6  Leachate Chemistry Data                                               E-183

     E-6.1     Introduction                                                 E-183
     E-6.2     Characterization of Landfill Leachate Chemistry                 E-183

              E-6.2.1  Introduction                                         E-183
              E-6.2.2  MSW                                              E-191
              E-6.2.3  HW                                                E-192
              E-6.2.4  ISW                                               E-193

                       E-6.2.4.1   MSW Ash                                E-193
                                      XXVII

-------
                       E-6.2.4.2 Coal Ash                                 E-193
                       E-6.2.4.3 C&DW                                  E-193

     E-6.3     Comparison to Published Data                                E-194
     E-6.4     Effect of Regulations on Leachate Chemistry                   E-194

E-7  Conclusions                                                          E-195

     E-7.1     Primary Liner Leakage Rates and Efficiencies                  E-195

              E-7.1.1   GM Primary Liners                                  E-195
              E-7.1.2   Composite Primary Liners                            E-197

     E-7.2     Leachate Generation Rates                                  E-198
     E-7.3     Leachate Chemistry                                         E-200

E-8  References                                                           E-203

Appendix F - Waste Containment System Problems and Lessons Learned

F-1  Introduction                                                          F-1

     F-1.1     Appendix Purpose and Scope                                F-1
     F-1.2     Appendix Organization                                       F-1
     F-1.3     Terminology                                                F-2

F-2  Data on Waste Containment System  Problems                           F-5

     F-2.1     Data Collection Methodology                                 F-5
     F-2.2     Detection of Problems                                       F-5
     F-2.3     Problem Classification                                       F-7
     F-2.4     Problem Description                                         F-8

F-3  Evaluation of Identified Problems                                        F-17

     F-3.1     Introduction                                                F-17
     F-3.2     Landfill Liner Construction                                    F-18

              F-3.2.1   Overview                                          F-18
              F-3.2.2   Leakage Through Defects in HOPE GM Top Liner      F-18
              F-3.2.3   Leakage at Pipe Penetration of Top Liner              F-20
              F-3.2.4   Severe Wrinkling of HOPE GM Liner                  F-21
              F-3.2.5   Migration of Landfill Gas Beyond Liner System and
                       to Groundwater                                     F-21
              F-3.2.6   Other Problems                                     F-22
                                     XXVIII

-------
F-3.3      Landfill Liner Degradation                                    F-22

          F-3.3.1  Overview                                           F-22
          F-3.3.2  Liner Damage by Fire                                F-23
          F-3.3.3  Liner Damage During Well Installation                 F-23
          F-3.3.4  Other Problems                                     F-24

F-3.4      Landfill LCRS or LDS Construction                            F-25

          F-3.4.1  Overview                                           F-25
          F-3.4.2  Rainwater Entering LDS Through Anchor Trench       F-25
          F-3.4.3  HOPE Pipe Separated at Joints                       F-26
          F-3.4.4  Other Problems                                     F-26

F-3.5      Landfill LCRS or LDS Degradation                            F-27

          F-3.5.1  Overview                                           F-27
          F-3.5.2  Erosion of Sand Layer on Sideslopes                  F-27
          F-3.5.3  Degradation of GT Filter Due to Outdoor Exposure      F-27
          F-3.5.4  Other Problems                                     F-28

F-3.6      Landfill LCRS or LDS Malfunction                             F-29

          F-3.6.1  Overview                                           F-29
          F-3.6.2  Clogging of GT in  LCRS Piping System                F-29
          F-3.6.3  Other Problems                                     F-30

F-3.7      Landfill LCRS or LDS Operation                               F-31

          F-3.7.1  Overview                                           F-31
          F-3.7.2  Malfunction of Leachate Pump or Flow Rate
                  Measuring System                                  F-31
          F-3.7.3  Other Problems                                     F-32

F-3.8      Landfill Liner System Stability                                 F-32

          F-3.8.1  Overview                                           F-32
          F-3.8.2  Liner System Instability Due to Static Loading          F-33
          F-3.8.3  Liner System Instability Due to an Earthquake          F-36

F-3.9      Landfill Liner System Displacement                            F-37

          F-3.9.1  Overview                                           F-37
          F-3.9.2  Uplift of Liner System Geosynthetics by Landfill Gas    F-38
          F-3.9.3  Uplift of Composite Liner by Surface-Water Infiltration   F-38
                                 XXIX

-------
    F-3.10    Cover System Construction                                  F-39
    F-3.11    Cover System Degradation                                  F-40
    F-3.12    Cover System Stability                                      F-41

              F-3.12.1 Overview                                          F-41
              F-3.12.2 Cover System Failure During Construction             F-42
              F-3.12.3 Cover System Failure After Rainfall or a Thaw         F-43
              F-3.12.4 Soil Cover Damage Due to an Earthquake             F-45

    F-3.13    Cover System Displacement                                 F-45
    F-3.14    Impoundment Liner Construction                             F-46

              F-3.14.1 Overview                                          F-46
              F-3.14.2 Leakage Through Defects in HOPE GM Liner          F-47
              F-3.14.3 Other Problems                                    F-47

    F-3.15    Impoundment Liner Degradation                             F-48
    F-3.16    Impoundment Liner System Stability                          F-49

F-4 Significance of Identified Problems                                     F-49

    F-4.1     Introduction                                                F-49
    F-4.2     Environmental Impacts                                      F-51

              F-4.2.1  Introduction                                        F-51
              F-4.2.2  Landfill Liner Systems                               F-51
              F-4.2.3  Cover Systems                                     F-61
              F-4.2.4  Impoundment Liner Systems                        F-61

    F-4.3     Construction, Operation and Maintenance Impacts              F-63
    F-4.4     Cost Impacts                                               F-64

F-5 Conclusions                                                          F-65

F-6 Recommendations to Reduce identified Problems                        F-72

    F-6.1     Introduction                                                F-72
    F-6.2     General                                                   F-72
    F-6.3     Liners and Barriers                                          F-73
    F-6.4     Drainage Systems                                          F-77
    F-6.5     Surface and Protection Layers                               F-80
    F-6.6     Liner System and Cover System Stability                      F-81
    F-6.7     Liner System and Cover System Displacements                F-83

F-7 References                                                          F-84
                                     xxx

-------
Attachment F-A Case Histories of Waste Containment System Problems         F-89

F-A.1    Introduction                                                      F-89
F-A.2    Landfill Liner Construction                                          F-89
F-A.3    Landfill Liner Degradation                                          F-108
F-A.4    Landfill LCRS or LDS Construction                                   F-119
F-A.5    Landfill LCRS or LDS Degradation                                   F-126
F-A.6    Landfill LCRS or LDS Malfunction                                    F-133
F-A.7    Landfill LCRS or LDS Operation                                     F-138
F-A.8    Landfill Liner System Stability                                       F-142
F-A.9    Landfill Liner System Displacement                                  F-165
F-A. 10  Cover System Construction                                         F-170
F-A. 11  Cover System Degradation                                         F-173
F-A. 12  Cover System Stability                                             F-177
F-A. 13  Cover System Displacement                                        F-203
F-A. 14  Impoundment Liner Construction                                    F-206
F-A. 15  Impoundment Liner Degradation                                     F-209
F-A. 16  Impoundment Liner System Stability                                 F-211
F-A. 17  References                                                      F-213

Appendix G - Long-Term Landfill Management

G-1 Introduction                                                          G-1

G-2 Strategies for Long-Term Landfill Management                           G-1

G-3 Incorporating Management Strategies into Design                        G-8

G-4 Landfill Maintenance, Monitoring, and Response Actions                   G-8

G-5 Conclusion                                                          G-9

G-6 References                                                          G-11
                                     XXXI

-------
                                  Chapter 1
                                Introduction

The environmentally safe and secure containment of wastes in landfills, waste piles,
and surface impoundments has been a major goal of the United States Environmental
Protection Agency (EPA) since the Agency's founding in 1970. To bring about a
systematic and effective approach to the design and installation of liner systems and
final cover systems, as integral components of modern waste containment systems, the
Agency has developed regulations, supporting guidance, and numerous reports on this
subject. The agency has likewise known that proper facility operation and maintenance
are as important as design and construction in achieving satisfactory long-term
containment system performance. This research report provides the results of the
evaluation of field performance data for existing waste containment systems across the
U.S.  Based on this evaluation, it is concluded that environmentally safe and effective
containment of waste is attainable. This  research report also presents the results of a
number of technical tasks that have led to recommendations for further improving the
performance of waste containment systems in comparison to the current state-of-
practice.

This first chapter of this report presents an overview of the goals of waste containment,
the regulatory framework for waste containment, and the various components that make
up typical waste containment systems. The chapter concludes with a description of the
specific performance-related issues and technical tasks addressed by the studies
described  in this research report.

1.1 Goals of Waste Containment
An EPA estimate of the amount of municipal solid waste (MSW) generated in the U.S.
for select years between 1960 and 1999  is presented in Table 1 -1. This table does not
include construction and demolition waste (C&DW),  incinerator ash, sludges, and
nonhazardous industrial waste, all of which add to the quantities shown in the table.

It should be recognized that waste reduction and recycling programs are having a
positive impact on reducing the quantities of waste generated and disposed,
respectively. Nevertheless, disposal in landfills containing engineered waste
containment systems continues to be the most widely used method in the U.S. for the
disposal of MSW and many other types of waste.

The following classes of waste materials, listed in descending order of approximate
degree of hazard, constitute the majority of solid waste material requiring management
and/or disposal in the United States today:

   •   low-level radioactive waste;
   •   hazardous waste;
                                      1-1

-------
       Table 1-1. Generation, Materials Recovery, Composting, Combustion, and Discard of MSW, 1960 to 1999 (In
                 millions of tons and percent of total generation) (from Municipal Solid Waste in the United States:
                 1999 Final Report, downloaded from EPA website at http://www.epa.gov/epaoswer/non-
                 hw/muncpl/pubs/mswfinal.pdf).
                    Criteria
1960
1970
1980
1990
1995
1999
IV)

Generation
Recovery for recycling
Recovery for composting3
Total materials recovery
Discards after recovery13
Combustion0
Discards to landfill, other disposal01

88.1
5.6
0.0
5.6
82.5
27.0
55.5

121.1
8.0
0.0
8.0
113.0
25.1
87.9
Millions of
151.6
14.5
0.0
14.5
137.1
13.7
123.4
Percent of Total
Generation
Recovery for recycling
Recovery for composting3
Total materials recovery
Discards after recovery13
Combustion0
Discards to landfill, other disposal01
100.0%
6.4%
0.0%
6.4%
93.6%
30.6%
63.0%
100.0%
6.6%
0.0%
6.6%
93.4%
26.1%
67.3%
100.0% 1
7.1%
0.0%
7.1%
92.9%
20.6%
72.4%
Tons6
205.2
29.0
4.2
33.2
172.0
31.9
140.1

211.4
45.3
9.6
54.9
156.5
35.5
120.9

229.9
50.8
13.1
63.9
166.0
34.0
131.9
Generation6
00.0%
7.7%
0.0%
7.7%
92.3%
14.4%
77.8%
100.0%
9.6%
0.0%
9.6%
90.4%
9.0%
81 .4%
100.0%
9.9%
0.0%
9.9%
90.1%
7.1%
82.9%
       aComposting of yard trimmings and food wastes.  Does not include mixed MSW composting or backyard composting.
       bDoes not include residues from recycling or composting processes.
       cDoes not include residues from recycling, composting, or combustion processes.
       dlncludes combustion of MSW in mass burn or refuse-derived fuel-form, and combustion with energy recovery of source separated
        materials in MSW (e.g., wood pallets and tire-derived fuel).
       eDetails may not add to totals due to rounding.

-------
   •  heap leach residual waste;
   •  hospital/research waste;
   •  MSW;
   •  incinerator ash;
   •  sewage treatment sludge;
   •  contaminated dredge soil;
   •  electric power-generation ash;
   •  mine spoil; and
   •  C&DW.

A primary performance goal for waste containment systems used at all of these types of
facilities is protection of groundwater quality. Historically, the use of liners to protect
groundwater quality has been practiced for some types of landfills in some parts of the
country from about the mid 1970s. Since that time, the use of waste containment
systems has become more and more widespread, and the capabilities of these systems
have progressively improved.

The need for waste containment systems in landfills is driven in large part by the need
to contain liquids and gases generated in the landfill.  Leachate generated in landfills
flows downward by gravity and, if not for the liner system, would continue its migration
out of the unit. Given a sufficient volume of leachate,  this liquid would eventually
migrate through the vadose zone, ultimately posing a  threat to groundwater quality and,
at some locations, nearby surface-water quality.  Both the quantity and quality of
leachate  are of concern. In  addition, for MSW landfills, the  biodegradation of
putrescible  organics in the waste creates landfill gas.  This gas can be an added source
of groundwater contamination if not contained in the landfill and then removed by
appropriate means. The gas can also create explosion hazards and contribute to air
pollution.

Liquid containment is also an important consideration for surface impoundments that
contain various process liquids and liquid wastes. As with landfills, the function of the
liner system beneath a surface impoundment is to contain impounded liquid and prevent
it from migrating through the subsurface and into the groundwater at a rate that would
cause an adverse impact to groundwater quality (or surface-water quality), or at a rate
that would not comply with a regulatory performance criterion. The potential for liquid
migration can be particularly significant for surface impoundments, due to the relatively
high liquid heads that may exist in these facilities.

With respect to abandoned dumps and remediation sites, the situation is different than
for a modern landfill because these types of sites already exist and often were operated
without benefit of an engineered liner system and other environmental controls.  One
way to remediate these types of sites is to install  a final cover system over the waste.
At some  locations,  a cover system by itself will be adequate to achieve the desired
                                       1-3

-------
performance levels.  Other locations will require additional components, such as
subsurface barriers (e.g., soil-bentonite cutoff wall) or liquid/gas extraction systems.

1.2 Regulations
In the U.S., MSW, hazardous waste, and certain other wastes are regulated under the
Resource Conservation and  Recovery Act (RCRA), including the Hazardous and Solid
Waste Amendments (HSWA) to RCRA. As used by EPA, the term hazardous waste
has a very specific, legal definition. As defined in Title 40 of the Code of Federal
Regulations (CFR), Part 261  (40 CFR 261), waste is hazardous if:

   1.  it is listed as a hazardous waste (listed hazardous wastes are specifically
      identified in 40 CFR 261, Subpart D);
   2.  it is mixed with or derived from a hazardous waste as defined by EPA;
   3.  it is not excluded (some wastes, such as MSW, are specifically identified and
      excluded as hazardous waste); and
   4.  it possesses any one  of four characteristics described in 40 CFR 261, Subpart C:
      (i) ignitability; (ii) corrosivity; (iii) reactivity; or (iv) toxicity as determined by the
      Toxicity Characteristic Leaching Procedure (TCLP) test.

Federal legislation applicable to MSW is contained in Subtitle D of RCRA. Federal
regulations applicable to MSW landfills (and  nonhazardous MSWcombustor ash
landfills (MSW ash landfills) are set forth in 40 CFR 258.  The basic regulations were
published on October 9, 1991.  These regulations are implemented by states and
territories with landfill regulations or laws that have been approved by the EPA.  Forty-
nine of the 50 states have an approved program.  Federal regulations specify that a
MSW or MSW ash landfill  liner system meet the minimum design standard  in 40 CFR
258.40(a)(2) or meet the performance standard in 40 CFR 258.40(a)(1).  The design
standard requires a single-composite liner system that consists of the following, from
top to bottom:

•   leachate collection and removal system (LCRS) that limits the head of leachate on
    the composite liner to  0.3 m or less;
•   0.75-mm thick geomembrane (GM) (1.5-mm thick if the GM is made of high density
    polyethylene (HOPE)) upper component of composite liner;  and
•   0.6-m thick compacted clay liner (CCL) lower component of composite  liner, with
    the CCL having a maximum hydraulic conductivity of 1 x 10~7 cm/s.

While the federal minimum design standard was adopted by many states, a few states
require that MSW landfills or MSW ash landfills have a double-liner system.

The performance standard requires a liner system design that is demonstrated to
achieve certain groundwater compliance standards (i.e., maximum  contaminant levels
(MCLs)) at a specified distance from the landfill  (i.e., a point of compliance). This
distance cannot exceed 150  m. Only the Director of an approved State can approve a
                                      1-4

-------
 design that meets the performance standard. The technical demonstration that a
 certain liner system meets the performance standard is often made using the EPA
 HELP and MULT I MED computer models. The modeling methods must be acceptable
 to the Director.

 Regardless of whether an MSW landfill has a liner system that meets the minimum
 design standard or the performance standard, groundwater monitoring and compliance
 in accordance with 40 CFR 258.50-58 is required for the facility.  If the liner system
 does not meet the minimum design standard, leachate recirculation on the liner systems
 is not allowed as specified in 40 CFR 258.28(a)(2).

 An example of a single-composite liner system for a MSW landfill is shown in Figure
 1 -1 (a).  The LCRS will often include a pipe network that drains to a sump at the low
 elevation of the landfill cell.  From the sump, leachate  is removed by a  submersible
 pump or gravity drainage pipe. Where pumps are used, the pump is lowered in vertical
 manholes that extend up through the waste mass or, more commonly,  in riser pipes that
 extend up the sideslope of the landfill. Generally, leachate generated by a landfill will
 need to be collected for the active life of the landfill plus a 30-year post-closure period.
 However, the 30-year period has yet to be reached for any landfill constructed under
 current EPA regulations.  Longer periods of leachate removal may be required for at
 least some sites, while for many modern sites, leachate generation should essentially
 cease prior to the end of the 30-year period.
0.15m          Filter
 0.3 m    LCRS, k > 0.01 cm/s
 f
'x'x'x'x'x'x'x'x'x'x'x'x'x'x'x'x'x'x
VxVxVxVxVxVxVxVxVx
                      f
                            GM
 °6.mVx-x-vvvx-x-x-x
   I  'x'x'x'
              (a)
0.15m
 T
 0.3 m   LCRS, k > 0.01 cm/s
                                       4
 0.3m     LDS, k>0.01 cm/s

                                                                /-
                                                                2
                                                                       GM
                                                                       (primary liner)
 rGM
J (secondary liner)
     \^x'x'x'x'x'x'x'x'x'x'x'xVx/xVx  Composite
 0.9 m //V^^SS^^Q^TP/^^ (-secondary
   I  'x'x'x'x'x'xVx'x'x'x'x'xVx'x'x'x I liner

               (b)
 Figure 1-1. Example of liner systems for:  (a) MSW landfills; and (b) hazardous
            waste landfills.
                                       1-5

-------
For waste materials considered to be hazardous as previously defined, the applicable
legislation is contained in Subtitle C of RCRA.  Specific EPA regulations for waste
containment systems at RCRA Subtitle C landfills, surface impoundments, and waste
piles are published in 40 CFR 264. These regulations require hazardous waste landfills
to have two independent liners with a leak detection system (LDS) between them and
LCRS above the primary (or top) liner.  The purpose of the LDS is to allow monitoring of
the primary liner (i.e., to identify whether, and to what extent, leakage is occurring
through the primary liner) and to provide a mechanism for removing liquids that enter
this system.  A double-liner system with an LDS is a hallmark of hazardous waste
landfill regulations in the United States. A major task of the project described in this
report was to evaluate the field effectiveness of landfills underlain by double-liner
systems with respect to leachate containment.

Regulatory requirements for hazardous waste landfill double-liner systems are given in
40 CFR 264.301.  The minimum liner system design standard generally considered to
meet these requirements includes, from top to bottom:

   •  LCRS that limits the head of leachate on the primary liner to 0.3 m or less;
   •  GM primary liner;
   •  0.3-m thick granular LDS drainage layer with a minimum hydraulic conductivity of
      1 x 10~2 cm/s or a geosynthetic LDS drainage layer with a minimum hydraulic
      transmissivity of 3 x 10~5 m2/s;
   •  GM upper component of a composite secondary liner; and
   •  0.9-m thick CCL lower component of the composite secondary liner, with the
      CCL having a maximum hydraulic conductivity of 1 x 10~7 cm/s.

An example of a double-liner system for a hazardous waste landfill is shown in Figure
1-1 (b).

Federal regulatory requirements exist for the disposal of waste types other than MSW
and hazardous waste.  While this report is not intended to provide an exhaustive survey
of these requirements, it is noted that requirements for landfill disposal of
polychlorinated biphenyls (PCBs) and PCB items under the Toxic Substances Control
Act (TSCA) are  contained  in 40 CFR 761.65, while requirements for land disposal of
uranium mill tailings under the Uranium Mill Tailings Radiation Control Act (UMTRCA)
are contained in 40 CFR 192.02.

Final cover systems are another important component of waste containment systems
used at landfills. While liner systems are installed beneath the waste, final cover (or
closure) systems are installed over the completed solid waste mass. For MSW, Subtitle
D regulations require that the final cover must be placed over the landfill within one year
after the waste reaches its final permitted height.  In terms of long-term landfill
performance and management, final cover systems are as important, and in some ways
                                      1-6

-------
more so, than the liner system (Bonaparte, 1995). The design, construction, and
maintenance of final cover systems should be practiced to the same level of care as for
liner systems.

Requirements for final cover systems for MSW and hazardous waste landfills are also
addressed in federal regulations. For liner systems of the type shown in Figures 1 -1 (a)
and (b), minimum final  cover system requirements are illustrated in Figure 1-2.  MSW
landfills must meet federal design criteria or performance-based design requirements
(40 CFR 258.60). The minimum design for a MSW landfill (which is underlain by a
composite liner) cover system includes the following components, from top to bottom:

   •  0.15-m thick soil surface  layer;
   •  0.5-mm thick GM upper component of composite barrier; and
   •  0.45-m thick CCL lower component of composite barrier, with the CCL having a
      maximum hydraulic conductivity of 1 x 10~5 cm/s.

Under Subtitle D, alternative cover system designs are allowed, however, these designs
must, at a minimum, be shown to perform equivalently to the federal design cover
system with respect to  reduction in percolation and erosion resistance.

It should be noted that  the federal requirements for final cover systems at MSW landfills
are only minimum requirements and do not represent "complete" designs for most
landfills since they do not address all important design criteria.  Some of these criteria
are addressed in EPA (1991), which is currently being updated. For example, the
minimum requirements do not include a drainage layer above the composite barrier or
an adequate thickness of cover  soil to allow sufficient water storage for healthy surface
vegetation.  As another example, the requirements do not include an adequate
thickness of soil above the CCL component of the final cover system to protect the CCL
from freeze-thaw damage for sites located in northern climates.  As a final example, the
requirements do not address the important matter of landfill gas transmission beneath
the final cover system.

For hazardous waste landfills, 40 CFR 264.310 requires that the landfill be closed  with a
final cover system that meets certain performance criteria, most notably, "Have a
permeability less than or equal to the permeability of any bottom liner system or natural
subsoils present."  The regulations do not contain minimum design  requirements for
final cover systems analogous to those for liner systems. However, EPA guidance
(EPA, 1989) recommends that final cover systems for hazardous waste landfills consist
of at least the following, from top to bottom:
                                      1-7

-------
      a top layer containing two components: (i) either a vegetated or armored surface
      layer; and (ii) a 0.6-m thick protection layer, comprising topsoil and/or fill soil, as
      appropriate;
      a 0.3-m thick granular drainage layer with a minimum hydraulic conductivity of
      1 x 10"2 cm/s; and
      a composite hydraulic barrier, consisting  of (i) a 0.5-mm thick GM upper
      component; and (ii) a 0.6-m thick CCL lower component, with the CCL having a
      minimum hydraulic conductivity of 1 x 10~7 cm/s.
                             CCL, k  <1(rcm/s
                      Composite barrier
           (a)
                 As required
                      T
            > Frost penetration
Topsoil layer

Cover soil layer i
                       .  ..•:••.•.•••.•.•. Sand drainage layer, .•••.-:••.•.•-
                 As required .••:•.•,••:•.'.-..•. . .  . . _.2   ,•••••• ••:••'•:•••:
                       L   oVv-VvV': k > 1 0 cm/s :y-;V-..v-
                     0.6 m 'v CjCL k^J^O^crn/s v^ f- Composite barrier
                      __^^_ .• _ .•_ .•_ .•_ . •_ . •_ . • . • . • . . . _• . _• . _•. _•. m i. m •. i _ i ^
                 AS required ^^Gas drainage
                      T'
           (b)
Figure 1-2.  Examples of final cover systems for: (a) MSW landfills; and (b)
            hazardous waste landfills.

It is noted that at the time of publication of this report, EPA is concurrently completing a
new technical guidance document titled, "Technical Guidance for RCRA/CERCLA Final
Covers" (Bonaparte et al., 2002). The reader is referred to this guidance document for
more detailed information on final cover systems for landfills and remediation sites.

With respect to abandoned dumps and remediation sites, the Superfund Amendments
and Reauthorization Act of 1986 (SARA) adopts and expands a provision in the
                                       1-8

-------
Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA)
of 1982 to require that remedial actions at sites being remediated under the Act must at
least attain applicable or relevant and appropriate requirements (ARARs). These
requirements for ARARs may derive from federal or state regulations.  ARARs may be
location-specific, action-specific, or chemical-specific.

RCRA Subtitle C or D requirements for treatment, storage, and disposal facilities
(TSDFs) will frequently be considered ARARs for CERCLA actions, because RCRA
regulates the same or similar wastes or constituents as found at  many CERCLA sites,
covers many of the same activities, and addresses releases and threatened releases
similar to those found at CERCLA sites.  When RCRA requirements are ARARs, only
the substantive requirements of RCRA must be met if a CERCLA action is to be
conducted on site.  Substantive requirements are those requirements that pertain
directly to actions or conditions in the environment.  Examples include performance
standards for incinerators (40 CFR 264.343), treatment standards for land disposal of
restricted waste (40 CFR  268), and concentration limits, such as maximum contaminant
levels (MCLs).  On-site actions do not require RCRA permits or compliance with
administrative requirements. Administrative requirements are those mechanisms that
facilitate the implementation of the substantive requirements of a statue or regulation.
Examples include the requirements for preparing a contingency plan, submitting a
petition to delist a listed hazardous waste, recordkeeping, and consultations.  CERCLA
actions to be conducted off  site must comply with both substantive and administrative
RCRA requirements. CERCLA MSW landfills represent a particular subset of CERCLA
sites for which EPA has established presumptive remedy guidance (EPA, 1993).

RCRA and CERCLA regulatory requirements provide flexibility for innovation and
alternatives by limiting the use of specific minimum design specifications in the
regulations, by providing performance criteria in lieu of design specifications, and/or by
providing administrative procedures for gaining approval of waivers from RCRA
mandatory requirements or  CERCLA ARARs.  When proposing an alternative design to
the performance-based and/or federal minimum design requirements contained in the
applicable regulation, the  proposal for the alternative design must often be supported
with a demonstration that the alternate is "technically equivalent" to a design meeting
the basic regulatory requirements.  Alternative design approaches may be used for any
one of a number of different waste containment system components or group of
components, including  liner systems, final cover systems, LCRSs, and LDSs.

1.3  Waste Containment System Components
Waste containment systems are generally considered to included liner systems, final
cover systems, subsurface barriers, and subsurface interceptors constructed of a range
of materials including soil, geosynthetics,  cement, and/or metals.  This report addresses
liner systems and final  cover systems constructed of soils and geosynthetics.  The
                                     1-9

-------
following material choices may be considered for the design of these types waste
containment systems:

   •  drainage layer:  geonet (GN), geocomposite (GC), or granular soil;
   •  filter layer: geotextile (GT) or granular soil;
   •  hydraulic barrier: GM, geosynthetic clay liner (GCL), or CCL, or a combination of
      the three;
   •  gas transmission layer: GT, GC, or granular soil;
   •  protection layer:  GT or soil, and;
   •  erosion control:  geosynthetic erosion control (GEC) materials, natural jute,
      gravel, asphalt,  riprap, or other materials.

An example of a liner system and final cover system for a landfill that incorporates some
of these materials (primarily geosynthetics) is illustrated in Figure 1-3.  It is of interest to
compare this figure to the liner and cover systems of Figures 1-1 (b) and 1-2(b). The
liner system shown in Figure 1-3 incorporates a double-liner system consisting of a
GM/GCL composite primary liner and a GM/CCL composite secondary liner.  The LDS
consists of a GT/GN/GT GC. The LCRS is gravel with a perforated pipe network
contained therein. A GT filter layer covers the entire LCRS and is intended to inhibit
clogging of the LCRS.  A GT cushion beneath the gravel LCRS protects the primary GM
from puncture by the overlying gravel.  On the  sideslopes, the LCRS is constructed of a
GT/GN/GT GC,  which transitions, at the sideslope toe, into the gravel LCRS on the
base.

The final cover system illustrated in Figure 1-3 contains a GM/GCL composite hydraulic
barrier.  A GT gas transmission layer is shown beneath the barrier and a GT/GN/GT GC
(or other type of geosynthetic composite) is shown above it. A GEC is installed on the
surface of the topsoil layer.  Both temporary and permanent types of GECs are
commercially available.

Additional information regarding each of the natural soil and geosynthetic components
of the waste containment systems illustrated in Figures 1-1, 1-2, and 1-3 are presented
in the  next section.

1.4 Liner System  and Final Cover System Components
This section presents relevant details on liner/barrier, drainage, filtration, and ancillary
materials typically used in the liner system and final cover system of waste containment
facilities.

1.4.1  Liner/Barrier Materials
The types of hydraulic  liner/barrier materials considered in this report are CCLs, GMs,
and GCLs.
                                      1-10

-------
                                                                                                        Geocomposite Drain
                                                                                       Geotextile Filter
                               Geosynthetic Erosion
                                  Control System
                                                                          Geotexti e Gas Vent
                                                Geosynthetic Clay Liner
                                                                                                 SOLID WAST
                                                                  Geonet


                                                                      Geosynthetic Clay Liner


                                                                              Primary Geomembrane
                                                                                                .
                                                                                         xxxxxxxxxxxxxxxxx
    Geotextile
    Filter       V
                                                Secondary
                                              Geomembrane
                                                                                                                                       Geosynthetic Erosion
                                                                                                                                         Control System
Geotextile Filter
Geocomposite/Geonet Drain
Geomembrane
Geosynthetic Clay Liner
Geotextile Gas Vent
                                                                              •!•!•!•!•!•!•!•!•!•!• Compacted Clay Liner  •!•!•!•!•!•!•!•!•!•_
   Geotextile Filter
   Gravel w/Perforated Pipe
   Geotextile Protection
 — Geomembrane(Primary)
   Geosynthetic Clay Liner
   Geotextile Filter/Retention Layer
   Geocomposite/Geonet Drain
   Geomembrane(Secondary)

  1 Compacted Clay Liner
Figure 1-3.  Idealized solid waste containment system  (with emphasis on geosynthetic material utilization) for a solid
                waste landfill.

-------
1.4.1.1   Compacted Clay Liners
CCLs are constructed primarily from natural soil materials that are rich in natural clay,
although the CCL may contain processed natural clay such as bentonite. CCLs are
constructed in layers called lifts that typically have a thickness after compaction of
0.15 m. On sideslopes equal to or flatter than about 3 horizontal:  1 vertical (3H:1V) lifts
are placed parallel to the slope.  However, parallel lifts are very difficult or impossible to
construct on sideslopes steeper than about 2.51-1:1 V. On steeper sideslopes, CCLs are
constructed using horizontal lifts.

For CCLs that must have a saturated hydraulic conductivity of not  more than 1 x 10'7
cm/s, it is recommended that the CCL material have the following characteristics:

   •  minimum percentage of fines:       from 30 to 50%
   •  minimum plasticity index:           from 7 to 15%
   •  maximum percentage of gravel:     from 20 to 50%
   •  maximum particle size:             from 25 to 50 mm (less for a lift placed in
                                        direct contact with a CCL)

The percentage of fines is defined as the percent by dry weight of  particles passing the
No. 200 sieve, which has 0.074-mm wide square openings. Percentage of fines is
typically determined by ASTM D422. Plasticity index, which is defined as the liquid limit
minus the plastic limit, may be determined by ASTM D4318.  Percentage of gravel  is
defined as the percent by dry weight retained on a No. 4 sieve (4.76 mm wide square
openings).  Local experience may dictate more stringent requirements and, for some
soils,  more restrictive criteria may be appropriate.  However, if the  criteria tabulated
above are not met, it is unlikely that a natural soil liner material will be suitable without
additives such as bentonite.

CCLs must be ductile, particularly when used in final cover systems (to accommodate
possible differential settlement), and must be resistant to cracking  from moisture
variations, e.g., desiccation. Sand-clay mixtures are ideal  materials if resistance to
shrinkage and desiccation induced cracking are important  (Daniel  and Wu, 1993).
Ductility is achieved by avoiding use of dense, dry soils, which tend to be brittle.

If suitable materials are unavailable, local soils can be blended with commercial clays,
e.g., bentonite, to achieve low hydraulic conductivity. A relatively small amount of
sodium bentonite (typically 2 to 6% by weight) can lower hydraulic conductivity as much
as several orders of magnitude. Such liners are usually called amended clay liners and,
in this report, are included in the CCL category.  The percent bentonite is usually
defined as the dry weight of bentonite divided by the dry weight of  soil to which
bentonite is added.  Soils with a broad range of grain sizes usually require a relatively
small amount of bentonite (less or equal to 6%).  Uniform sized soils, such as concrete
sand, usually require more bentonite (up to 10 to 15%).  Sometimes materials are
                                       1-12

-------
blended to provide a material with a broad range of grain sizes, thus minimizing the
amount of bentonite amendment needed.

Some of the significant issues for CCLs are: (i) the accuracy of field hydraulic
conductivity assessment using laboratory tests on small undisturbed sample of the
constructed CCL; (ii) the compaction criteria to achieve the required CCL hydraulic
conductivity; and (iii) the long-term hydraulic performance of CCLs in final cover
systems. A major task of this project focused on these topics.

1.4.1.2   Geomembranes
GMs are thin,  factory-manufactured polymeric materials that are widely used as
hydraulic barriers in liner and final cover systems due to their non-porous structure,
flexibility, and ease of installation. GMs have the advantages of extremely low rates of
water and gas permeation through intact GMs and, depending on the material, the
ability to stretch and deform without tearing. They also protect underlying CCLs from
desiccation. Disadvantages of GMs include leakage through occasional GM
imperfections, relatively high diffusion potential by certain concentrated organic liquids,
potential for slippage along interfaces between GMs and adjacent materials, and
material embrittlement over time.

GMs form an essential component of most liner/barrier layers. Of the factory
manufactured polymeric GMs that are commercially available, the types most commonly
used in waste containment systems are:

   •  HOPE;
   •  very  flexible polyethylene (VFPE) [this classification includes linear low density
      polyethylene (LLDPE), low density linear polyethylene (LDLPE), and very low
      density polyethylene (VLDPE)];
   •  polyvinyl  chloride (PVC);
   •  flexible polypropylene (fPP); and
   •  ethylene  propylene diene monomer (EPDM).

Most of these GMs are available with textured surfaces on one or both sides for
increased frictional resistance when needed to achieve slope stability design criteria.
Additionally, spray-on elastomeric GMs are available, as are bituminous GMs.
However, these materials are rarely used in waste containment applications in
comparison to those previously itemized.

GMs are most often used as liquid and gas barriers, both  in liner systems and final
cover systems.  The mechanism for liquid or gas mass transfer through an intact GM is
one of molecular diffusion.  Water vapor transmission rates for several typical GMs
based on testing performed in accordance with ASTM E96 are as follows:
                                      1-13

-------
   •  for 1.0-mm thick HOPE:  water vapor transmission rate « 0.020 g/m2/day;
   •  for 0.75-mm thick PVC: water vapor transmission rate « 1.8 g/m2/day;
   •  for 1.0-mm thick HOPE:  solvent vapor transmission rate « 0.20 to 20 g/m2/day
      (depends on the solvent type).

Note that 1.0 g/m2/day « 10 liter/ha/day; thus the rate for water diffusion is extremely
low.  In contrast, the rate of diffusion for some chemicals, particularly certain volatile
organic compounds (VOCs) can be quite high.  Fortunately, leachate from modern
landfills typically contains only trace concentrations of VOCs and, as a consequence,
VOC diffusive mass transfer rates will typically be low.  A second mechanism for liquid
transport through GMs, is flow through GM holes causes by punctures, tears, flawed
seams, etc.  The rate of flow through a given size GM hole is dependent on the
hydraulic head acting on top of the hole, the permeability of the soil material underlying
the GM, and other factors.  The leakage rate through a GM hole where the GM is
underlain by a relatively permeable soil (e.g., sand) will be much larger than for a GM
hole where the GM is underlain by a CCL, all other factors being equal.

Regarding the shearing resistance of the interfaces between GMs and adjacent
materials,  interface strengths can be very low when smooth,  relatively rigid GMs  are
used. Strengths can be significantly increased through the use of textured GMs. There
are a number of manufacturing methods available  to  provide texturing:

   •  co-extrusion for blown film manufacturing;
   •  impingement for flat die manufacturing;
   •  lamination for flat die manufacturing; and
   •  structuring via a heated calendar for flat die manufacturing.

The texturing processes result in an increase in  peak interface shear strength compared
to the interface shear strength for a smooth GM. This increase may be in the range of
10 to 20 degrees for GM/GT interfaces. The difference may be of the same magnitude,
or less, for GM/soil interfaces, depending largely on the characteristics of the soil. The
difference  in interface strength is typically smaller when large displacement interface
strengths are considered.  Testing and experience has shown that the behavior of
geosynthetic/geosynthetic and soil/geosynthetic interfaces can be complex.  Product-
specific and  project-specific interface shear tests are always recommended. Interface
shear testing of geosynthetics is usually carried  out in a direct shear testing apparatus
in accordance with ASTM D5321.

1.4.1.3   Geosynthetic Clay Liners
GCLs consist of factory-manufactured rolls of bentonite placed between GTs or bonded
to a GM.  The bentonite is the low hydraulic conductivity (or permeability) component of
this composite material. The geosynthetics are stitch bonded, needlepunched, or
adhesively bonded to the bentonite to create self-contained products suitable for
                                      1-14

-------
handling, transportation, and placement as a barrier material. The fibrous structuring of
needlepunched and stitchbonded materials also results in increased internal shear
strength for use of GCLs on sideslopes. The application of GCLs as a barrier by itself,
or as a composite barrier with an overlying GM, is rapidly growing in its use and
acceptance. Three EPA workshop reports are available on GCLs (see Daniel and
Scranton, 1996)).

Bentonite is the critical component of GCLs and gives rise to the material's very low
hydraulic conductivity (permeability). Bentonite is a naturally occurring, mined clay
mineral that is extremely hydrophilic. When placed in the vicinity of water (or even
water vapor), the bentonite attracts water molecules into a complex configuration that
leaves  little free water space in the voids. This significantly decreases the hydraulic
conductivity of the bentonite. The hydraulic conductivity of most sodium bentonite
GCLs is in the vicinity of 1 x 1O9 to 5 x  1O9 cm/s (Estornell and Daniel, 1992).

The various GCL products are manufactured such that the following types are most
commonly used:

   •  bentonite adhesively  bonded between two GTs;
   •  bentonite stitch bonded between two GTs;
   •  bentonite needlepunched between two GTs; and
   •  bentonite adhesively  bonded onto a GM.

While the low hydraulic conductivity of GCLs gives rise to its' favorable comparison to
CCLs on the basis of a flow  rate or (flux) calculation, the assessment of full technical
equivalency is much more complicated. Koerner and  Daniel (1994) have proposed a
comparative assessment of GCLs to CCLs to be made on the basis of numerous
hydraulic, physical/mechanical, and construction criteria.

Using the above mentioned  criteria, GCL's are generally equivalent or superior to CCLs
with the exception of certain field installation issues, e.g., subgrade preparation,
puncturing, and direct contact by construction vehicles; with respect to certain hydraulic
issues  such as time-of-travel and degradation due to cation exchange; and with respect
to mass transport issues, such as diffusion and retardation. It is suggested that with
proper subgrade preparation and soil covering in a timely manner and of sufficient
thickness, GCLs can be adequately installed.  Equivalency with respect to the hydraulic
and other design criteria must be determined on a project-by-project basis. The issue of
the cation exchange potential of GCLs  has recently received much attention and the
reader is referred to Shackelford et al. (2000)  and Jo et al. (2001) for additional
information.

One of the more significant issues associated with the use of GCLs is that of adequate
shear strength when GCLs are installed on sideslopes. A major task of this project
                                      1-15

-------
focused on this topic.  Constructability issues involving GCLs are also important with
respect to composite liners, i.e., GM/GCL intimate contact.

1.4.1.4   Composite Liners
While any of the three liner materials just described (CCL, GM, and GCL) can be used
as a barrier material by itself, it is the combination of two or more of the components
that has proven to be most effective in terms of liquid and gas containment. In each
case of a composite liner, the GM forms the upper component, with the soil or GCL
being the lower component(s).  From practical experience, most composite liners fall
into one of the following categories:

   •  GM over CCL (GM/CCL);
   •  GM over GCL (GM/GCL); or
   •  GM over GCL over CCL (GM/GCL/CCL).

In all cases, the basic premise of using a composite liner is that leakage through a hole
or defect in the GM is  impeded by the presence of the CCL or GCL. Figure 1-4
illustrates the concept. If a CCL or GCL is used alone, liquid migration can occur over
the entire area of the liner that is subject to a hydraulic head.  If a GM  is used alone  and
is placed on a permeable substrate, the rate of flow through a hole in the GM can
approach the rate of flow through a similarly-sized orifice. In a composite liner,  leakage
will only occur at the location of the GM hole, but  it will be much slower than flow
through an orifice due to the hydraulic impedance provided by the CCL or GCL. The
level of impedance provided by the CCL or GCL is a function of the hydraulic
conductivity of that material, and the amount of lateral flow at the interface between  the
GM and CCL or GCL.  The amount of interface flow is a function of the "intimacy" of the
contact between the GM and CCL or GCL components (Giroud and Bonaparte, 1989;
Gross et al., 1990). Both theoretical investigations and field performance studies have
shown that leakage through composite liners is much less than leakage through GMs
alone or soil liners alone (Bonaparte and Othman, 1995).  Due to their superior
performance capabilities, in comparison to GMs, CCLs, or GCLs alone, composite liners
have been incorporated into federal minimum requirements for both MSWand
hazardous waste landfills, and they are being increasingly used in a wide variety of
waste containment system applications.

In considering the use of composite liners, design engineers are often faced with
evaluating the relative merits of using a GM/CCL  composite liner versus a GM/GCL
composite liner.  Technical, cost,  constructability,  and disposal capacity (i.e., airspace)
considerations will govern liner selection on a project-by-project basis.  An important
concept in comparing  GM/CCL and GM/GCL composite liners is "technical
equivalency."  Establishing the technical equivalency of a GM/GCL barrier to a GM/CCL
barrier on a specific project requires consideration of a number of design and
performance criteria.  In some cases, it may be advantageous to consider a three-
                                     1-16

-------
                        Leachate
                                                CCL or GCL
              (a) Flow through entire area of CCL or GCL
                        Leachate
                               \\\\\\\\\\\
            \\\\\\\\\\ '-.>. ^X\ \1\\\\\\\\\\\\
           ,v,v,v,v^v,x,KV,Vi,v,v,v,v,v f GM/CCL Composite
             (b) Flow through CCL only from hole in GM
                        Leachate
              (c) Flow through GCL only from hole in GM
Figure 1-4. Composite liner flow minimization concept.
                                  1-17

-------
component composite liner.  A three-component composite liner may be appropriate, for
example, where clay material capable of achieving the required hydraulic conductivity
performance criterion is not available, but use of a GCL by itself is not adequate for the
lower component of the composite liner (due to, for example, the need for a CCL
component to address issues related to time-of-travel, cation exchange,  or puncture
potential).

With respect to liquid  migration through a GM hole in a GM/GCL composite liner,
concern has been expressed with respect to the potential magnitude of interface flow
within the GT that covers the GCL. This concern, however, has been shown
experimentally to be of only minor consequence (Wilson-Fahmy and Koerner, 1995).
The reason behind this finding is that the bentonite of the GCL hydrates  and either
extrudes or intrudes into the covering GT.  A more significant issue than high
transmissivity within the covering GT is one  of possible lower interface shear strength
with the material above. The same issue holds for materials that are beneath the GCL.

GMs undergo expansion and contraction in response to exposure to sunlight and
temperature when placed and seamed together in the field.  If seaming occurs with the
GM taut, tensile stresses are induced in the  GM when the temperature decreases (e.g.,
after the GM is covered with soil) and the GM contracts.  GMs transitioning from
sideslopes to the flat interior of a landfill cell  have  lifted off the ground in  a trampoline-
like manner due to contractive stresses. To avoid trampolining, a GM  may be placed
with some slack such that during subsequent contraction at cooler temperatures, the
material will lie flat with essentially no internal tensile stress.  Slack is incorporated in
the form of waves, or wrinkles.  However, with this approach there is always a concern
that soil will be placed over the GM at a time when the waves still exist.  The issue of
the disposition of these waves after backfilling has been investigated, and the results of
the investigation are presented in this report. It should be mentioned that all GM types
(except reinforced GMs) have similar thermal coefficients of expansion.  However, stiffer
and thicker GMs, such as the polyethylene GMs, concentrate the waves and hence the
waves  are more pronounced and visible. Polyethylene GMs were the focus of the
investigation described herein.

1.4.2 Drainage Materials
Fluid collection,  conveyance, and removal represent another critical function of waste
containment systems for landfills, surface impoundments, and waste piles. The fluid to
be collected, conveyed, and removed will be leachate, water, impounded wastewater,
industrial liquid,  or landfill gas. There are five typical locations where drainage materials
may be required within a waste containment system:

   •  LCRS beneath solid waste;
   •  LDS between primary and secondary liners of a double-liner system;
   •  internal drainage layer above the barrier in  a final cover system;
                                      1-18

-------
   •  gas transmission layer beneath the barrier in a final cover system; and
   •  pore pressure relief system in areas of high groundwater.

Candidate drainage materials include soils, GNs, GTs, and/or GCs, and alternative
materials, such as tire chips. Granular soil and geosynthetic (i.e., GN, GC, and GT)
drainage layers are described below.

1.4.2.1   Granular Soils
Granular drainage materials are normally composed of relatively clean sand or gravel.
Gravel is material that does not pass through the 4.74-mm wide openings of a No. 4
sieve.  Sand consists of material that passes through the No. 4 sieve but not through
the 0.075-mm wide openings of a No. 200 sieve.  "Clean" sand or gravel refers to sand
or gravel that contains very little or no material that passes through the openings of a
No. 200 sieve.  Clean sands and gravels are often produced by washing natural sands
and gravels to remove any "fines," which are particles that pass through the openings of
a No. 200 sieve.

The drainage layer should meet filter criteria with the overlying layer of soil or waste.  If
the drainage layer material does not meet these criteria, a granular soil or GT filter will
be required.

Specifications for granular materials often require:

   •  no more than 2 to 5% (dry-weight basis) of material passing the No. 200 sieve; a
      "fines" content at the lower end of this range is usually preferable;
   •  a maximum particle size on the order of 25 to 50 mm; however, smaller particles
      will typically be required if a GM will underlie the drainage layer; alternatively, a
      GT cushion layer can be used;
   •  restrictions on gradation, stated in terms of allowable percentages for specified
      sieve sizes (these restrictions may exist for various purposes, including filtration
      considerations);
   •  limitations on mineralogy (often the drainage material is required to be a non-
      carbonaceous material,  with a limit on the amount of calcium carbonate in the
      material, although hard evidence that carbonaceous materials are truly
      unsuitable is lacking);
   •  restrictions on the angularity of the material, if the material will be in contact with
      geosynthetics, which are vulnerable to puncture by large, sharp objects (or,
      alternatively, a GT cushion may be employed);
   •  that no deleterious material be present; and
   •  a minimum acceptable saturated hydraulic conductivity.

The specified material requirements attempt to ensure that the materials will not
puncture adjacent geosynthetics, will be chemically stable, and will  provide adequate
drainage.
                                       1-19

-------
The required thickness and hydraulic conductivity of natural soil drainage layers should
always be established on the basis of site-specific and material-specific considerations.
It is not recommended that regulatory-suggested minimum values be used without
verifying by calculations that such values are adequate.  For example, regulatory
minimum hydraulic conductivity values of 1 x 10~2 cm/s (and in some states 1 x 10~3
cm/s) are often too low to satisfy rationally-based design criteria. The use of granular
drainage layers with permeabilities that are too low can lead to hydraulic head buildup
on liners or barriers and, in some cases, result in seepage-induced slope instability.
Lower permeability lateral drainage layers are also more prone to clogging and result in
longer leak detection times when used in an  LDS.  Higher hydraulic heads associated
with lower permeability drainage materials also increase the potential for liquid migration
out of the waste management unit.

The required flow capacity, qc (m3/s/m), of a  granular drainage layer must be equal to or
greater than the product of the maximum flow rate, qm (m3/s/m), obtained from the
design analyses and the factor of safety,  FS  (dimensionless):

                                  qc^qmFS                              (Eq.  1-1)

The maximum flow rate for design  should be established by appropriate analysis as
discussed below.  The FS selected for design should be based on the level of
uncertainly inherent in the design input parameters and the consequences of failure. A
minimum FS value of 2.5 is recommended for cases where the uncertainly in input
parameters is low and the consequences of failure are small (e.g., no slope instability
for a final cover system, little potential for increased percolation or leakage). For some
situations, a larger FS may be appropriate.  Koerner and Daniel (1997) have
recommended using  a FS value of at least 5 to 10 to account for the uncertainties
typically inherent in the assessment of waste containment system hydraulic conditions.

For granular drainage layers, the drainage layer hydraulic conductivity is selected to
provide adequate flow capacity and unconfined flow conditions.  For geosynthetic
drainage layers (discussed below), the drainage  layer hydraulic transmissivity is
selected to provide adequate flow capacity and unconfined flow conditions. For all
drainage layer materials, the required field hydraulic properties for design are evaluated
considering the material properties measured in the laboratory and reduction factors
that consider the potential for reduction in the property over time due to long-term
clogging, deformation, etc., in  the field.

For granular drainage layers, the field hydraulic conductivity can be computed as:


                             kf=k,f	1                          (Eq.  1-2)
                              1     '                                     V. ~1     /
                                      1-20

-------
where: kfieid = long-term field hydraulic conductivity of granular drainage layer (m/s); ki =
hydraulic conductivity of granular drainage layer (m/s) measured in the laboratory; RFCc
= reduction factor for chemical clogging (dimensionless); and RFBc = reduction factor for
biological clogging (dimensionless).

For geosynthetic drainage layers (discussed below), the field hydraulic transmissivity
can be computed as:


                       ef=e,f	1                        (Eq.  1-3)
                        f    '  RP  RP  RP RP                           V.  ~1    /
                             ^KI-,NKI-CRKI-CCKI-BC )

where: Of = long-term field hydraulic transmissivity of geosynthetic drainage layer
(m3/s/m); 6i = hydraulic transmissivity of geosynthetic drainage layer (m3/s/m) measured
in the laboratory; RF|N  = reduction factor for elastic deformation and/or or intrusion of the
adjacent geosynthetics into the drainage layer (dimensionless); RFCR = reduction factor
for creep deformation of the drainage layer and/or creep deformation of adjacent
materials into the drainage layer (dimensionless); and all other variables are as defined
previously.

It may occasionally be necessary to consider other reduction factors, such as factors for
installation damage or elevated temperature effects. If necessary, they can be included
on a site-specific basis. On the other hand, if the reduction factor has been included
some way in the test procedure for measuring the hydraulic property, the reduction
factor would appear in the foregoing formulation as a value of unity.

For design of LCRSs, the EPA Hydrologic Evaluation of Landfill Performance (HELP)
water balance model (Schroeder et al., 1994) is widely employed to obtain a leachate
generation rate for use in establishing an LCRS design flow rate (i.e., to establish qm).
The authors believe that the HELP model is useful for this purpose and as a design tool
for comparing different design scenarios. Limitations of the model as a predictive tool
are discussed subsequently in this report.  To establish the design maximum flow rate
for LDSs, a primary liner leakage rate must be assumed. Maximum primary liner
leakage rates are sometimes taken as regulatory action leakage rates (ALRs)  or are
established using an arbitrary conservative value (e.g., 1000  liters/ha-day) for  purposes
of hydraulic design.  A more rational approach has been presented by Giroud et al.
(1997).  The 2002 update to the EPA technical guidance document for RCRA/CERCLA
final cover systems  (i.e., Bonaparte et al., 2002)  provides a detailed discussion of
procedures to obtain the design  maximum flow rate for internal drainage layers in final
cover systems.

14.2.2   Geosynthetics
A number of different types of geosynthetics have been used as drainage layers in
waste containment systems.  Geosynthetic drainage materials that have been used in
                                      1-21

-------
these applications include:

   •   GNs of solid ribs with diamond-shaped apertures;
   •   GN of foamed ribs with diamond-shaped apertures;
   •   "high flow" GNs of solid ribs in a parallel orientation;
   •   drainage cores of single cuspations or dimples;
   •   drainage cores of double cuspations or dimples;
   •   drainage cores of built-up columns;
   •   drainage cores of stiff three-dimensional entangled mesh;
   •   needlepunched nonwoven GTs; and
   •   resin-bonded nonwoven GTs.

Like granular drainage layers, a geosynthetic drainage layer should meet filter criteria
with the overlying protection layer. Unless a GN or other core drainage material is
sandwiched between GMs, drainage cores require a GT filter to keep the overlying
material from directly clogging the apertures of the drain.  Furthermore, if a GM
hydraulic barrier underlies a GN or core drainage  layer, as is often the case, a GT may
be required between the drain and GM to provide  higher interface shear resistance on
sideslopes and, possibly, reduce deformation-related intrusion of the GM into the drain
and/or protect the GM from puncture or other damage by the drain. Often the GT is
heat bonded or glued to the GN or drainage core,  creating a GC, to enhance interface
shear strength,  decrease the potential for fugitive  soil particles to enter the drain during
construction, and facilitate installation.  If a GT drainage  layer is used, it should be
selected to meet filter criteria with  the overlying material.

Specifications for geosynthetic drainage layers often require:

   •   resin  and additive requirements;
   •   minimum thickness;
   •   minimum mass per unit area;
   •   minimum hydraulic transmissivity at a specified normal stress and  hydraulic
       gradient;
   •   minimum strength  requirements to survive  installation;
   •   if the  drainage material is a GN or core, inclusion of a GT filter above the drain;
       and
   •   if the  drainage material is a GN or core, inclusion of a GT beneath the drain, if
       necessary, to increase interface shear resistance, reduce deformation-related
       intrusion of an underlying hydraulic barrier  material into the drain, and/or protect
       the hydraulic barrier from puncture or other damage by the drain.

As with the hydraulic conductivity of a granular drainage layer, no specific minimum
hydraulic transmissivity can be recommended for  a geosynthetic drainage material
because the required value is site dependent. To minimize the potential for excessive
                                       1-22

-------
erosion and slope instability, however, the drainage layer should be able to convey the
maximum flow rate entirely in the layer without buildup of excess head.  It is noted that a
geosynthetic drainage layer is generally required to have a higher transmissivity than
that for a granular drainage layer to convey the required  design flow rate under
unconfined flow conditions. As discussed by Giroud et al. (2000), the geosynthetic
drainage layer hydraulic transmissivity that is equivalent  to a granular drainage layer
hydraulic transmissivity for these conditions can be calculated as:

                             edg = Eeds=Ekdstds                           (Eq. 1-4)

where:  9dg = geosynthetic drainage layer transmissivity (m3/s/m); E = equivalency factor
(dimensionless); 9ds = granular drainage layer transmissivity (m3/s/m); kds = granular
drainage layer hydraulic conductivity (m/s); and tds = granular drainage layer thickness
(m).  The equivalency factor can be approximated as (Giroud et al., 2000):
                         E =
                              1
                            0.88
(Eq. 1-5)
where:  Ld = length of drainage layer flow path (m), and all other terms are as defined
previously.

The hydraulic transmissivity of geosynthetic drainage layers can be measured in the
laboratory using ASTM D4716.  The test setup should simulate the actual field
conditions as closely as possible in terms of boundary conditions, stresses, and
gradient.

1.4.3  Filtration Materials
To prevent clogging of drainage layers, it is often necessary to install a granular or GT
filter layer directly over the drainage layer material. The function of the filter is to limit
the migration of fines from the overlying soil into the underlying drainage layer, while
allowing unimpeded flow of liquid through the filter and into to the drainage layer.  If
gravel is used as the drainage material, a filter is generally needed as a transition
between the overlying waste or soil and the gravel due to dissimilar particle sizes of the
respective soils. The filter can be either sand or a GT.  If a geosynthetic is used  as the
drainage material, the filter will always be a GT.

Filter  criteria establish the relationship of grain sizes necessary to retain adjacent
materials and prevent clogging of a drainage layer, while allowing unimpeded
percolation.
                                       1-23

-------
1.4.3.1    Granular Soils
Soil filters usually consist of fine to medium sand when placed over coarse sand or
gravel drainage layers.  The filter particle size distribution must be carefully selected.
Fortunately, there is a considerable body of information available to use in selecting a
filter particle size distribution (see Koerner and Daniel (1997)). Typically, the criteria
described in Cedergren (1989) are used.

To prevent piping from the overlying soil or waste layer into the filter, and from the filter
into the drainage layer, these criteria require, respectively:

                       D-I5 (filter)/ DSS (cover soil) < 4 to 5, and               (Eq. 1-6)

                       DIS (drainage layer)/ DSS (filter) < 4 to 5               (Eq. 1-7)

To maintain adequate permeability of the filter layer and drainage layer, respectively:

                           (filter)/ DI 5 (cover soil) > 4 to 5, and               (Eq. 1 -8)

                           (drainage layer)/ DIS (filter) > 4 to 5              (Eq. 1-9)
where: DSS = particle size at which 85% by dry weight of the soil particles are smaller
(mm); and DIS = particle size at which 15% by dry weight of the soil particles are
smaller (mm). The criteria should be satisfied for all layers or media in the drainage
system, including protection soil, filter material, and drainage material.

14.3.2    Geotextiles
A GT filter must be installed over a GN or GC drainage core when the adjacent material
is soil or waste.  GT filters are also commonly placed over granular soil drainage layers.
As with soil filter layers, GT filters must allow water or leachate to pass unimpeded into
the drainage layer while retaining the overlying material and limiting the migration of
fines into the drainage material. As with soil filter layers, the design of GT filters
involves a two-step process: first to assess permeability (or permittivity); and second to
evaluate soil retention  (or apparent opening size).

The first step in design of a GT filter is to establish the GT permittivity criterion.  The
approach  to defining this criterion involves first obtaining the permittivity required to
achieve unimpeded flow from the material overlying the GT (\j/req) and then applying a
factor of safety to obtain the minimum acceptable GT permittivity for the purpose of
establishing the construction specification requirement (\j/min). The following equations
may be used:
                             req
                                        1-24

-------
                                TJO                                    (Eq. 1-11)
                          Vmm=FSyreq

where: \j/ = GT permittivity (s~1); k0 = GT saturated hydraulic conductivity of overlying
material (m/s); t = thickness of GT at a specified normal pressure (m); and FS = factor
of safety.  A minimum factor of safety of 5 is recommended.

The testing of a GT for permittivity is conceptually similar to the testing of granular soils
permeability.  In the U.S., the testing is usually performed using the permittivity test,
ASTM D4491.  Alternatively, some design engineers prefer to work directly with
permeability and require the GT's permeability to be some multiple of the adjacent soil's
permeability (e.g., a minimum of 5 times higher).

The second step of the design of a GT filter is intended to assure adequate retention of
the upgradient soil. There are several methods available for establishing the soil
retention requirements of GT filters. Most of the available approaches involve a
comparison of the upstream material particle size characteristics and compare them to
the 95% opening size of the GT (i.e., defined as 695 of the GT). The 695 is  the
approximate largest soil particle size that can pass through the GT. Various test
methods are used to estimate 695:  (i) in the U.S., wet sieving is used and the value
thus obtained is called the apparent opening size (AOS), ASTM D4751; (ii) in Canada
and some European countries, hydrodynamic sieving is used and the value thus
obtained is called the Filtration Opening Size (FOS); and (iii) in other European
countries, wet sieving is used.

The simplest of the design methods compares the GT AOS to standard soil particle
sizes  as follows (Koerner, 1998):

   •  for soil with < 50% passing the No. 200 sieve (0.074 mm):  Ogs < 0.59 mm (i.e.,
      AOS of the GT > No.  30 sieve); and
   •  for soil with > 50% passing the No. 200 sieve:  Ogs < 0.33 mm (i.e., AOS of the
      GT > No. 50 sieve).

Alternatively, a series of direct comparisons of GT opening size (Ogs , OSQ , or Ois) can
be made to some soil particle size to be retained (Dgo, DSS , or DIS). The numeric value
depends on the GT type, soil type, flow regime, etc.  For example, Carroll (1983)
recommends the following relationship:

                          095 < (2 or 3) D85                            (Eq. 1-12)

where: DSS = particle size at which 85% by dry weight of the soil particles are smaller
(mm); and Ogs = the 95% opening size of the GT (mm).
                                      1-25

-------
However, as shown by Giroud (1982, 1996), this relationship should only be used if the
coefficient of uniformity of the soil to be protected is less than four. General procedures,
applicable for all values of the coefficient of uniformity of the soil to be protected, are
available, see Giroud (1982), Lafleuret al. (1989), and Luettich et al.  (1992).

Occasionally, a drainage layer is placed directly against a GCL. For GT-encased GCLs,
the GT components may not be adequate to prevent migration of bentonite into the
drainage layer.  The required filter criteria for this condition are under study, and the
manufacturer's and technical literature should be consulted.  One study indicated that a
350 g/m2 nonwoven, needlepunched GT provided adequate protection from bentonite
migration for all GCLs investigated (Estornell and Daniel, 1992).

1.4.4  Ancillary Materials and Components
There are a number of other geosynthetic materials that are occasionally or sometimes
used in waste containment systems.  These other materials  are briefly described below.

1.4.4.1   Plastic Pipe (aka Geopipe)
Plastic drainage pipe may be used for a variety of purposes  in a waste containment
system:

   •  leachate conveyance and removal within the LCRS;
   •  liquid conveyance and removal within the LDS;
   •  percolation water removal within the final cover system internal drainage layer;
   •  landfill gas transmission  and removal within a final cover system gas
      transmission layer;
   •  gas extraction wells in a  waste mass; and
   •  leachate injection into a waste mass where leachate recirculation is practiced.

The locations in a landfill where pipes are subjected to the highest compressive
stresses are in the LCRS and LDS. These collection systems typically underlie the
deepest parts of a landfill, and the compressive strength of the pipe may not be
adequate in landfills having large  depths of waste.

The allowable overburden stress that can be applied to a given plastic pipe is usually
governed by a limiting deflection criterion which design engineers often evaluate using
the Iowa State formula (Moser,  1990). This formula uses the full prism weight of the
height of overburden and is believed to be conservative (i.e., the formula does not
account for soil arching). The subject of plastic pipe capacity will be addressed in this
report.

The potential for pipe clogging must also be considered by landfill design engineers.
Several  states require annual pipe inspection and cleanout as a means to demonstrate
that a landfill piping system (or at least part of it) remains functional. While pipe
inspections provide information on conditions within the pipe, they do not provide
                                      1-26

-------
information on the condition of the pipe backfill or the condition of any filter layer
surrounding the pipe backfill. The long-term performance of these waste containment
system components with respect to clogging is a subject that merits further
investigation.

For piping systems above or within the solid waste (e.g., pipes in the final cover system
and pipes used for gas removal or leachate injection), a key design criterion is pipe
flexibility.  This flexibility is required to accommodate waste settlement (both total and
differential) and, in seismic impact zones, seismically-induced deformations.
Corrugated or profiled drainage pipe exhibits a high degree of flexibility compared to
rigid wall plastic pipe.  Corrugated pipe is, however, less strong than  rigid wall pipe, and
the performance of connections and outlet details must be adequately considered.

14.4.2    GM Protection
When granular soil is used to construct an LCRS, it is placed directly above the
hydraulic barrier layer, as shown in Figures 1-1 and 1-3.  If the soil consists of a coarse
sand or gravel, and if the hydraulic barrier includes a  GM, the possibility of puncturing of
the GM  exists. For this situation, a cushion layer may need to be placed between the
granular soil and GM.  Needlepunched nonwoven GTs are typically used in this
application. The key design parameter for GT cushions is the required mass per unit
area.  An investigation of GM puncture protection was undertaken during the course of
this study, and a design methodology was developed for calculating the required  mass
per unit area of GT needed to prevent puncture.  The results of this study are presented
in Chapter 2.

1.4.4.3    Erosion Control
For many final cover systems, the establishment of plant species may be aided by
placing a natural or GEC layer on the surface before, during, or after seeding. The fact
that the final cover system construction is often completed late in the year (i.e., often
occurs at the  end of the growing season) adds to the need for proactive erosion control
measures. Erosion can be harmful in more ways than simply adding to maintenance
costs.  For example, erosion can lead to clogging of toe drains and exposure of the final
cover system internal drainage and/or barrier to unanticipated physical and climatic
stresses.

The selection of erosion control materials is based upon the slope angle, slope length,
hydrology, time of year, etc. Indeed, there are many  such materials to fulfill site-specific
needs.  Theisen (1992) categorizes the materials, and each is further described in
Koerner and Daniel (1997).  The field performance of several GEC materials was
evaluated as  a component of the GCL test plot program described in this report.

1.5 Issues Evaluated in This Study
During the course of this four-year study, various concerns regarding the design
construction,  and performance of waste containment  systems were investigated.  By
                                      1-27

-------
association with the different principal investigators, these concerns were divided into
three broad areas: (i) geosynthetic materials; (ii) natural soil materials; and (iii) field
performance. Each area will be described briefly in this section and will then be
elaborated upon in the individual chapters and appendices of this report.  The
remainder of this report is structured as follows:

   •  geosynthetic material studies are described  in Chapter 2 along with Appendices
      A and B; the principal investigator for these studies was Professor Robert M.
      Koerner, P.E.;
   •  natural soil material studies are described in Chapters 3 and 4 along with
      Appendices C and D; the principal investigator for these studies was Professor
      David E. Daniel, P.E.;
   •  field performance studies are described in Chapter 5 along with Appendices E
      and F; the principal investigator for these studies was Dr. Rudy Bonaparte, P.E.;
   •  a summary of the project is presented in Chapter 6; and
   •  long-term landfill management strategies are presented in Appendix G.

1.5.1 Geosynthetic Materials Tasks
Since at least 1982, when the EPA first promulgated regulations requiring the use of
GMs in hazardous waste landfill liner systems, a number of different geosynthetic
materials have been used in waste containment systems.  While geosynthetic materials
and design methods have advanced greatly since that time, a number of important
technical issues remain. The geosynthetic tasks were undertaken to address five such
issues.

1.5.1.1   Puncture Protection of GMs
The  possible puncture of GMs from underlying stones in the soil subgrade or from
overlying granular drainage materials was experimentally evaluated. The focus of the
evaluation was on HOPE GMs since this type of GM is widely used as a liner material
beneath the waste mass (where stresses are the highest).

Based on the results of the experimental investigation,  a design methodology was
developed that can be used to calculate the mass per unit  area required for a
needlepunched nonwoven GT to prevent puncture of an adjacent GM by a certain size
particle.

1.5.1.2   Wave Behavior in HOPE GMs
An experimental evaluation of the fate and disposition of waves, or wrinkles, in HOPE
GMs was undertaken. As previously discussed, waves can prevent intimate contact
between the GM and natural soil or GCL components of a  composite liner and disrupt
LCRS and/or LDS flow paths.  If severe, waves could also produce unacceptable
residual stresses in the GM, which can adversely impact GM service life. The
disposition of waves when subsequently covered with soil was evaluated through a
large-scale laboratory-testing program.  The effects of wave height, applied stress, GM
                                      1-28

-------
thickness, and ambient temperature on wave disposition were assessed. All of the
tests incorporated strain gaging on the GMs so as to evaluate residual stresses in the
test specimens.

All tests were conducted for 1,000 hours, except for one test performed for 10,000
hours.  A viscoelastic model was used to extrapolate GM tensile strains measured
during the 1,000 hour tests out to 10,000 hours. A second model was used to convert
the resulting strains into residual stresses. The results are informative and lead to
improved recommendations for GM installation.

1.5.1.3   Plastic Pipe Behavior Under High Overburden Stresses
With the current tendency toward large regional landfills, the height of landfilled wastes
is steadily increasing.  Fifty-meter high landfills are commonplace, and 100-m high
landfills are known to exist.  Pipes within the LCRS and LDS beneath such high landfills
must be able to function under the high imposed overburden stresses, or, alternatively,
the performance limits of these materials (in terms of maximum allowable overburden
stress) must be defined.

In this study, the behavior of plastic pipe with respect to excessive deformation was
evaluated. A finite element model (FEM) was developed to model the stress-
deformation  behavior of plastic pipe under high overburden stress. A number of pipe
and bedding configurations, including different pipe wall thicknesses, were evaluated.
Graphs of waste height versus deformation under both short-term and long-term
conditions were developed and calibrated against available test data.  Design
recommendations are provided.

1.5.1.4   Prediction of GT Service Life
An experimental study to provide data to predict the service life of polypropylene (PP)
GTs, polyethylene (PE) geogrids, and polyester (PET) GTs was  initiated and is ongoing.
The study involves incubation  in forced air ovens (oxidation) for the PP GTs and PE
geogrids, and in water baths (hydrolysis) for the PET GTs. All incubations are at
elevated temperatures,  i.e., 50°C to 85°C. The data resulting from post-immersion tests
will be extrapolated to site-specific temperatures to estimate the time for 50%
degradation  of some engineering property (i.e., the halflife) for each material.

As part of this task, a side issue was investigated. This side issue had to do with the
potential for auto-catalytic degradation of GTs that had already been partially-degraded
by exposure to sunlight after the GTs were backfilled and protected from further
exposure to ultraviolet light.  The result of the experiments indicates that degradation
does not continue after the GT is buried,  i.e., the mechanism  is not auto-catalytic.
                                      1-29

-------
1.5.1.5   Prediction of GM Service Life
An experimental and analytical program was undertaken to develop estimates of the
service life of HOPE GMs.  As a first step in this effort, three stages in the degradation
process were identified:

       •  antioxidant depletion;
       •  induction; and
       •  half-life (i.e., 50% degradation) of an engineering property.

Laboratory incubation chambers, designed to simulate oxidizing conditions below the
GM and liquid exposure above the BM and a compressive stress equivalent to a 30 m
high landfill, were used to obtain data to estimate the time durations of the first two
stages of the  degradation process. Work is still ongoing to further define the duration of
the third stage under the test condition. Also, several additional incubation scenarios
are still under investigation, i.e., exposure of the GM to moving water, exposure in air,
and exposure under simulated sunlight.

1.5.2 Natural Materials Tasks
Natural materials (clays and granular soils) are widely used for a variety of functions in
waste containment systems.  A number of tasks were undertaken to address issues and
questions remaining with respect to the use of these natural materials in waste
containment systems.

1.5.2.1   GCL Test Plots in Cincinnati, Ohio
This task involved design, construction, and performance monitoring of 14 full-scale
final  cover system test plots, all containing a GCL hydraulic barrier component. The
test plots were constructed on both 3H:1V slopes and 2H:1V slopes. The goal of this
task was to evaluate the internal shear strength of three of the four types of
commercially-available GCLs (see Section 1.4.1.3), namely:

   •  needlepunched reinforced GCL;
   •  stitch-bonded reinforced GCL; and
   •  unreinforced GM/bentonite composite GCL.

Four different commercially-available products were evaluated.  The test plot slopes
were constructed in November of 1994, and internal stresses were mobilized by cutting
the overlying  geosynthetics in the spring of 1995. Monitoring (using subgrade moisture
sensors, bentonite moisture sensors, and deformation gages  on the upper and lower
surfaces of the GCLs) has been ongoing. This has resulted in a number of important
technical findings, recommendations regarding the design of GCLs for final cover
system applications, and recommendations for using GCL materials on sideslopes.
                                      1-30

-------
1.5.2.2   CCL Test Pad Analysis
This task involved collecting and analyzing data on the field performance of CCL test
pads that had been constructed and monitored. The test pads were located throughout
the U.S.  In all, data from 102 test pad projects were obtained and analyzed.  Eighty-
seven of the test pads were constructed to verify that a hydraulic conductivity of 1 x 10~7
cm/s or less could be achieved using the proposed project soil material and
construction methods. Test pad results were correlated to a number of different
variables, including:

   •  index properties;
   •  particle-size distribution;
   •  compaction moisture content;
   •  degree of saturation;
   •  compaction density; and
   •  total thickness.

This task also involved the analysis of laboratory hydraulic conductivity testing of soils
from the test pad sites (whenever available)  to establish correlation between results
from laboratory and field hydraulic conductivity tests.

1.5.2.3   Admixed Liners
This task focused on the use of soil-bentonite mixtures in admixed natural soil liners to
achieve a hydraulic conductivity of 1 x 10'7 cm/s or less.  Admixtures are often used
when local borrow soil alone is not capable of meeting the hydraulic conductivity
criterion. In many cases, the addition of bentonite into the soil, typically at a dry weight
application rate of 2 to 12%, will produce an  admixed soil liner capable of meeting the
hydraulic conductivity criterion.

For this task, a database of 12  case studies  was developed. The case study
information is presented and analyzed. Comparisons to the findings for CCL liners are
also presented.

1.5.2.4   CCLs in Final Covers
Federal regulations and most state  regulations allow the use of CCLs either alone, or in
combination with a GM, as a component  of landfill final cover systems. Concerns
associated with the use of CCLs in final cover systems include:

   •  degradation due to freeze-thaw;
   •  degradation due to shrink-swell;
   •  cracking from differential settlement; and
   •  deformations when placed on steep sideslopes.
                                      1-31

-------
For this task a number of field case histories were collected and analyzed.  Results of
the analysis are presented and  recommendations with respect to the use of CCLs in
final cover systems are presented.

1.5.3  Field Performance Tasks
Solid waste containment facility regulations have been in place for a number of years
and assessment of the field performance of facilities meeting these regulations and
especially the more recent regulations (e.g., land disposal restrictions of 40 CFR 268,
which were progressively implemented from 1986 to 1994), is both timely and  essential.
These tasks are focused on providing information and insight into the field performance
of waste containment systems,  particularly liner and final cover systems for landfills.
The four field performance tasks performed for this  project are described below.

1.5.3.1   Review of Published Information
Available published information on the field performance of modern waste containment
systems generally designed and constructed to current standards was collected and
reviewed.  The collected information, including approximately 100 technical papers and
reports, is related to the performance of liner systems, final cover systems, LCRSs, and
entire waste management units. The state of knowledge with respect to the field
performance of these systems has been assessed and is included in the following field
performance tasks.

1.5.3.2   Data Collection and Analysis
Data related to the performance of liner systems for double-lined waste management
facilities designed and constructed to current standards have been collected and
analyzed for 54  landfill facilities, representing a total of 189 landfill cells.  The data cover
more than an eight-year monitoring period for some facilities. The data was assembled
into a database that includes: (i) general facility information (e.g., location, average
annual rainfall, and subsurface soil type); (ii) general cell information (e.g., waste type,
cell area, dates of construction, operation, and closure); (iii) details of the liner system
and final cover system (e.g., material type, thickness, and hydraulic conductivity of each
layer); (iv) LCRS flow quantities and chemical constituent concentrations; and  (v) LDS
flow quantities and chemical constituent concentrations.  The results of this task are
summarized and analyzed, and conclusions are drawn with respect to leachate
generation  rates, GM and composite liner performance capabilities, and leachate
chemical constituents.

1.5.3.3   Assessment of Problem Facilities
Through the work conducted as part of the previous two tasks, as well as a
supplemental survey of the technical literature and interviews with regulatory personnel,
waste containment system problems were identified at 66 landfill and five surface
impoundment facilities.  The problems generally deal with the following areas;  (i) slope
instability or excessive deformation of liner systems or cover systems; (ii) defective as-
                                      1-32

-------
built components of liner systems or final cover systems; (iii) degraded components of
liner systems or final cover systems; and (iv) malfunction of LCRSs or LDSs, or
operational problems with these systems. The primary human factor contributing to the
problem is classified as design, construction, or operation related. Case histories of the
problems are provided.  The case histories document the observed problem, the
design, construction, and/or operational factors that led to the problem, and any
implemented solution and present specific lessons that can be learned from the
problem. Based on an evaluation of the data for the facilities, recommendations are
developed for reducing such problems in the future.

15.3.4   Comparison of Actual and HELP Model Predicted LCRS Flow Rates
Measured LCRS flow rates for eight landfill cells are compared to flow rates predicted
using Version 3.04a of the EPA HELP computer code. Cells were selected for
evaluation based on: (i) the completeness of design,  operation,  and LCRS flow rate
data for the cells; and (ii) waste type and geographic location of  the cells.  The
measured LCRS flow rates are compared to flow rates simulated using the HELP model
to assess whether the observed trends in measured flows from cells with different waste
types and  in different geographic locations are reasonably predicted. The HELP model
simulations were performed using estimated hydraulic properties for the landfill liner
system components  and waste and either: (i) synthetic solar radiation, rainfall, and
temperature data generated for the site using the HELP model, or (ii) synthetic solar
radiation data and actual rainfall and temperature data recorded in the vicinity of the
site. A parametric analysis is also performed to develop general guidelines on the
selection of HELP model input parameters to better predict LCRS flow rates.

1.6  References
ASTM D422 - Test Method for Particle-Size Analysis of Soils.
ASTM D4318 - Test Methods for Liquid Limit, Plastic  Limit, and  Plasticity Index of Soils
ASTM D4491 - Test Methods for Water Permeability of Geotextiles by Permittivity.
ASTM D4716 - Test Method for Constant Head Hydraulic Transmissivity (In-Plane Flow)
  of Geotextiles and Geotextile Related Products.
ASTM D4751 - Test Method for Determining Apparent Opening Size of a Geotextile.
ASTM D5321 - Test Method for Determining the Coefficient of Soil and Geosynthetic or
  Geosynthetic and  Geosynthetic Friction by the Direct Shear Method.
ASTM E96 - Test Methods for Water Vapor Transmission of Materials.
Bonaparte, R.,  Gross, B. A., Daniel, D. E., Koerner, R. M., and Dwyer, S. F. (2002),
  "Technical Guidance for RCRA/CERCLA Final Covers," U.S. EPA Office of Solid
  Waste and Emergency Response, Washington, D.C. (in final review)
Bonaparte, R. (1995), "Long-Term Performance of Landfills," Proceedings of the ASCE
  Specialty Conference Geoenvironment 2000, ASCE Geotechnical Special Publication
  No. 46, Vol.  1, pp. 415-553.
Bonaparte, R. and Othman, M. A. (1995), "Characteristics of Modern MSW Landfill
  Performance," Geotechnical News, Vol. 13, No. 1, pp. 25-30.
                                     1-33

-------
Carroll, R. G., Jr. (1983), "Geotextile Filter Criteria," Transportation Research Record
  916, Transportation Research Board, Washington, DC, pp. 46-53.
Cedergren, H. R. (1989), Seepage, Drainage, and Flow Nets, 3rd Edition, John Wiley &
  Sons, Inc., New York, NY, 464 pgs.
Daniel, D.  E. and Scranton, H. G. (1996), Report of 1995 Workshop on Geosynthetic
  Clay Liners, EPA/600/R-96/149, June, 93 pgs.
Daniel, D.  E. and Wu, Y. K. (1993), "Compacted Clay Liners and Covers for Arid Sites,"
  Journal of Geotechnical Engineering, Vol. 119, No. 2, pp. 223-227.
EPA (1989), Final Covers on Hazardous Waste Landfills and Surface Impoundments,
  Technical Guidance Document, EPA/530/SW-89/047, U.S. EPA, Office of Solid
  Waste and Emergency Response, Washington, D.C., 39 pgs.
EPA (1991), Design and Construction of RCRA/CERCLA Final Covers, Seminar
  Publication, EPA/625/4-91/025, U.S. EPA, Office of Research and Development,
  Washington, D.C., 145 pgs.
EPA (1993), Presumptive Remedy for CERCLA Municipal Landfill Sites, OSWER
  Directive No. 9355.0-49FS, EPA/540/F-93/035,  U.S. EPA, Office of Solid Waste and
  Emergency Response, Washington, D.C., 14 pgs.
Estornell, P. M. and Daniel, D. E. (1992), "Hydraulic Conductivity of Three Geosynthetic
  Clay Liners," Journal of Geotechnical Engineering, Vol. 118, No.  10, pp. 1592-1606.
Giroud, J.  P. (1982), "Filter Criteria for Geotextiles," Proceedings of the Second
  International Conference on Geotextiles, Las Vegas, NV, Vol. 1, Industrial Fabrics
  Association International, St. Paul, MN, pp. 103-108.
Giroud, J.  P. (1996), "Granular Filters and Geotextile Filters," Proceedings ofGeoFilters
  '96, Lafleur, J. and Rollin, A. L, Editors, Montreal, Canada, May 1996, pp. 565-680.
Giroud, J.  P. and Bonaparte, R.  (1989), "Leakage Through Liners Constructed with
  Geomembranes. Part II: Composite Liners," Geotextiles and Geomembranes, Vol.
  8, No. 2, pp. 77-111.
Giroud, J.  P., Gross,  B. A., Bonaparte, R. and McKelvey, J. A. (1997), "Leachate Flow in
  Leakage Collection Layers Due to Defects in Geomembrane Liners," Geosynthetics
  International, Vol. 4, Nos. 3 and 4, pp. 215-292.
Giroud, J.P., Zhao, A., and Bonaparte, R.B. (2000), "The Myth of Hydraulic
  Transmissivity Equivalency Between Geosynthetic and Granular Liquid Collection
  Layers", Geosynthetics International, Vol. 7, Nos. 4-6, pp. 381-401.
Gross, B. A., Bonaparte, R. and Giroud, J. P. (1990), "Evaluation of Flow from Landfill
  Leakage Detection Layers," Proceedings of the 4th International Conference on
  Geotextiles and Geomembranes and Related Products, The Hague, Netherlands,
  May 28 - June 1, 1990, pp. 481 -486.
Jo, H. Y., Katsumi, T., Benson, C. H., Edil, T. B. (2001), "Hydraulic Conductivity and
  Swelling of Nonprehydrated GCLs Permeated with Single-Species Salt Solutions",
  Journal of Geotechnical and Geoenvironmental Engineering, ASCE, Vol.  127, No. 7,
  pp. 557-567.
Koerner, R. M. (1998), Designing with Geosynthetics, 4th Edition, Prentice Hall,
  Englewood Cliffs, NJ, 761 pgs.
                                     1-34

-------
Koerner, R. M. and Daniel, D. E. (1994), "A suggested methodology for assessing the
  technical equivalency of GCLs to CCLs", Proceedings of Geosynthetics Research
  International-7 Conference on Geosynthetic Liner Systems, IFAI, St. Paul, MN, pp.
  255-275.
Koerner, R. M. and Daniel, D. E. (1997), Final Covers for Solid Waste Landfills and
  Abandoned Dumps, ASCE Press, Reston, VA, 321  pgs.
Lafleur, J., Mlynarek, J. and Rollin, A. L. (1989),  "Filtration of Broadly Graded
  Cohesionless Soils," Journal of Geotechnical  Engineering, ASCE, Vol. 115, No. 12,
  pp. 1747-1768.
Luettich, S. M., Giroud, J. P.  and Bachus, R. C. (1992), "Geotextile Filter Design Guide,"
  Geotextiles and Geomembranes, Vol. 11, pp.  355-370.
Moser, A. P. (1990), Buried Pipe Design, McGraw-Hill, New York, NY.
Schroeder, P. R., Dozier, T. S., Zappi, P. A., McEnroe, B. M., Sjostrom, J. W. and
  Peyton R. L., (1994) The Hydrologic Evaluation of Landfill Performance (HELP)
  Model:  Engineering Documentation for Version 3, EPA/600/R-94/168b, U. S.  EPA,
  Risk Reduction Engineering Laboratory, Cincinnati, OH.
Shackelford, C. D., Benson, C. H., Katsumi, T., Edil, T. B., and Lin, L.  (2000),
  "Evaluating the Hydraulic Conductivity of GCLs Permeated with Non-standard
  Liquids", Geotextiles and Geomembranes, Elsevier, Amsterdam, Vol. 18 Nos. 2-4,
  pp. 133-161.
Theisen, M.S. (1992), "The Role of Geosynthetics in Erosion and Sediment Control: An
  Overview," Journal of Geotextiles and Geomembranes, Vol.  11, Nos. 4-6, pp.  199-
  214.
Wilson-Fahmy, R. and Koerner, R. M. (1995), "Leakage  Rates Through a Hole in a GM
  Overlying a Geosynthetic Clay Liner," Proceedings of Geosynthetics '95, IFAI, St.
  Paul, MN, pp. 655-668.
                                     1-35

-------
                                  Chapter 2
                            Geosynthetic Tasks

This chapter presents the results of tasks that were directed specifically toward
geosynthetics. Five different topics were investigated on the basis of concerns shared
by the Agency and the principal investigators. The five topics are the following:

  • puncture protection of GMs;
  • wave behavior in HOPE GMs (with additional detail in Appendix A);
  • plastic pipe behavior under high vertical stresses;
  • lifetime prediction of GTs; and
  • lifetime prediction of GMs (with additional detail in Appendix B).

2.1  Puncture Protection of GMs
GMs are used as barrier layers in various applications such as, in liner systems and
cover systems for landfills, waste piles, and surface impoundments; as liners for oil and
gas tank secondary containment systems; and for other environmental applications.
Because GMs are located at the base of these contained materials, remediation of a
puncture failure would generally be very difficult and expensive. This, in addition to the
fact that detection of failure is  not easily accomplished and, in some cases, may not
even be possible, emphasizes the need to adequately protect GMs against puncture.
Leakage through the GM in many of these applications may pose a risk to human
health and the environment.

2.1.1  Overview
One of the mechanisms by which the hydraulic barrier function of a GM may be
compromised is puncture.  Puncture holes in a GM will increase the potential for
leakage through it, whether the GM is the entire hydraulic barrier or part of a composite
liner (Giroud and Bonaparte, 1989).

In GM applications, the puncturing object may be a stone in the subgrade or leachate
collection soil placed above the GM. Because of a lack of a rational design method,
puncture design is currently considered in a rather arbitrary manner. For example, the
current state-of-practice for landfill liner system applications in the U.S. involves the use
of a needlepunched nonwoven GT with a mass per unit area in the range of 250 to 600
g/m2 as a protection material depending upon the maximum size of material in contact
with the GM.  On the other hand, significantly heavier needlepunched nonwoven GTs
(mass per unit area of 1000 to 3000 g/m2) are used in Germany despite the fact that
only rounded gravel is allowed above the GM. The discrepancy between these two
practices emphasizes the need for a rational design method capable of providing a
puncture free GM as cost effective as possible.
                                      2-1

-------
In this chapter, a theoretical framework for evaluating the puncture behavior of GMs is
presented first.  The framework is applicable to unprotected GMs as well as GMs
protected using needlepunched nonwoven GTs. The puncturing object is characterized
by its shape and height above a firm subgrade. The GM behavior is considered in
terms of its tensile load-elongation behavior.  The protection material is characterized
by both its thickness and its load-elongation behavior.

Following development of the theoretical framework, the results of an experimental
study are presented in which the performance of a variety of protection materials is
evaluated.  The tested materials included nonwoven and woven GTs (both virgin and
recycled),  GCLs, used carpets (both industrial and domestic), and shredded tire rubber
mats. The main focus is, however, on virgin needlepunched nonwoven GTs since
these are the most commonly used materials for GM puncture protection. The GM
used in most of the study is a 1.5-mm thick HOPE GM.

Based on the data obtained from the experimental investigation and supplemented by
predictions using the theoretical analysis, a design methodology, with examples,  is
presented for use with needlepunched nonwoven GTs used as protection layers under
a variety of possible field conditions.

2.1.2  Theoretical Aspects of GM Puncture
A framework for considering GM puncture has been developed by considering
axisymmetric conditions on a single isolated protrusion (Figure 2-1). The loading is
assumed to be hydrostatic allowing for the deformed catenary shape of the GM to
gradually decrease into the underlying void space. Simultaneously, a larger portion of
the GM conforms to the tip of the protrusion.  To obtain a solution, several assumptions
are necessary:

  • the GM possesses no bending stiffness;
  • the load-extension behavior of the GM is linear elastic;
  • the GM in contact with the protrusion tip is assumed to be in a  state of equal stress
    by analogy with membrane theory for a GM subjected to hydrostatic pressure;
  • the contact between the GM and the protrusion tip is frictionless;
  • tensile strains cease to occur in the GM after it conforms to the subgrade, i.e., the
    portion of the GM in contact with the subgrade becomes fixed in its position;
  • the suspended portion of the GM  is in a state of equal tension, i.e.,  the force per
    unit width at any radius multiplied by the circumference at that  radius is constant;
    and
  •  Poisson's ratio effect is neglected.

From the equilibrium of an infinitesimal element, the following equation results:

                                                                       (Eq. 2-1)
                                      2-2

-------
where:

(Ri-x)
ds
i
Pi
Fi
radius from protrusion center to an infinitesimal element;
chord length of the element;
subscript referring to the instantaneous position of the GM;
pressure applied from overlying material;
total force in GM, i.e., force/unit width at any radius multiplied by
the circumference at that radius; and
change in tangent slope angle.
                                ds
                                    Detail A
i-irm son
d
-^

i
1
1
RO '

, -M Ij^,
f ^1
, -*-i r i
RI' ' !
wr \
^* 1
                                                                  Deformed
                                                                 void space
Figure 2-1.  Geometry used in theoretical analysis for GM puncture.
As part of this research study, Equation 2-1 was solved with various boundary
conditions and material (GM and GT) properties. From the result of this study, the
following conclusions were reached by Wilson-Fahmy et al. (1996):
                                       2-3

-------
  •  The protrusion height simulates an isolated stone on a soil subgrade.
  •  Decreasing the tip radius is indicative of sharper stones or other puncturing
     objects.
  •  Ro/H ratios greater than 4.0 are representative of an isolated stone case.
  •  For a closely spaced assemblage of cones, e.g., a gravel drainage layer, a
     protrusion height equal to one-half of the maximum stone size is a reasonable
     approximation.
  •  The required puncture resistance is decreased for smaller stones rather than large
     ones, rounded stones rather than angular ones, and a bed of stones as compared
     to isolated  stones.
  •  The puncture strength of a GM is improved with greater thickness, use of a GT,
     use of relatively thick GTs, and increased tensile strength of the GT.

2.1.3 Experimental Aspects of GM Puncture
A truncated cone arrangement was used to simulate a worst-case field condition. An
array of three such cones was placed in a pressure vessel and backfilled with sand,
leaving a protrusion height of a given amount (Figure 2-2). Using this device,
approximately 200 tests were conducted for a range of conditions to evaluate the effects
of GM puncture behavior. These tests consisted of the following material variations:

  •  HOPE GMs-1.0 to 2.5 mm thick;
  •  PP and PET needlepunched nonwoven GTs - mass per unit area of 130 to 1350
     g/m2;
  •  continuous filament and staple fiber GTs;
  •  an assortment of other  GTs made from virgin and post consumer plastics;
  •  discarded carpets; and
  •  rubber tire  mats.

The test results  demonstrated the improved puncture resistance of GMs protected by
needlepunched  nonwoven GTs compared to GMs alone.  Both PP and  PET, comprised
of either continuous or staple fibers, produced a set of curves with very uniform trends
(Figure 2-3). The test results for these materials were used to develop an equation that
forms the basis  of the design method:

                                      M
                          Pallow=450—                               (Eq. 2-2)
                                     i-r
where:

Pailow   = allowable bearing pressure for GT-protected 1.5-mm thick
          HOPE  GM (kPa);
M      = mass per unit area of the protection  GT (g/m2); and
H      = effective height of the protruding object (mm).
                                      2-4

-------
                         Water
                          inlet
Pressure
 gage
                                  o
                              (a) Pressure vessel
                                                           Cone height (CH)
                                                           Sand height varies
                                                 1^ 85 mm^|
                          (b) Details of truncated cones
Figure 2-2.  Details of the hydraulic pressure vessel and truncated cones for
            geomembrane puncture evaluation tests.
                                       2-5

-------
                  1000
                                                        CH=12mm

                                                        CH=25mm

                                                        C/-/=38mm

                                                        CH=50mm
                       0    200   400   600   800   1000   1200  1400

                                 Mass per unit area (g/m2)

Figure 2-3.  Failure pressure versus mass per unit area for nonwoven needle
            punched GTs at different cone heights. Data includes both PET
            continuous filament (not encircled) and PP staple fiber (encircled)
            GTs.
The failure pressures predicted using Equation 2-2 are plotted against the actual failure
pressures from the laboratory tests in Figure 2-4, which is seen to result in a correlation
coefficient of 0.973.

Equation 2-2 was then adjusted from the worst-case truncated cones used in the
experiments to actual soils through introduction of a number of different experimentally-
obtained modification factors to account for particle shape, packing density, arching,
and reduction factors to account for GT creep and long-term degradation. The final
form of the equation, as presented by Narejo et al. (1996), is:
where:

Callow
M
H
MFS
                                                                        (Eq. 2-3)
modified allowable pressure for GT-protected 1.5-mm thick HOPE GM (kPa);
mass per unit area of the protection GT (g/m2);
effective height of the protruding object (mm);
modification factor for protrusion shape (> 1.0);
modification factor for packing density (> 1.0);
                                       2-6

-------
        = modification factor for arching (> 1 .0);
        = reduction factor for creep (< 1 .0); and
        = reduction factor for chemical/biological degradation (<1 .0).
Finally, the factor of safety is formulated in the traditional manner:
                          pg_  Callow
                               ^applied
                                                                    (Eq. 2-4)
where:
FS       = factor of safety; and
Pappiied  = applied pressure (kPa) (i.e., the maximum pressure exerted on the GM).
         1000
                   Best Line Fit Based on Regression Analysis
               Pressure (empirical) = 2.56 + 1.02 x Pressure (measured)  RA2 = 0.973
     •5T   800
     Q_
     §
     w
     o>    600
     Q.

     I
     'co
          400
o
T3
2
Q_
          200
                                           Pressure (empirical) = Pressure (measured)
                         200         400         600         800

                             Measured Failure Pressure (kPa)
                                                                    1000
Figure 2-4.  Measured values versus empirically predicted failure pressures for all
            nonwoven needle punched GTs evaluated in association with 1.5 mm
            HOPE GMs. Data includes both PET continuous filament (not
            encircled) and PP staple fiber (encircled) GTs.
                                       2-7

-------
2.1.4  Puncture Protection Design Methodology
Equations 2-3 and 2-4 provide the basis for a puncture protection design methodology.
If the protection GT mass per unit area is known, the two equations are solved to obtain
a factor of safety against puncture.  If the protection GT is not known (as is usually the
case in design), a factor of safety is selected and the equations are solved for the
unknown, M, the required mass per unit area of the GT.  For purposes of the design
examples in this chapter, a factor of safety of 3.0 has been used.  This factor of safety
value is arbitrary, and the design engineer will need to select a value considering project
specific criteria. Given the uncertainties associated with any real design case, a
minimum factor of safety of 2.5 is recommended.

2.1.5  Examples
The following illustrative examples are taken from Koerner et al. (1996).

(a) A 30-m high landfill using 25 to 38-mm size gravel as a leachate collection layer on
   a 1.5-mm thick HOPE GM requires a protective cushion consisting of a GT
   protection layer having a mass per unit area of 590 g/m2 to achieve a FS = 3.0.
(b) The same problem, but now for 100 m of waste, requires a protective cushion
   consisting of a GT protection layer having a mass unit area of 1,600 g/m2 to achieve
   aFS = 3.0.
(c) Using a 2000 g/m2 GT under 16 to 32 mm rounded  leachate collection stone (as
   required in Germany) for a 30-m high  landfill results in  FS « 15.
(d) A 10-m deep surface impoundment with 12-mm size stones in the soil subgrade
   under the GM requires a 330 g/m2 GT, while the same conditions but 25-mm size
   stones requires a 550 g/m2 GT.  Both examples use a  FS = 3.0.

Using the available design methodology and a site-specific FS, the required mass per
unit area of a needlepunched nonwoven protection GT over or under an HOPE GM can
be obtained. It should be noted that puncture during installation is not addressed by this
method. GTs require a minimum mass per unit area based on installation
considerations alone (Richardson, 1996).

2.2 Wave Behavior in GMs
It is a frequent occurrence to see thermally-induced waves in seamed GMs after
installation and prior to covering or backfilling.  The fate and disposition of these waves
after backfilling were studied in this task.  The study involved laboratory modeling under
controlled conditions. Due to their relative stiffness and thickness, waves are of more
concern for HOPE GMs than for most other commercially-available products. Thus,
HOPE GMs were the focus of the study.  It should be noted, however, that all GMs in
commercial use have coefficients of expansion/contraction within an order of magnitude
of one another, and the problem  of excess slack is common to all types of GM products.
                                      2-8

-------
2.2.1  Large-Scale Experiments
A large-scale (1.8 m by 1.0 m by 1.0 m) experimental test box was constructed in the
laboratory to evaluate the behavior of HOPE GM waves. Smooth HOPE GM
specimens, 1.5 mm thick, of different lengths were placed in the box to create waves of
different sizes,  and sand  was placed over the specimens (Figure 2-5). The box had
plastic front and rear windows for visual monitoring of wave deformation under normal
stress. The results for the initial series of tests on large, moderate and small waves, as
quantified by the height-to-width (H/W) ratio, are given in Figure 2-6.  Note that even
under the highest normal stress that could be exerted by the box (e.g., 70 kPa), the
waves remained and the H/W ratios increased considerably over the  original H/W ratios.

By adjusting the ends of the waves to be located at the edges of the viewing windows
and applying normal stress,  it was seen that the wave end points did  not move,
irrespective of the amount of GM lying horizontally beyond the wave.  It was concluded
that as normal  stress is applied to the entire GM, the flat GM surfaces are held in
position by mobilized frictional forces between the GM and overlying and underlying
sand layers.  This was the case for all tests. Thus, the only wave movement possible is
in decreasing the void space beneath the original size wave.  There is no meaningful
lateral deformation beyond the wave itself.  One consequence of this observation is that
the frictional  forces on the upper and lower GM surfaces restrict the ability of GM waves
to dissipate laterally as normal stresses are applied  in the field.  A second
consequence, applicable to laboratory testing, is that smaller laboratory test boxes can
be used with full sized waves as the GM lengths beyond the wave end points need not
be large. It is easier to apply high normal stresses (e.g., greater than 1,000 kPa) to a
smaller box.  In addition,  the smaller setups could be housed in an environmental room.
As a result, four steel boxes measuring  300 mm by 300 mm by  300 mm were
assembled with thick plastic faces for visual observation.  The large box was
subsequently used for a 10,000 hour control test.

2.2.2  Small-Scale Experiments and Results
Using the smaller boxes, a number of variables were investigated (Table 2-1). In all of
these tests, at  least six electrical resistance strain gages were bonded to the surfaces of
the smooth HOPE GM near the crest and inflection points. The gages were bonded
onto the side of the GM subject to extensional deformations.  All tests were conducted
for 1,000 hours.  The results of these tests  are summarized below:

Regarding the  original  wave heights (which varied from 14 to 80 mm):
  • wave height decreased with increasing normal stress;
  •  an average reduction in wave heights  of 40% was observed after 1,000 hours;
  •  GM thickness had a negligible effect on the decrease of wave height with normal
    stress;
  • there was a slight decrease in wave height with increasing temperature;
  • final wave heights were 5 to 47 mm; and
  •  intimate contact with the soil subgrade was never achieved.
                                      2-9

-------
                                       1.8m
            Sand below
             /
*   05m     \  Geomembrane
                                                                0.6m
                                                                 0.4 m

                                                      To/fata acquisition
Figure 2-5.  Photograph and schematic illustration of the large-scale experimental

           test box used in wave study.
                                    2-10

-------
 (a) Relatively Large Wave
^\
Original
(0 kPa)
Final
(70 kPa)
Ht.
240 mm
125mm
H/W
0.5
2.0
              250 mm
                                                                   200 mm
                                                                   150 mm
                                                                   100 mm
                                                                   0 mm
(b) Moderate Wave
^^
Original
(0 kPa)
Final
(70 kPa)
Ht.
130 mm
70 mm
H/W
0.4
0.9
                                         s
X
              250 mm
                                                                   200 mm
                                                                   150 mm
                                                                   100 mm
                                                                   50 mm
                                                                   '0 mm
 (c) Relatively Small Wave
^\
Original
(0 kPa)
Final
(70 kPa)
Ht.
80 mm
35 mm
H/W
0.25
0.4
                                                                   250 mm
                                                                   200 mm
                                                                   150 mm
                                                                   100 mm
                                                                   50 mm
                                                                   0 mm
Figure 2-6. Results of profile-tracing of three preliminary tests on different
           size waves in 1.5 mm thick HOPE GMs subjected to increasing
           vertical pressures.
                                   2-11

-------
Table 2-1.  Experiments Conducted Using Small-Scale Test Boxes.	
  Experimental  	Experimental Conditions	
   Parameter    Normal Stress   Original Height of   GM Thickness  Temperature
   Evaluated        (kPa)	Wave, (mm)	(mm)          (°C)
Normal Stress
(kPa)
Height of Wave
(mm)
GM Thickness
(mm)
Testing
Temperature
180
360
700
1,100
700
700
700
60
14
20
40
60
80
60
14
20
40
60
1.5
1.5
1.0
1.5
2.0
2.5
1.5
23
23
23
23
42
55
Regarding the original H/W values for the waves (which varied from 0.17 to 0.33):

  •  H/W increased with increasing normal stress;
  •  H/W increased approximately linearly with increasing original wave height;
  •  H/W decreased approximately linearly with increasing GM thickness;
  •  H/W decreased slightly with increasing temperature; and
  •  final H/W values varied from 0.14 to 0.65.

Regarding the tensile strains measured along the top of the GM near the crest of the
wave and the bottom  of the GM near the inflection points of the wave at its sides:

  •  strains increased with increasing normal stress;
  •  strains increased with increasing original wave height;
  •  strains increased with increasing GM thickness;
  •  strains increased slightly with increasing temperature; and
  •  maximum strains within each test series varied from 3.2% to 4.9%.

2.2.3 Data Extrapolation and Analysis
The results of a 10,000-hour control test performed in the large box were compared to
predicted results extrapolated using experimental data up to 1,000 hours and the
Kelvin-chain model. This model has been shown to be applicable for extrapolating
physical property test results for a wide range of polymeric geosynthetic materials,
                                      2-12

-------
Soong and Koerner (1998).  The extrapolated data showed good agreement with the
experimental data. Therefore, it was considered reasonable to extrapolate all of the
data from the 1,000-hour experiments to 10,000 hours using the Kelvin-chain model.
The resulting values of maximum tensile strains were converted  to stresses using
temperature adjusted moduli values. The resulting tensile stresses were adjusted
downward to account for stress relaxation using a Maxwell-Weichert model, see Soong
and Lord (1998). The effect of stress relaxation was to significantly decrease the
residual stress values compared to the non-adjusted values.  However, the residual
stresses were still significant. Table 2-2 gives the residual stresses remaining in the
waves after 10,000 hours, and the corresponding percentage of  short-term yield stress
values.

2.2.4  Discussion
These laboratory tests appear to indicate that waves remaining in field-seamed HOPE
GMs at the time of covering  or backfilling do not disappear. Even the smallest wave (14
mm), at the highest normal stress (1100 kPa), for the thinnest GM (1.0 mm), at the
highest temperature (55°C),  remained elevated above the soil subgrade.  The authors'
conclusions with respect to the possible significance of these findings are as follows:

  •  intimate contact with the soil subgrade is not achieved when even only small
    waves remain in an HOPE GM upon backfilling, and even when the GM is
    subjected to relatively high normal stresses;
  • all waves not in contact with the subgrade have some amount of residual tensile
    stress, the amount depending primarily on the size and shape of the wave in its
    final configuration;
  • the waves probably form some retardation to flow of leachate on top of the GM, the
    implications of which have not been evaluated; and
  • possible long-term implications of trapped  waves in HOPE (and other types) GMs
    have not been evaluated and are beyond the scope of this project.

It is the authors' belief that the results of this task may have important ramifications for
the way in which GM liners are installed,  see Eith and  Koerner (1997).  The complete
results of this study are  given in Appendix A along with some of the possible
recommendations for minimizing GM waves during installation.

2.3 Plastic Pipe Behavior  Under High Vertical Stresses
A network of perforated pipes is generally required for the transmission of leachate in
the leachate collection and removal system (LCRS) of waste containment facilities.
HOPE or PVC pipes are normally used. The performance of plastic pipes under the
stresses imposed by large heights of waste in landfills is relatively unknown due to the
fact that experience with plastic pipes under high overburden stresses is limited. The
problem is further complicated by the fact that the behavior of plastics is time dependent
and their adequate performance is  required as long as there is a possibility of
generation of leachate within the landfill, which  can be for a considerable number of
                                      2-13

-------
 years. This is in contrast to the use of plastic pipes in transportation or agriculture
 applications where the design lifetime is relatively short, the live loads exerted on the
 pipe are mainly of temporary  nature, and the depth of burial is relatively small. For this
 reason, the applicability of classical design methods for flexible pipes in landfill
 applications can be questioned.  One of the limitations in applying classical design
 methods is that the effect of arching in alleviating the loads on pipes is probably
 underestimated (or ignored completely). Hence design according to classical theories
 may be overconservative. The response of plastic pipes under simulated conditions
 was investigated as a task of this study using a finite element modeling approach.

 Table 2-2. Residual Stresses (after 10,000 hours) in the HOPE GM Test
	Specimens for the Various Experiments Performed in this Study.	
     Experimental Variables and Conditions
Residual Stress
     (kPa)
Residual Stress
  (% of yield)
180kPa
Normal Stress 360 kPa
700 kPa
1100kPa
14 mm
20 mm
Original Height of Wave 40 mm
60 mm
80 mm
1.0 mm
Thickness of GM 1.5mm
2.0 mm
2.5 mm
23°C
14mm-42°C
55°C
23°C
Testing Temperature 20 mm - 42°C
55°C
23°C
40 mm - 42°C
55°C
23°C
60 mm - 42°C
55°C
1200
1300
2000
2100
130
740
1500
2000
2300
1600
2000
1600
1800
130
250
440
740
850
750
1500
1600
690
2000
2600
1600
8.5
9.2
13.7
14.4
0.9
5.1
9.8
13.7
15.7
10.5
13.7
13.7
11.8
0.8
2.1
4.5
4.9
7.3
8.0
9.5
13.7
7.4
13.2
22.0
17.5
 2.3.1  Leachate Removal Configurations
 As shown in Figure 1-3, the leachate removal system at the bottom of a landfill usually
 consists of a layer of sand or gravel (0.3 to 0.6 m thick) in which a perforated pipe
                                       2-14

-------
removal system is embedded. There are a number of possible configurations for the
pipe and its embedment, three of which are illustrated in Figure 2-7. The pipes are
arranged in a variety of patterns, e.g., a header pipe in the center of the landfill with a
series of feeder pipes located at right angles or at acute angles in a herringbone
fashion. The spacing is such that the mounded head on the liner is no more than the
regulatory maximum, 0.3 m.  Giroud and Houlihan (1997) describe  the analytic behavior
of this leachate mound. Critical in this regard is the minimum slope of the base of the
landfill, which should be larger than 0.5% after settlement or consolidation, since the
entire system drains by gravity to the low point of the cell.  At this location, a sump is
installed and the leachate is removed as required, recall Figure 1-5. The leachate
collection pipes are typically 150 to 200 mm in diameter.

2.3.2  Characteristics of Plastic Pipe
There are many candidate pipe materials for use in a leachate removal system beneath
a solid waste landfill. However, HOPE or PVC plastic pipe, also known as geopipe, has
emerged as the material of choice. There are a number of advantages to geopipe, but
also several disadvantages.

Advantages of Plastic Pipe
  • good flow characteristics;
  • no corrosion;
  • good resistance in chemically and biologically active environments;
  • light weight, which facilitates handling, storage, and installation;
  • pipe can bend along longitudinal axis; and
  • economical compared to other types of pipe.

Disadvantages of Plastic Pipe
  • low resistance to circumferential distortion;
  • due to their visco-elastic nature, creep and/or stress relaxation may constitute a
    problem under  long-term conditions;
  • strength and stiffness are temperature dependent; and
  • potentially susceptible to stress cracking if proper choice of resin and processing is
    not carefully considered.

Pipes from all thermoplastic materials are manufactured to the standard outside
diameters of traditional  pipe sizes.  For waste containment applications, PVC pipe  is
often specified to have a wall thickness corresponding to that of Schedule 80 pipe. The
dimensioning system often used with HOPE pipe is that in accordance with the standard
dimension ratio (SDR) which is equal to the ratio of the outside diameter to the minimum
wall thickness. An SDR of 11 is  often specified for HOPE pipes in landfills.  An SDR of
9 is the thickest wall section commonly used in the USA, however,  some German
landfills use HOPE pipe with an SDR of 6.
                                      2-15

-------
                                                                Drainage
                                                                Layer

                                                                Geotextile
                                                                (Protection
                                                                  Layer)

                                                                Geomembrane
                                                                 Compacted or
                                                                 Geosynthetic
                                                                 Clay Liner
                               (a) Trench Type of Installation
                    Waste
                           Pipe
Coarse
Drainage
Stone
                    Waste
                            (b) Embankment Type of Installation
                                                Coarse
                                                Drainage
                                                Stone
                                                                 Drainage
                                                                 Layer


                                                                 Geotextile
                                                                 (Protection
                                                                  Layer)

                                                                 Geomembrane

                                                                 Compacted or
                                                                ' Geosynthetic
                                                                 Clay Liner
                                                                 Drainage
                                                                 Layer

                                                                  Geotextile
                                                                 (Protection
                                                                   Layer)

                                                                 Geomembrane
                                                                 Compacted or
                                                                 Geosynthetic
                                                                 Clay Liner
                        (c) Embankment with V-Trench Type of Installation

  Figure 2-7.  Various pipe removal schemes within leachate collection layers.
     Figure 2-7. Various pipe removal schemes within leachate collection layers.

When perforated, the perforations are either in the form of slots or circular holes.
General practice is to use circular perforations for smooth walled pipes and slotted
                                               2-16

-------
perforations for corrugated pipes. The pattern of perforations consists of two or more
rows of holes of diameters ranging between 10 to 15 mm and spaced at distances
between 100 and 300 mm. The rows of holes are located symmetrically though offset
from each other in the lower 180 degrees or less of the pipe circumference.  With
perforated pipes, the slots are located at the valleys of the corrugations.

Various methods are used for joining of pipes.  They can be grouped into three
categories: butt fusion or seaming, overlap connections, and special couplings.

In general, butt fusion or seaming is used for thick-walled HOPE pipe (either solid or
perforated). The ends of the pipe are brought together with a heated plate placed
between them. A small force brings the ends of each pipe against opposite sides of the
plate. When adequate thermal energy is realized and the pipe ends become viscous,
the heat plate is removed and  the pipe ends are quickly brought together.  Adequate
force is applied to the opposing pipes to extrude a slight amount of the molten  material
out of the seam area. After cooling, the force is released and the seam is completed.
PVC pipe, on the other hand, is usually chemically seamed  using a solvent on  the pipe
ends before the pipe ends are drawn together.

The overlap type of connections can only be  made if the pipe thickness is adequate to
machine the pipe ends so as to accept one another.  To make a tight connection,
gaskets are sometimes  used which reside in  slotted seats of the thicker section of the
connection. Extrusion seaming can be used  from the outside of small diameter pipes or
from the inside of large diameter pipes to make a leak-free connection.

Special couplings are used to connect the ends of profile-wall (i.e., corrugated) pipes.
Each of these couplings must be mated to the type of pipe for which they were
designed.  It is not acceptable practice to use couplings made for one style of profiled
pipe on a different style.  Electro-fusion couplings are also used with smooth HOPE
pipe.

It should be noted that the influence of holes  (perforations or slots) and connections (of
all types) are not routinely accounted for as part of the design process. Design
engineers sometimes attempt to account for holes by assuming that the normal force on
the pipe is applied over  an area reduced by the size of the holes (i.e.,  an increased
normal stress is considered). The design method to follow is  based on the pipe itself,
not holes or connections, which represent an area of future  research activity.

2.3.3  Design by the Iowa State Formula
Design of plastic pipes (i.e., the calculation of pipe deflection) in most applications is
based on the modified Iowa State formula, which was originally developed in 1941, see
Spangler (1971). It was later modified by Watkins and Spangler (1958). Variations to
the Iowa State formula as well as other analytical approaches have been proposed for
                                      2-17

-------
predicting pipe deflection. However, such methods have not been generally accepted in
practice and hence only the Iowa State formula is presented herein. The formula takes
one of the following two forms:

                                 D, KW
                          A=-E—~	                            (Eq.2-5)
                              — + 0.061 E'
                              R3
or
                          A     D,KW~
                          - = -EJ—^	§	                            (Eq. 2-6)
                          D      +Q.061  E'
                               R3
where:
A    =   change in pipe diameter, m (A is used interchangeably in design for the
         horizontal and vertical deflections, Ax and Ay respectively as per ASTM
         D2412; in the derivation of the formula, A is the horizontal deflection and the
         deflected pipe is assumed to take an elliptic shape);
D|_   =   deflection lag factor (dimensionless);
K    =   bedding constant (dimensionless);
W   =   load per unit length of pipe (kN/m);
Ws   =   load per unit area (kPa);
R    =   mean radius of pipe (m);
D    =   mean diameter of pipe (m);
A/D  =   deflection ratio (dimensionless);
E    =   modulus of elasticity of pipe material (kPa);
I     =   moment of inertia of pipe wall per unit length (m4/m); and
E'    =   modulus of soil reaction (kPa).

The deflection  lag factor "D|_" is a result of soil compression at the sides of the pipe
whereby additional load may be exerted on the pipe with time. A value of 1.5 for the
deflection lag factor was originally proposed.  However, due to the inherent
conservatism in the formula, it has more recently been suggested that a value of 1.0 be
used (Moser, 1990). Note that in design,  the load W is taken  as the full prism load over
the pipe which, in the case of no variation of unit weight with height above the pipe, will
be equal to unit weight times the pipe diameter times the full height above pipe.
Accordingly, the load per unit area "Ws", will be equal to the overburden pressure above
the pipe. Thus, in using this approach, the effect of arching in relieving pipe stress is
not addressed, nor considered.  It has been pointed out by Moser (1990) that the long
term load will never exceed the prism load.

The bedding constant  "K" varies with the bedding angle.  However, a value of 0.1 is
often assumed in calculations since other parameters are much more significant.
                                      2-18

-------
Howard (1977) gives various values of E' for different soil types under different
compaction efforts which range between 3500 kPa (for poor quality fine grained backfill
soils) to 20,000 kPa (for good quality granular backfill soils). These values are
commonly used in design and are often referred to as  U.S. Bureau of Reclamation
(BuRec) values.

A maximum deflection, which is generally limited to 10%, is usually specified for flexible
pipes.  Excessive deflection of pipes can lead to reversal of curvature of the pipe ring
and substantial loss in flow capacity.  Once reversal of curvature occurs, fluctuations in
soil pressures can cause progressive ring deformation, which could lead to eventual
collapse (Watkins, 1987).

A parametric evaluation of the modified Iowa State formula under an  overburden
pressure of 1100 kPa is given in Figure 2-8 in which the deflection ratio Ay/D is plotted
against pipe stiffness for various values of the modulus of soil reaction. For a municipal
waste unit weight of 11.8 kN/m3, the chosen overburden pressure corresponds to a
waste height of about 90 m, and for a hazardous waste unit weight of 16.5 kN/m3 it
corresponds to a height of about 67 m.  Figure 2-8 indicates that the  deflection ratio is
more sensitive to pipe stiffness at  low values of the modulus of soil reaction than at high
values.  This emphasizes the importance of soil type and compaction effort in the zone
around the pipe.  Note that the pipe stiffness for HOPE pipes of SDR 11,9 and 6 are
approximately equal to 3400, 6600 and 26,400 kPa, respectively. As already
mentioned, pipe of SDR 11 is commonly used in the U.S. and pipe of SDR 6 has been
used in Germany.

2.3.4 Design by Finite Element Model
The finite element method has proven to be a versatile tool for many types of
geotechnical and structural analyses. The methods can be adapted to soil-pipe
interaction problems.

Figure  2-9 shows the discretization scheme used  in this particular finite element
analysis. Because  of symmetry, only half the geometry needs to be considered in the
analysis. In order to reduce the amount of input data,  a mesh generation subroutine is
incorporated in the  program. The  user specifies the number and  lengths of the vertical
and horizontal subdivisions around the pipe.  The numbering and coordinates of the
nodal points and the numbering of the elements are then automatically determined.

The soil and waste  are modeled using quadrilateral and triangular continuum elements.
The quadrilateral elements are four-node isoparametric elements, and the triangular
elements are constant strain elements.  Both elements have two degrees of freedom
per node and are compatible with  each  other. The pipe section is represented by
twelve  frame elements with three degrees of freedom per node.  The derivation of the
stiffness matrices of the frame elements and the triangular and quadrilateral elements
                                      2-19

-------
can be found in most finite element textbooks. The soil-pipe interface is represented by
twelve quadrilateral elements of zero thickness, which allow relative displacement only
in a direction parallel to the soil-pipe interface.
             20
03
DC.
c
o
"o

I
Q
             18


             16


             14


             12


             10


              8
E'=3500 kPa

E'=7000 kPa

E'=10000kPa

E'=14000 kPa

E'=17000kPa

E'=20000 kPa
                       5000     10000    15000     20000

                                   Pipe Stiffness (kPa)
      25000
                                                              30000
Figure 2-8.  Relationships between deflection ratio and pipe stiffness for various
            moduli of soil reaction using the Iowa State formula.
The boundary conditions are automatically set in order to minimize the size of the input
data.  The nodes lying on the bottom horizontal boundary are restricted from movement
in both the horizontal and vertical directions.  Symmetry is simulated by allowing only
vertical movement for the nodes lying on the pipe centerline with the exception of the
nodal point at the intersection with the bottom boundary, which is also restricted from
movement in the vertical direction.  Also, no rotation is allowed at the two pipe nodes at
                                       2-20

-------
     Ce
nter
      line
                                                             Continuum Elements
                                                              (Soil or Waste)
 Frame
 Element
Figure 2-9.  Finite element mesh developed and used in this study.
                                       2-21

-------
the invert and crown. The nodes lying on the vertical boundary on the right hand side of
the mesh are restricted from movement in the horizontal direction.

The stress-strain properties of the pipe, soil, solid waste and soil-pipe interface are
idealized in the finite element program using various available models. A brief
description follows:

   •  The plastic pipe is modeled as in a linear elastic material.  Thus, the modulus is
     critical in assessing both short-term and long-term deflection behavior.
   •  The soil around the pipe is modeled  using the hyperbolic model developed by
     Duncan and Chang (1970).
   •  The solid waste is also reproduced with a hyperbolic model, but with different
     modulus and strength parameters.
   •  The soil-pipe interface is represented using a shear stress versus relative
     displacement relationship presented by Clough and Duncan (1971).
   •  The program sequence,  along with the many details involved in the analysis, is
     given in Wilson-Fahmy and Koerner (1994).

2.3.5 Comparison of Design Methods
In order to illustrate the difference in pipe  deflection behavior as calculated by the Iowa
State formula versus this particular version of a FEM approach, a numeric example is
presented.

This example simulates the trench type of installation of Figure 2-7 (a).  Two different
HOPE pipes of 150 mm nominal diameter (external  diameter = 168 mm) are considered
in the analysis.  One has an SDR of 11 and the other has an SDR of 9.  The pipe
modulus is taken equal to 750 MPa to represent short-term conditions and 150 MPa to
represent long-term conditions. The pipe stiffness under short-term conditions is equal
to 3400 kPa for the SDR 11 pipe and 6600 kPa for the SDR 9 pipe. Other details are
given in Wilson-Fahmy and Koerner (1994).

Figure 2-10 shows the relationship between overburden stress and the diameter
changes in the vertical and horizontal directions under short and  long-term conditions
using the FEM approach.  As expected, the SDR 9 pipe deflects less than the SDR 11
pipe, the difference being larger under short-term conditions.

For comparison with the modified Iowa State formula, Table 2-3 gives the deflection
ratios predicted  using both the FEM and the Iowa State formula at 1100 kPa overburden
stress under both short-term and long-term conditions.  In applying the Iowa State
formula, the value of E' was taken equal to 21,000 kPa as representative of a coarse
stone of the type often used around plastic pipe in landfills.
                                      2-22

-------
CD
O)
c
CO
.c
O
CD
E
CO

b
     5.0
     4.0
3.0
     2.0
     1.0
     0.01
    10.0
 CD
 O)
 c
 CO
 .c
 O
 "CD
 E
 CO
 b
     2.0 -
     0.0
            SDR 11 (V)
                SDR 11 (H)
                          V = diameter decrease in vertical direction

                          H = diameter increase in horizontal direction

                       j	i	I	i	I	i	|_
                200
                      400        600

                        Pressure (kPa)
800
1000
                (a) Short-term pipe modulus = 750,000 kPa
                 200
                       400        600

                        Pressure (kPa)
 800
 1000
                 (b) Long-term pipe modulus = 150,000 kPa
 Figure 2-10.  Relationships between overburden stress and vertical and

              horizontal diameter changes using the FEM approach.
                                   2-23

-------
Table 2-3.  Deflection Ratios Predicted Using Iowa State Formula and
           Finite Element Analysis.
   Method of         SDR            E              Deflection Ratio (%)
    Analysis	(psi)	Short-Term	Long-Term
   Iowa State           11          21,OOOkPa           6.3              8.1
    Formula

 Finite Element         11             N/A              4.9              9.7
Iowa State
Formula
Finite Element
9
9
21,OOOkPa
N/A
4.9
3.7
7.6
8.2
where N/A = not applicable


The values in Table 2-3 indicate that, in comparison to FEM predictions, the Iowa State
formula overestimated the deflection ratio under short-term conditions (which is
conservative). Conversely, the formula slightly underestimated the vertical deflection
ratio under long-term conditions (which is slightly unconservative, but typically
accommodated in design by the incorporation of a factor of safety). In all cases, the
deflection ratio does not exceed 10% with the short-term deflection being less than
7.5%. Note that the short-term deflection predicted using the FEM is less than 5%.

The effect of arching of the soil above the pipe is clearly shown in Figure 2-11 where it
can be seen that under short term conditions (Es = 750,000 kPa) the effect is quite
small, i.e., the overburden stress on the pipe is approximately 80 to 90% of that
predicted using the  Iowa State formula. However, under long-term conditions (Es =
150,000 kPa), it is significant amounting to a decrease in overburden stress of the order
of 50%.  Furthermore, Figure 2-11 suggests that increasing  the pipe stiffness by
increasing the modulus and/or increasing thickness (using low SDR) results in less
potential effects of arching.

2.4  Prediction of GT Service Lifetime
A frequently asked question involving GTs is "how long will they last"? A study task was
developed to provide insight into this question. The task was subdivided into three
subtasks:

   • behavior of partially ultraviolet-degraded GTs;
   • oxidative degradation of PP and PE GT yarns and PE geogrid ribs; and
   • hydrolytic degradation of PET GT yarns.

Each subtask is briefly described in this report.  At the time of report preparation,
however,  only the first subtask has been completed.
                                       2-24

-------
             1000  -
             800  -
         03
         Q_
         CD
         Q.
         Q.  600  -
         c
         o
         I
2
Q_
             400  _
             200  -
                                  Overburden pressure
                                  (independent of pipe,
                                  per Iowa State formula)
                   SDR 9 (E=750,000 kPa)

                   SDR 11 (E=750,000 kPa)
                                             DR 9 (E=150,000 kPa)

                                            SDR 11 (E=150,000 kPa)
                            20
                         30
40
50
60
70
80    90
                                  Height above pipe (m)
Figure 2-11.  Reduction in overburden stress using FEM as compared to Iowa
             State formula, i.e., the effect of arching.
2.4.1  Behavior of Partially Ultraviolet Degraded GTs (PP and PET)
During any type of construction that involves the use of GTs, there will generally be the
potential for exposure to sunlight.  Within the sunlight spectrum, the ultraviolet (UV)
portion is most harmful to polymers used in the manufacture of GTs, causing a photo-
oxidative reaction to occur.  Thus it is necessary to cover or backfill GTs intended for
long-term service in a landfill in a timely manner. If not covered promptly, UV-
degradation will begin to occur.  A concern has been expressed that if such degradation
is initiated,  the reaction may continue to propagate within the GT even after covering or
backfilling.  This subtask was directed at addressing this issue, i.e., whether or not UV-
initiated degradation is auto-catalytic. The subtask involved incubation of GT
specimens in a laboratory UV fluorescent acceleration device per ASTM D 5208.  This
method was selected since the laboratory device provides for an economical approach
for providing UV exposures.  It also can be correlated to field exposure in the usual
manner, see Hsuan and Koerner (1993).
                                      2-25

-------
Twelve different types of GTs were included in the experimental program.  Nine were
PP and three were PET.  Figure 2-12 illustrates the overall experimental procedure, with
incubation, removal, and testing of the various samples. Details of the experimental
program are available in Hsuan et al. (1994).  Results for one of the PP GTs is given in
Figure 2-13(a) and for one of the PET GTs in  Figure 2-13(b).
              #1 - Light PP Needled
              #2 - Light PP Heatbonded
              #3 - Light PET Needled
#4 - Black PP Needled
#5 - Black PP Woven
#6 - Black PP Slit Film
                        UV (on/off)
                       (60°C/50°C)
     UV (on/off)
    (70°C/50°C)
                               Samples at 70%
                               Strength Retained
                         PP Samples    >
                      (#11 #2, #4, #5, #6)
                             &
                    UJnexposed Samples^
                        Forced Air Oven
                        (60°C for 70°C)
                              for
                          3000 hours
 Water Immersion
      (60°C)
        for
    3000 hours
Figure 2.12. Design flow chart of the complete study (after Hsuan et al., 1994).
The results of the UV exposure tests can be summarized as follows:

  •  UV exposure is the major contributor to the degradation of exposed GTs as
     opposed to thermal-oxidative or hydrolytic degradation. PP GTs were found to be
     slightly more susceptible to UV degradation than PET GTs.
  •  The rate of tensile strength reduction due to UV exposure (as a fraction of the
     original strength) is inversely proportional to the mass per unit area of the GT.  This
     is due to the degradation mechanism, which involves the initiation of photo-
     oxidation on the exposed surface of the GT, which moves progressively inwards as
     the duration of the exposure increases.
  •  The 70% UV-degraded GTs show similar trends as those unexposed GTs
     regarding the strength retained property within 3000 hours at 60°C.
                                     2-26

-------
               110
    10.


    -10
         o  UV degradation at 60°C
         A  Forced air oven at 60°C
         A  70% UV degraded and then 60°C oven
                         	A	
                      Removed from UV and
                      placed in oven at 60°C
                           1000       2000       3000

                             Exposure Time (hours)
                                            4000
                          (a) Polypropylene geotextiles
T3

ro
-t-«
ttl
_c
-t-J
D)
OJ

OT
            (D
            D.
               110-
                90.
                70.
                50-
                30
                      0  UV degradation at 60°C
                     ^   Water bath at 60°C
                     A   70% UV degraded and then 60°C water
                        Removed from UV and
                        placed in water at 60°C
                   0     1000    2000    3000   4000    5000
                              Exposure Time (hours)

                            (b) Polyester geotextiles


Figure 2-13. Strength retained graphs from geotextile incubated under
              three different test conditions.
                                    2-27

-------
  •  Photo-oxidative degradation ceased in all of the tested GTs (PP and PET) once
     the UV source was removed.

While the above findings are of technical interest, it is well known that "timely cover" is
necessary. Even a thin, 150-mm thick layer of soil is adequate to eliminate the potential
for UV degradation.  In this regard, specifications may require that the GT to be covered
within two to four weeks after placement unless the GT has been shown to resist
degradation for a longer time period. Pending further research,  the results of the testing
presented herein indicate that this maximum time frame is reasonable.

2.4.2 Oxidative Degradation  of PP GT Yarns and PE Geogrid Ribs
This subtask is being used to evaluate the oxidative degradation of two types of PP GT
yarns and two types of HOPE geogrid ribs.  The incubation uses forced air ovens at
elevated temperatures for varying times.  Samples are periodically removed and tested
for their retained strength and elongation.  These values are then compared to the
unaged strength and elongation for a percent retained value, thereby indicating the
degree of degradation.

Furthermore,  by incubating at several elevated temperatures (75, 65 and 55°C are used
in this study), an Arrhenius plot can  be developed.  Using the slope of the resulting
curve, extrapolation  down to site-specific temperature can be made, resulting in an
estimate of the service lifetime for these materials based on a thermal oxidation failure
criterion.

A separate oven is required for each incubation temperature.  Samples are periodically
removed and tested in tension.  Several of the resulting data available at the time of the
preparation of this report are presented in Figure 2-14 for PP GT yarns  and Figure 2-15
for PE geogrid ribs, respectively.  Each point is the average of five replicate tests.

This project is projected to continue for an additional three-year period,  thereby
achieving an estimated five-year maximum incubation time. The reason for this lengthy
duration of incubation is that shorter tests would require excessively high incubation
temperatures that would unrealistically bias the predicted service lifetime, i.e.,
temperatures that are too high may significantly underpredict the service lifetime.

2.4.3 Hydrolytic Degradation of PET GT Yarns
This subtask is being used to evaluate the hydrolytic degradation of eight types of PET
GT yarns in water baths at elevated temperatures for varying times.  Immersed samples
are periodically removed and tested for their retained  strengths and elongation.  These
values are then compared to the unaged strength and elongation for a percent retained
value, thereby indicating the degree of degradation.
                                      2-28

-------
               110
ro

ro
CD
           T3
            
-------
3

-------
Furthermore, by incubation at several elevated temperatures (65, 55 and 45°C are used
in this study), an Arrhenius plot can be developed.  Using the slope of the resulting
curve, extrapolation down to site-specific temperature can be made resulting in an
estimate of the service lifetime of these materials based upon a hydrolytic degradation
criterion.

A separate bath is required for each temperature incubation.  Samples are periodically
removed and tested in tension.  Some of the resulting data are presented in Figure
2-16.  Each point is the average of five replicate tests.

This project is projected to continue for an additional three-year period, thereby
achieving an estimated five-year maximum incubation time. As with the oxidation study
previously described, the reason for the lengthy duration is that excessively high
incubation temperatures may unrealistically bias the predicted service lifetime.

2.5 Prediction of GM Service Lifetime
One of the  most frequently asked question involving any type of GM is, "how long will it
last"? Since HOPE is the type of GM  most commonly used in waste containment
systems, it  is the focus of this task.  The steps involved in the task are as follows:

  • understand the mechanisms that are involved in the degradation process;
  • simulate the application(s) in the laboratory as closely as possible;
  • perform the incubations under the simulated conditions at elevated temperatures
    (at least three and preferably four temperatures);
  • remove the samples periodically and test them for changes from  their as-received
    properties; and
  • perform Arrhenius modeling to arrive at an estimated lifetime for the site specific
    temperature.

A detailed discussion of this task is  presented in Appendix  B of the report.  A brief
summary is given below.

2.5.1  Degradation of HOPE GMs
HOPE GMs are formulations consisting of PE resin (« 97%), carbon black (« 2%), and
antioxidants (« 1%). The long-term aging  process involves three discrete stages, see
Figure 2-17(a):

  • depletion time of antioxidants;
  • induction time; and
  • time to reach a specified reduction  in the value of a significant engineering
    property, e.g., elongation, modulus, strength, etc.; for  the purposes of this task, the
    numeric value of the specified  property reduction is taken as 50%.
                                      2-31

-------
               110
           •a
           a>
            co
           a:

           •a
           to
           o
100
                                 4      6       8      10


                                    Aging Time (month)
                                            12
14
               110
                80
                                 4      6       8      10


                                     Aging Time (month)
                                            12
14
Figure 2-16.  Behavior of PET yarns, sample A, after water incubation.
                                      2-32

-------
       100
     •a
     a>
     g
     'to
     •£
     a:
     a>
     a.
     o
        50
                       Various stages
                  -A
A = antioxidant depletion time
B = induction time
C = 50% property degradation
   time (i.e., the "halflife")
         100
      •a
      a>
      a:
      O)
      Q.
      O
          50
                                 Aging Time (log scale)
                  (a) The various stages of HOPE geomembrane aging
                                     TS
                            Aging Time (log scale)
                                                         '55
             (b) Degradation with stage "C" under varying temperature
Figure 2-17. Aging and degradation behavior of HOPE (and other polyolefins)
             over time under elevated temperature incubation.
                                         2-33

-------
These three stages of degradation are shown conceptually on Figure 2-17(a).  Although
not evaluated herein, this type of generalized behavior is also characteristic of low
density PE and flexible PP GMs.

The antioxidants are extremely important to the aging process,  since they react with
oxygen diffusing into the polymer structure and thereby inhibit oxidation from occurring.
When depleted at the end of Stage A (as indicated by zero oxidative induction time
(OIT) using a differential scanning calorimetry test), the induction time stage begins.

The induction time represents a time period required  to initiate a measurable amount of
oxidation-induced chain scission of the polymer structure, i.e., Stage B.  It is the least
understood of the three degradation mechanisms but it is clearly present.  For example,
polymers (like milk jugs) with no long-term antioxidants will not begin to  degrade
immediately.  While the relative induction time  may be short, its quantification should be
included in a lifetime assessment.

The oxidation process continues into Stage C such that engineering properties begin to
change. Typically, the break elongation will decrease, the modulus will  increase, and
the break strength will slightly increase, then decrease. In general, the yield elongation
and strength of HOPE will not show signs of change since these values  are small  in
comparison to the break properties. The above events, of course, signify that the
polymer is transitioning from a ductile to a brittle material. Embrittlement represents a
physical manifestation of the degradation process. As shown in Figure 2-17(b), the
response is strongly temperature dependent.  A 50% change in properties is usually
taken by polymer engineers as being  a significant change and is called the "halflife".  It
is arbitrarily assumed in this report to  signify the end  of the service life of the material.

2.5.2 Simulated Applications
There are a large number of GM applications that could be simulated in the incubation
process.  The applications targeted in this study are:

  •  landfill liners;
  •  surface impoundment liners; and
  •  landfill covers.

Each application is modeled in the simulation through selection of an incubation
medium, an applied stress (if any), and specific values of elevated temperatures for the
exposures. Table 2-4 provides the various simulation series that are ongoing in this
particular task. This report will focus only on Series No. Ill, which is the most important
series for the purposes of this project since it simulates the base liner of a landfill. A
series using leachate as the incubation medium was  started, but was subsequently
terminated due to leachate variation.
                                      2-34

-------
Table 2-4. HOPE GM Simulation Series.
Incubation
Series
1
II
Incubation
Method
water
(both sides)
air
(both sides)
Applied
Stress
none
none
Simulated GM Application
surface impoundments below
liquid level
landfill covers and waste pile
covers
IV
water above/air
beneath

water
(both sides)
260 kPa
(compression)

30% yield stress
(tension)	
                                                 landfills liners beneath waste
surface impoundments along
side slopes below liquid level
The incubations for Series No. Ill are performed in 20 identical chambers as shown in
Figure 2-18. Five chambers are maintained at each of four temperatures, i.e., 85, 75,
65 and 55°C. Samples are periodically removed and tested.

2.5.3  Antioxidant Depletion Time
Upon  removal of the incubated samples from the chambers shown in Figure
2-18, the samples are tested for their OIT.  Two options are available: (i) standard OIT
per ASTM D3895; and (ii) high pressure OIT per ASTM D5885.  Both of these OIT
methods utilize a calorimeter to evaluate the length of time the polymer melt can sustain
an oxygen environment.  The OIT (time in minutes) is related to the amount and type of
antioxidants that are used in the formulation to protect the resin from degradation. The
curves of Figure 2-19(a) were  generated using data obtained over a period of 24-
months.  Note the strong influence of elevated temperature on the OIT depletion times.
As shown in Figure 2-19(b), semi-logarithmic plots of the data result in straight line
relationships between OIT and incubation period.  The slopes of these straight lines  (for
extrapolation to site-specific temperature) for both OIT-tests are  shown in Figure 2-20.

Using these slopes and extrapolating down to site-specific temperature results in Table
2-5. The selection of the actual value is obviously site-specific.  However, data from
MSW landfill monitoring in Pennsylvania, California, and Florida  are now becoming
available. These data  indicate that 20°C is a typical value for the in-situ temperature of
HOPE GMs in liner systems for MSW landfills. As seen in Table 2-5, a value of 20°C
results in antioxidant depletion times for the type of GM evaluated herein of 192 years
based on standard OIT (Std OIT) tests and 196 years based on  HP-OIT tests.
                                      2-35

-------
               Insulation-
                 heat
tape — -&4
             thermocouple-

           readout box!
                               Saturated
                                 Sand
                                                  10
                                   I
                                   Load

                          Piezometer
                         . perforated steel
                           loading plate

                          geomembrane
                          sample under
                          compression
Figure 2-18.  Photograph and schematic diagram of a typical compression
           column for incubation Series No. III.
                                   2-36

-------
 E,
 H
 o
     20
                            10        15
                        Incubation Time (month)
                           20
25
                    (a) Arithmetic response curves
 E,
 H
 O
 £   2
            - -e-
55°C
65°C
75°C
85°C
                                                      O:
                           10        15        20       25
                        Incubation Time (month)

                    (b) Semi-log plotting of above curves
Figure 2-19. Standard oxidative induction time test results for Series No.
            Ill incubations on HOPE GM samples.
                                     2-37

-------
         CD

         2

         o
         '-t-t
         0)
         CL
         CD
         T3
             -1.0
             -2.0-
-3.0-
         o
         —   -4.0
             -5.0
                              A  Std-OIT y = 17.045-6798.0X RA2 = 0.953
                                                           2 = 0.943
               0.0027
              0.0028
0.0029
0.0030
0.0031
 Figure 2-20. Arrhenius plot for incubation Series No.
             compression stress).
                                         (water above/air beneath-
 Table 2-5. Extrapolation of Depletion of Antioxidants Trends to Various In-Situ
	Temperatures.	
       In-Situ Temp.
                       Std OIT
                      HPOIT
           30°C
           25
           20
           15
           10
            5
                        90 yrs.
                       130
                       192
                       286
                       432
                       663
                       89 yrs.
                      131
                      196
                      296
                      455
                      709
 2.5.4  Induction Time
 Stage B in Figure 2-17(a) represents the time that it takes an unstabilized polymer (i.e.,
 one with no antioxidants) to manifest a measurable amount of chain scisson. Hence, to
 evaluate this stage it would be appropriate to select a PE material with a minimum
 antioxidant content and monitor its engineering properties over time to determine the
 induction time.

 Milk and water containers represent commercial HOPE products that do not contain
 antioxidants because of their limited shelf life. Some aged milk and water containers
                                      2-38

-------
were retrieved from the waste mass of a MSW landfill.  The age of these retrieved
containers was approximately 25 years based on the dates shown on newspaper and
canceled checks that were retrieved at the same location of the landfill. The oxidative
induction time and tensile properties of the aged samples were evaluated.  The results
were compared to those obtained from unaged containers,  i.e., purchased at a grocery
store prior to the test. The data are shown in Table 2-6(a) and (b) for water and milk
containers, respectively. For this comparison, it was assumed that the aged and
unaged containers were made using the same polymer resins and manufacturing
processes. This may or may not be the case.
Table 2-6(a).  Properties of Aged and Unaged Water Containers.
         Property
 Unaged
Container
  Aged
Container
% Change
Modulus (MPa)
Yield Stress (MPa)
Yield Elongation (%)
Break Strength (MPa)
Break Elongation (%)
   650
    25
    11
    35
  1700
   580
    24
    11
    22
   879
    nil
    nil
    nil
  -37%
  -43%
Table 2-6(b). Properties of Aged and Unaged Milk Containers.
Property
Modulus (MPa)
Yield Elongation (MPa)
Yield Strain (%)
Break Strength (MPa)
Break Elongation (%)
Unaged
Container
550
24
11
22
990
Aged
Container
507
22
11
14
730
% Change
nil
nil
nil
-36%
-26%
Based on this limited data for 25-year old HOPE containers, and assuming the aged
and unaged containers had the same initial properties, it is seen that yield stress, yield
elongation, and modulus have essentially remained unchanged in a landfill atmosphere.
Only the break properties (strength and elongation) have begun to decrease.  Thus, it is
estimated that the induction time for HOPE is on the order of 20-years.

2.5.5 Half life of Engineering Properties
Stage C in Figure 2-17(a) represents the time for a HOPE GM to reach 50% change in
its engineering properties after depletion of antioxidants and induction time occurs.  The
material properties that are being monitored in this part of the study are listed  in Table
2-7.
                                     2-39

-------
Table 2-7.  Engineering Properties Being Evaluated.
           Test	ASTM Method	Property	
         Density                    D 1505                  Crystallinity
        Melt Index                   D 1238               Molecular Weight
         Tensile                    D 638               Yield, Modulus, Break
The methodology used to estimate the halflife of properties is similar to that used to
predict the lifetime of antioxidants in the HOPE GMs, i.e., the Arrhenius model. The
properties listed in Table 2-7 were monitored over increasing incubation times at
incubation temperatures of 85, 75, 65, and 55°C. The monitoring results are evaluated
by plotting percentage of the original engineering property remaining at a given
incubation time against that time, as shown in Figure 2-17(b).  The incubation time
corresponding to a percent retained of 50% is the halflife of the material at that
particular incubation temperature. The inverse of the lifetime is the reaction rate.  Once
the reaction rates at the four elevated temperatures are obtained, the data is
extrapolated by utilizing an Arrhenius plot, as shown in Figure  2-21.  Subsequently, the
reaction rate at a lower site-specific temperature, such as 20°C, can  be predicted. The
estimated time to  reach halflife of the property can be calculated as the inverse of the
reaction rate at this temperature.

Since the current test results of the incubations in Table 2-4 have not shown reduction
in material properties in the majority of the incubated samples, the halflife of the GMs
cannot be evaluated based on actual test data generated in the course of this study.
Thus, Figure 2-21 presents no actual data.  In order to estimate the potential halflife,
data from published literature are utilized.  Viebke et al. (1994) found that the activation
energy of the degradation mechanism of a unstabilized PE pipe is 80 kJ/mol.  (This
represents the slope of the Arrhenius plot shown in Figure 2-21). During the incubation
process, the pipe was exposed to water inside and circulating  air outside at constant
temperatures ranging from 70 to 105°C.

Using the Viebke et al. (1994) data, halflife can be estimated using Equation 2-7.  (Note
that the gas constant R = 8.314 J/mol and Rr represents the reaction rate from Figure
2-21  at the temperature indicated).


                        Rr@115 _c  R  Lns+273  20+273
                         Rr@20
                                      -80,000r  1	1_~
                          Rr@115     8.314  Ls88 293.
                           Rr@20
                                    e
                                      2-40

-------
                           R
                             r@20
           Rr.
           R
           R
           R
   85
 r-75
 r-65
 r-55
    "CD
     c
     o
     ti
     03
     CD
     ce:
R
 r-20
Ea/R
, I I I
, I I I
, I I I
, I I I
I I I I
f






                          1/85  1/75   1/65   1/55
                                      1/Temperature
                                                             1/20
Figure 2-21.  Arrhenius plot to analyze data and extrapolate to site-specific
             temperature.
Since half life at 115°C is 90 days which is 3027 times faster than the incubation
temperature at 20°C, the halflife at 20°C will be:
                                   = (90)(3027)
                                    = 272,000 days
                                     = 746 years
                                                          (Eq. 2-8)
This value represents the halflife of the engineering property monitored within Stage C
of the overall lifetime as illustrated in Figure 2-17.

2.5.6  Summary of Lifetime Prediction
Using the conceptual behavior model shown in Figure 2-17(a), the lifetime of a GM
consists of three-stages; antioxidant depletion, induction time, and halflife of
engineering  properties. For the 1.5-mm thick HOPE GM being evaluated in this study
                                      2-41

-------
under simulated landfill conditions, Table 2-8 represents the current best-estimate of the
lifetime prediction value.

Table 2-8. Estimated Lifetime of HDPE GM Being Evaluated in this Study.
Stage
A
B
C
Total
Description
Antioxidant Depletion
Induction Time
Halflife of Engineering Property
Lifetime Estimate
Duration (years)
200
20
750
970
Based on the methodology presented herein, the estimated service lifetime of a 1.5-mm
thick HDPE GM under the simulated test conditions is on the order of 1,000 years.  Note
that the existence of wrinkles will reduce this estimated service lifetime. No attempt has
been made for this report to estimate the degree to which wrinkles will reduce the
service lifetime. The amount remains for further research. Also remaining for further
research is an investigation as to the lifetime of GMs other than HDPE.

2.6 References
ASTM D638, Test Method for Tensile Properties of Plastics.
ASTM D1238, Test Method for Flow Rates of Thermoplastics by Extrusion Plastometer.
ASTM D1505, Test Method for Density of Plastics by the Density-Gradient Technique.
ASTM D2412, Determination of External Loading Characteristics of Plastic Pipe by
   Parallel-Plate Loading.
ASTM D3895, Test Method for Oxidative-lnduction Time of Polyolefins by Differential
   Scanning Calorimetry.
ASTM D5208, Practice for Operating Fluorescent UV and Condensation Apparatus for
   Exposure of Photodegradable Plastics.
ASTM D5885, Test Method for Oxidative Induction Time of Polyolefin Geosynthetics by
   High Pressure Differential Calorimetry.
Clough,  R. W. and Duncan, J. M. (1971), "Finite Element Analysis of Retaining Wall
   Behavior, " Journal of Soil Mechanics and Foundation Engineering Division, ASCE,
   Vol. 97, SM12, pp.  1657-1673.
Duncan, J. M. and Chang, C. Y.  (1970), "Non-Linear Analysis of Stress and Strain in
   Soils, " Journal of Soil Mechanics and Foundation Engineering Division, ASCE, Vol.
   96, SMS, pp. 1629-1653.
Eith, T. A. and Koerner, G. R. (1997), "HDPE GM Waves in the Field," Proceedings
   GRI-11 on Field Installation of GMs, Geosynthetic Research Institute, Philadelphia,
   PA, pp.  101-114.
Giroud, J. P. and Bonaparte, R. (1989),  "Leakage Through Liners Constructed with
   Geomembranes. Part II:  Composite Liners," Geotextiles and Geomembranes, Vol.
   8, No. 2, pp. 77-111.
                                     2-42

-------
Giroud, J. P. and Houlihan, M. F. (1997), "Design of Leachate Collection Layers, "
  Proceedings Sardinia '95, CISA, Cagliari, Italy, pp. 613-640.
Howard, A.  K. (1977), "Soil Reaction for Buried Flexible Pipe, " Journal of Geotechnical.
  Engineering Division, ASCE, Vol. 103, No. GT1, January, pp. 33-43.
Hsuan, Y. G. and Koerner, R. M. (1993), "Can Outdoor Degradation be Predicted by
  Laboratory Accelerated Weathering?" Geotechnical Fabrics Report, Vol. 11, No. 8,
  pp. 12-16.
Hsuan, Y. G., Koerner, R.  M. and Soong, T.-Y. (1994), "Behavior of Partially Ultraviolet
  Degraded Geotextiles," Proceedings 5th International Conference on Geosynthetics,
  Singapore, Vol. 3, pp. 1209-1212.
Koerner, G. R.  and Koerner, R. M. (1995), "Temperature Behavior of Field Deployed
  HOPE Geomembranes," Proceedings Geosynthetics '95, IFAI,  pp. 921-937.
Koerner, R. M., Wilson-Fahmy, R. F. and Narejo, D. (1996), "Puncture Protection of
  Geomembranes Part III: Examples, " Geosynthetics International, Vol. 3, No. 5, pp.
  655-676.
Moser, A. P. (1990), Buried Pipe Design, McGraw-Hill, New York, NY.
Narejo, D., Koerner, R. M. and Wilson-Fahmy, R. F. (1996), "Puncture Protection of
  Geomembranes Part II:  Experimental," Geosynthetics International, Vol. 3, No. 5,
  pp. 629-653.
Richardson, G. N. (1996),  "Field Evaluation of Geosynthetic Protection Cushions,"
  Geotechnical Fabrics Report, Vol. 14,  No. 2, pp. 20-25.
Soong, T.-Y. and Koerner, R. M.  (1998), "Modeling and Extrapolation of Creep Behavior
  of Geosynthetics, " Proceedings 6th International Conference on Geosynthetics, IFAI,
  St. Paul,  MN, pp. 707-710.
Soong, T. -Y. and Lord, A. E., Jr. (1998), "Slow Strain Rate Modulus via Stress
  Relaxation Experiments, " Proceedings 6th International Conference on
  Geosynthetics,  IFAI, St. Paul, MN, pp. 711-714..
Spangler, M. G. (1971), Soil Engineering, 2nd Ed.,  International Textbook Co.,
  Scranton, PA.
Viebke, J., Ifwarson, M. and Gedde, U. W. (1994),  "Degradation of Unstabilized
  Medium-Density Polyethylene Pipes in Hot-Water Applications, " Polymer
  Engineering and Science, Vol. 34, No. 17,  pp. 1354-1361.
Watkins, R. K.  (1987), Structural Performance of Perforated and Slotted HOPE Pipes
  under High Soil Cover, Report to King County Solid Waste, Seattle,  WA.
Watkins, R. K.  and Spangler, M. G. (1958), "Some  Characteristics of the Modulus of
  Passive Resistance of Soil - A Study in Similitude, " Highway Research Board
  Proceedings, Vol. 37, pp. 576-583.
Wilson-Fahmy, R. F. and Koerner, R. M.  (1994),  Finite Element Analysis of Plastic Pipe
  Behavior in Leachate Collection and Removal System, GRI Report #12, Geosynthetic
  Research Institute, Drexel Univ., Philadelphia, PA.
Wilson-Fahmy, R. F., Narejo,  D. and Koerner, R. M. (1996), "Puncture Protection of
  Geomembranes Part I:  Theory," Geosynthetics  International, Vol. 3, No. 5, pp. 605-
  628.
                                     2-43

-------
                                  Chapter 3
     Slope Stability of Full-Scale Field Test Plots Containing GCLs
                     to Simulate Final Cover Systems

3.1  Introduction
GCLs consist of a thin layer of bentonite encased between two GTs or mixed with an
adhesive and attached to a GM.  GCLs  are a relatively new type of liner material, having
first been used in a landfill in 1986 (Schubert, 1987).  Although GCLs are relatively new,
their use in waste containment facilities  has increased steadily because of the extremely
low hydraulic conductivity of bentonite, the low cost of GCLs, the ease and speed of
installation compared to CCLs, and the  low volume occupied by GCLs compared to
much  thicker CCLs.

GCLs enjoy several favorable hydraulic  characteristics,  including self-healing properties
(Shan and Daniel, 1991; Estornell and Daniel, 1992), ability to withstand differential
settlement (Koerner et al., 1996;  Lagatta et al., 1997), ability to self-heal after
desiccation (Boardman and Daniel, 1996), and resistance to the potentially damaging
effects of freezing temperatures (Hewitt  and Daniel,  1997; Kraus et al.,  1997).
Bentonite is subject to increases in  hydraulic conductivity caused by chemical
alterations, particularly when calcium is  leached from cover soils under conditions of low
overburden stress, such as in secondary containment linings (Dobras and Elzea, 1993)
or final cover systems (James et al., 1997).  In liner systems, where the overburden
stress on the GCL is much  greater,  alterations in hydraulic conductivity, if any, tend to
be small for GCLs permeated with actual landfill leachate (Ruhl and Daniel, 1997).

The favorable hydraulic properties of GCLs are tempered by the low shear strength of
hydrated bentonite (Mesri and Olson, 1970;  Olson, 1974; Gilbert et al.,  1996; Stark and
Eid, 1997; Stark et al., 1998; Fox et al.,  1998) and low bearing capacity of hydrated
GCLs (Koerner and Narejo, 1995; Fox et al., 1996).  When bentonite is hydrated and
sheared, angles of internal  friction as low as 5 to 10° may result.  Because bentonite is
so well known for its low shear strength, caution is appropriate when employing
materials such as GCLs that contain bentonite on slopes.

A shearing failure involving a GCL can occur at three possible locations (Figure 3-1):  (1)
the external interface between the top of the GCL and the overlying material (soil or
geosynthetic); (2) internally within the GCL; and (3) the  external interface between the
bottom of the GCL and the  underlying material (soil or geosynthetic). If failure is
internal, the failure may be  bentonite-to-bentonite (e.g.,  at the mid-plane of the GCL), or
it may be at the internal interface between the bentonite and either the upper or lower
geosynthetic component (if present).
                                      3-1

-------
                    Potential Failure Surfaces:
                    1.  Interface between upper surface
                        of GCL and overlying material
                    2.  Internal failure within GCL (can
                        be within bentonite or at the
                        internal interface between
                        bentonite and a geosynthetic)
                    3.  Interface between lower surface
                        of GCL and overlying material
       GCL
                          Material Overlying GCL
                         (Material Underlying GCJJl
 i -
/•—•
 1
Figure 3-1.  Potential failure surfaces for a GCL.
Current engineering design practice is to establish appropriate internal and interface
shear strength parameters for design of GCLs on slopes using direct shear tests on
300-mm square test specimens, and to employ traditional limit equilibrium techniques
for analyzing slope stability.  However, the low shear strength of bentonite, the limited
number of laboratory test results available, the inherent limitations of laboratory direct
shear tests, the uncertainty over use of peak versus residual shear strength, the relative
newness of GCLs, and the lack of field experience with GCLs all lead to questions
about the long-term stability of GCLs on relatively steep slopes.

To provide field-scale data on the stability of GCLs on slopes, field test plots were
constructed. It was recognized that it would not be possible to construct and instrument
a full-scale landfill lined with GCLs, but it was possible to construct and instrument
prototype landfill covers.  Therefore, test plots were constructed to evaluate the stability
of field test plots containing GCLs.

This chapter summarizes the test plots, data collected from the test plots, and
conclusions from the test plots.  Appendix D provides additional details.
                                      3-2

-------
3.2 Background on GCLs

3.2.1  Introduction
A GCL consists of approximately 5 kg/m2 of sodium bentonite sandwiched between two
GTs or attached to a GM with an adhesive. Figure 3-2 shows two general types of
GCLs. Both types were used for this research.
(A)

Geotextile-Encased GCL

(Bentonite Sandwiched between Two Geotextiles)
Geotextile

Bentonite |



Geotextile

(B) Geomembrane Supported GCL
(Bentonite Glued to Geomembrane)

Bentonite |||

^^^^^l C-ienmemhrane


J

Figure 3-2. Two general types of GCLs.
The specific types of GCLs that were available when this project was initiated in 1994
are shown in Figure 3-3. Only the unreinforced, GT-encased, GCL was not included in
the field test plots (all the others were included). The unreinforced, GT-encased GCL
was omitted because this type of GCL is not intended for relatively steep slopes - a
reinforced, GT-encased GCL would be recommended instead.
                                     3-3

-------
  Reinforced. Geotextile-Encased. Needlepunched GCL
    Woven or Nonwoven
    Geotextile ^^
Needlepunched
Fibers \












Sc


dii


m


Be


nt(


nil


B




1




                  Nonwoven Geotextile
         Unreinforced. Geotextile-Encased GCL
                            Woven Geotextile



Sodium Bentonite Mixed
with an Adhesive



        Woven Geotextile
  Reinforced. Geotextile-Encased. Stitch-Bonded GCL
                           Woven Geotextile






Sodium Be
with an




ntonite Mixed
Adhesive






          Woven Geotextile
                           Sewn Stitches
      Unreinforced. Geomembrane-Supported GCL

Sodium Bentonite Mixed
with an Adhesive


        Geomembrane
Figure 3-3.  Cross sections of GCLs available at the time of this study.
                              3-4

-------
The specific product types included in this testing program were as follows.  Bentomat®
and Bentofix® are reinforced, GT-encased, needlepunched GCLs that consist of dry
sodium bentonite sandwiched between two GTs.  One GT is nonwoven while the other
GT can be either woven or nonwoven.  The entire assembly is needle punched
together.  Bentomat ST, which contains nonwoven and woven GTs on the two surfaces
of the GCL, was used in the research program. Bentofix NS (also with nonwoven and
woven GT components on the two surfaces of the GCL) and Bentofix NW (nonwoven
GTs on both surfaces) were also used in the  research program.  Claymax®  5000SP
                                                                 /5\
was the GT-encased,  stitch-bonded GCL used in the test plots. Gundseal  was the
GM-supported GCL employed in the testing program.

3.2.2  Advantages and Disadvantages of GCLs
GCLs enjoy numerous advantages and disadvantages (Daniel and Boardman, 1993),
which are summarized here.  The principal advantages of GCLs (as compared with
CCLs) are favorable cost, convenience of installation, and outstanding hydraulic
properties.

The installed cost of GCLs is typically equal to or less than that of CCLs, particularly if
clay must be shipped from off site or if bentonite must be blended with soil to form the
clay liner material. In  addition, a GCL occupies less volume than  a CCL, which can
result in more landfill volume becoming available for waste disposal when a GCL is
used.  Because GCLs can be installed far more rapidly than CCLs, construction time is
less with GCLs, which can significantly reduce overall construction costs.

GCLs are convenient for owners of waste containment facilities because they can be
bid with the other geosynthetic components and installed by the same organization that
installs the other geosynthetics.  The cost of a GCL is more predictable than that of
CCLs, and the much more rapid installation time of GCLs is usually attractive to the
project owner.

The other major advantage of GCLs is their favorable hydraulic characteristics. The
hydraulic conductivity  of GCLs is typically in the range of 1 to 5 x 10~9 cm/s,  which is one
to two orders of magnitude lower than the typical hydraulic conductivities assumed for
CCLs. This makes the hydraulic performance of GCLs potentially superior to CCLs.  In
addition, GCLs have excellent self-healing properties (Shan and Daniel, 1991; Estornell
and Daniel, 1992), excellent ability to withstand differential settlement (Koerner et al.,
1996; Lagatta et al., 1997), ability to self-heal after desiccation (Boardman and Daniel,
1996), and resistance to the potentially damaging effects of freezing temperatures
(Hewitt and Daniel, 1997; Kraus et al., 1997).
                                     3-5

-------
GCLs also suffer from several disadvantages. Perhaps the three most significant
disadvantages of GCLs are the low shear strength of hydrated bentonite, vulnerability to
chemical alterations, and the thinness of GCLs.

Bentonite is famous among geotechnical engineers and geologists for its very low
strength when hydrated. Potential problems arise when GCLs are placed on slopes.
Also, bentonite may be for locally squeezed (thus thinning the GCL) by sharp stones or
an uneven subgrade. It is generally accepted that  GCLs can be safely placed on landfill
cover slopes inclined at 101-1:1 V (5.7°) or flatter without any need for internal
reinforcement or slope stability analysis. Steeper slopes may or may not be stable,
depending on the type of GCL and specific conditions relevant to the slope.

The second disadvantage of GCLs relates to potential chemical alterations that could
increase hydraulic conductivity (Ruhl and Daniel, 1997).  Particular concern exists for
landfill covers containing calcium-rich soils (James et al.,  1997).

A third concern about GCLs is related to the thinness of GCLs.  GCLs are nominally
about 10 mm thick. Like any thin liner, GCLs are vulnerable to puncture, e.g., as
described for a case involving accidental puncture  of a GM/GCL composite liner by a
piece of maintenance equipment (Daniel and Gilbert, 1996).  The thinness of GCLs also
makes them less able to adsorb and attenuate chemicals than much thicker CCLs, and
less resistant to chemical diffusion than much thicker CCLs (Foose et al., 1996; 1999).

3.2.3  Shear Strength of GCLs
One disadvantage of GCLs is the low shear strength of hydrated bentonite. Shear test
data on  unreinforced, hydrated GCLs result in friction angles of about 10° at
intermediate normal stress.  In EPA workshops on  GCLs, the shear strength of the
bentonite in GCLs, which controls the internal shear strength of unreinforced GCLs, was
cited as a primary technical concern in the use of GCLs in waste containment systems
(Daniel and Boardman,  1993). The main factors affecting the internal shear strength of
GCLs include the magnitude of normal stress, water content  of the bentonite, type of
hydrating liquid, rate of shearing, reinforcement, amount of deformation, and effects of
seismic  loading.  These factors are reviewed  below.

3.2.3.1  Magnitude of Normal Stress
The classical Mohr-Coulomb failure criterion for the shear strength of soil is:

      T  = c +  a tancj)                                                    (Eq. 3-1)

where: T is the shear stress (Pa), c is the cohesion (Pa), a is  the normal stress (Pa), and
c|) is the angle  of internal friction (degrees). The concept is illustrated in Figure 3-4. The
                                      3-6

-------
ideal Mohr-Coulomb failure envelope is linear.  However, the relationship between
shear stress and normal stress for bentonite is not always linear (Figure 3-5).
 tn
 -—
 C/3
 l_
 03
 CD
 .C
 C/D
                                 Normal Stress (a)


      Figure 3-4. Mohr-Coulomb failure envelope.
     w
     w
    -—
    OD
    s_
    CD
    CD
    .C
    C/)
                                   Assumed Linear Failure Envelope
                                   Tangent to Actual Failure Envelope
t
Linear Failure Envelope
Gives Correct Strength for
the Normal Stress of Interest

1
c
\
I
'
^^^X Failure Er
^^/yC '
^r >^ ^^*^^L Assumed Secant Failure Envelope
^^ yX^ 1^ that Gives Correct Shear Strength
X >X for the Normal Stress of Interest by
1 / ' Assuming that c = 0
{/ \
t
Normal Stress
of Interest
Normal Stress
ivel
(a)
      Figure 3-5. Curved Mohr-Coulomb failure envelope.
                                        3-7

-------
The cohesion of the GCL can be very important, particularly for internally reinforced
GCLs (i.e., needlepunched or stitch-bonded GCLs) employed in situations with low
normal stress, such as landfill covers. Work is underway to correlate peel strength with
cohesion in the  hope of identifying a relatively simple index test that will correlate with
cohesion.

3.2.3.2  Water Content
The shear strength of bentonite is sensitive to water content.  The angle of internal
friction decreases with increasing water content.  For example, shear tests that were
performed on an unreinforced GCL at The University of Texas showed that at a water
content of 20%, the angle of internal friction was 22°, but when the water content was
increased to 50%, the friction angle of the unreinforced GCL decreased to 7° (Daniel et
al., 1993).  Hydrated bentonite is significantly weaker than dry bentonite.

When hydrated  GCLs are tested in direct shear boxes, the GCL may either be hydrated
at low normal stress and then consolidated in the shear box to the desired  normal stress
for shear testing, or the GCL may be immediately subjected to the final normal stress
and hydrated under that stress. The recommended procedure is normally to apply
stresses and hydration water in a manner that will simulate the conditions in the field.
However, the procedure is also impacted by the practicality of testing. Because 300
mm by 300 mm shear boxes are very expensive, it is customary practice to minimize
the amount of time that the boxes are committed to any one test. To accomplish this,
the GCL is often hydrated in a separate apparatus at a comparatively low normal stress
of about 12 kPa, and then transferred to the shear box for consolidation and shearing.

3.2.3.3  Type of Hydrating Liquid
The type of hydrating liquid relates to the bentonite particle's adsorption capability. This
is evidenced by both hydraulic conductivity and shear strength, the latter being the
focus of this study. The greater the adsorptive capability of the hydrating liquid, the
lower the shear strength of the bentonite. The GCL's  shear strength should be
evaluated with the site-specific liquid that will hydrate the bentonite.

3.2.3.4  Rate of Loading
The rate of loading of GCLs affects the shear strength of the GCL.  The general
experience with bentonite is the slower the loading, the lower the internal shear strength
of the GCL (Daniel et al., 1993). Thus,  care should be taken in testing GCLs so as not
to shear the  GCL too quickly.

3.2.3.5  Reinforcement
Many commercial GCLs are reinforced  to enhance the internal shear strength of the
GCL. The reinforced GCLs used in the field test plots included Bentomat, Claymax
                                       3-8

-------
500SP, and Bentofix. When a reinforced GCL is sheared internally, the needlepunched
fibers or sewn stitches are put into tension as shearing occurs, which enhances internal
shear strength. However, there are limitations on the benefits of this reinforcing, as
discussed below in the next subsection.

3.2.3.6 Amount of Deformation
The peak shear strength is the maximum shear strength measured during shear.
Typically, however, many materials "strain soften" after the peak strength is reached.
The residual shear strength is the minimum post-peak shear stress, which typically
occurs at a very large displacement compared to the displacement at which the peak
strength is generated.  Figure 3-6 illustrates the difference between peak and residual
shear strength.
CO
-1— '
CO ;
O)
c
CD
.c
CO
Stress-Strain Curve
^ A (Peak)
/ V^^
/ ~ • B (Residual)
Deformation
       CO
       CO
      CO
       O)
       _
       CO
       CD
       .C
       CO
 Mohr-Coulomb
Failure Envelopes
                                 Peak
                                                 Residual
                            B
                         Normal Stress, a
Figure 3-6. Peak and residual shear strength.
                                      3-9

-------
If a reinforced GCL is loaded to very large shearing displacements, reinforcing fibers
may pullout from one or both of the GTs, break, or creep. If the reinforcing fibers fail,
the strength of the reinforced GCL may be about the same as that of an unreinforced
GCL. The key issue is how much deformation will actually occur in the field, and
whether there is a risk of residual conditions actually developing.

3.2.3.7  Seismic Loading
Data on effects of cyclic loading on the internal shear strength of GCLs is very limited.
The tests indicate that cyclic loading causes a slight increase in the internal shear
strength of dry,  unreinforced GCLs.  The increase in strength is the result of a slight
densification of the dry bentonite during cyclic loading.  However, the tests indicate that
unreinforced, saturated bentonite undergoes a reduction in strength from cyclic loading.
The reduction increases with increasing number of cycles of loading.  Results are
described by Lai et al. (1998).

3.2.4 Interface Shear Strength
The interface shear strength of the GCL with an adjacent material can be the most
critical (i.e., lowest) shear strength. The shear strength of GCLs at interfaces can be
affected by several factors.  One factor is the interfacing materials.  For example, the
friction angle between a GCL and subsoil will be different from the friction angle
between a GCL and a GM.  Also, a textured GM will typically have a higher interface
friction angle with  a  GCL than a smooth GM.

Another factor affecting the  interface shear strength of GT-encased GCLs is the
different types of GTs  used  in making GCLs. An interface involving a woven GT may
have a lower shear strength than an interface involving a nonwoven GT.  A third factor
is the degree of hydration of the bentonite and its  potential mobility through the GT
components of the GCL.  If  hydrated bentonite can swell through the GTs, the  hydrated
bentonite may "lubricate" the interface with an adjacent material. In addition, the level of
normal stress and amount of deformation can  influence interface shear strengths.
Distortion before or during deformation could also be a factor.  Finally, the amount of
deformation can influence interface shear strength. The large-displacement interface
shear strength is generally less than the peak value.

3.3 Field Test Plots
Fourteen field test plots were constructed in Cincinnati, Ohio.  The layout of the plots is
shown in Figure 3-7. Five plots (plots A-E) were constructed on a 3H:1V slope, and
nine plots (plots F to L, N, and P) were built on a 2H:1V slope.  The 3H:1V slopes were
part of an actual final cover  system over a closed  section of the landfill, and the 21-1:1 V
slopes were along an excavated slope on one side of a large pit located adjacent to the
landfill.  Plot P was built where plot G had originally been located, after a slide
                                       3-10

-------
                                Geomembrane Placed over GCL
                                -*	*~
                 Crest of Slope
  3H:1V Plots:
                   Toe of Slope

A

B

C

D

E
f                                      Bentomat  f   Bentofix   f
                                         ST     I     NS     I
                                Gundseal      Claymax      Gundseal
                                (Bentonite       500SP       (Bentonite
                                  Up)                      Down)
                                           Note:

                                           All Test plots Were Nominally
                                           9 m Wide and either 20 m Long
                                           (21-1:1 V Slopes) or 29m Long
                                           (31-1:1 V Slopes)
Geomembrane Placed
     over GCL
                                                    No Geomembrane
                       Geomembrane
                        Placed over
                           GCL
          Crest of Slope •
  2H:1V Plots:
           Toe of Slope —•»
G,
P

H

I

J

K

L

M

N
                        Gundseal
                        (Bentonite
                          Up)
         Claymax
          500SP
Bentomat
   ST
                                           Bentofix
                                             NW
Bentofix
  NW
Bentofix
  NS
                          Bentomat ST (Plot G);
                          Gundseal (Plot P) with
                           Bentonite Up
                            Claymax
                            500SP
                   No GCL
                   (Erosion
                   Control
                     Plot)
Figure 3-7.  Layout of field test plots.

-------
occurred at plot G.  An additional plot (M) did not contain a GCL; this plot was
constructed to study erosion of surface soils. Plots on the 21-1:1 V slope were nominally
20 m long while those on the 3H:1V slope were 29 m long. All plots were two GCL
panel widths (« 9 m) wide and were 0.9 m thick.

A typical cross section of a test plot is shown in Figure 3-8. Most of the test plots were
constructed with a GM overlying the GCL, which would be typical of a final cover system
for a landfill. However, GCLs are also used in final cover systems without GMs.
Hence, three plots were constructed with no GM.  The plots were drained internally
using either a GC drainage layer (GT/GN/GT system) or, for the three plots that did not
contain a GM, a sand drainage layer. The drainage layer was included to limit build-up
of water pressure in the overlying soils and, by doing so, to enhance the stability of the
cover soils.
    Geonet
   Drainage
   Material
                 Geomembrane
                               GCL
Figure 3-8. Typical cross-section for 3H:1V test plot.
Most of the GCLs were intentionally placed with bentonite in direct contact with the
subgrade soil on the assumption that the bentonite would absorb water from the
subgrade soils and hydrate in this manner. Laboratory experiments have shown that
GCLs placed on damp or moist soils will hydrate in this fashion (Daniel et al., 1993). It
                                      3-12

-------
was desired that the bentonite hydrate in the test plots so that the most critical condition
(i.e., hydrated bentonite) could be investigated.

3.3.1  Rationale for 2H:1V and 3H:1V Slopes
The rationale for selecting the 2H:1V and 3H:1V slope inclinations was as follows.  The
31-1:1 V slope was selected to be representative of typical final cover systems for landfills
in use at the time the study was initiated (1994).  In order to confirm that GCLs are safe
against internal failure on 3H:1V slopes, it must be shown that they are not only stable,
but are stable with an adequate factor of safety.  For an infinite slope consisting of
cohesionless interfaces with no seepage, the factor of safety (FS) is:

      FS  = tan(<|>) / tan(p)                                              (Eq. 3-2)

where: (j> = the angle of internal friction (degrees); and p = the slope angle.  Engineers
typically design permanent slopes for landfills to have a minimum factor of safety for
static loading of about 1.5.  The ratio of tan pfor a 2H: 1V slope to tan p of a 3H:1V
slope is  1.5. Subject to the assumptions listed above, if a GCL is theoretically stable on
a 2H:1V slope (i.e., FS > 1.0), the same GCL is demonstrated to be stable on a 31-1:1 V
slope with  FS >  1.5.  Therefore, the 2H:1V slopes were chosen to demonstrate internal
stability of GCLs on in 3H:1V slopes with FS > 1.5. However, it was recognized that
constructing 2H:1V slopes was pushing the GCLs to (and possibly beyond) their limits
of stability, if not with respect to the internal shear strength the GCLs, then certainly with
respect to  the various interfaces within the system.

3.3.2  GCLs
Table 3-1 summarizes the type of GCL installed in each plot, the targeted and actual
inclinations of the slopes, and the dimensions and cross  section of each test plot.
Bentofix and Bentomat are GT-encased,  needle-punched GCLs. Bentofix NS and
Bentomat ST consist of bentonite sandwiched between nonwoven  and woven GTs.
Bentofix NW consists of bentonite sandwiched between two nonwoven GTs.  One
surface of  Bentofix is heat burnished (the side with the woven GT for Bentofix NS  and
one of the  sides with a nonwoven GT for Bentofix NW). As indicated in Table 3-1,  either
the woven or nonwoven GT faced upward, depending on the GCL and test plot. Which
GT component (woven or nonwoven) was in contact with a textured GM turned out to
be very important. Figure  3-9A depicts the cross section for the two test plots (B and G)
in which the woven GT component of the GCL faced  upward. Figure 3-9B illustrates the
cross section in  Plots D and N with the nonwoven GT component facing upward.  Plots I
and L, which contained Bentofix NW, also had a nonwoven GT component of the GCL
facing upward, but the lower GT component of this GCL was also a nonwoven GT.
                                      3-13

-------
Table 3-1.  Information on Test Plots.
Test
Plot
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
Type of
GCL
Gundseal
Bentomat ST
Claymax
500SP
Bentofix NS
Gundseal
Gundseal
Bentomat ST
Claymax
500SP
Bentofix NW
Bentomat ST
Claymax
500SP
Bentofix NW
Erosion
Control
Bentofix NS
Gundseal
Nominal
Slope
(H:V)
3:1
3:1
3:1
3:1
3:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
Target
Slope
Angle
(deg)
18.4

18.4
18.4
18.4
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
Actual
Slope
Angle
(deg)
16.9
17.8
17.6
17.5
17.7
23.6
23.5
24.7
24.8
24.8
25.5
24.9
23.5
22.9
24.7
Actual
Slope
Length
(m)
28.9
28.9
28.9
28.9
28.9
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
Actual
Plot
Width
(m)
10.5
9.0
8.1
9.1
10.5
10.5
9.0
8.1
9.1
9.0
8.1
9.1
7.6
9.1
9.0
Cross Section
(Top to Bottom)
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GDL/GCL
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
Soil/GT/Sand/GCL
Soil/GT/Sand/GCL
Soil/GT/Sand/GCL
Soil
Soil/GDL/GM/GCL
Soil/GDL/GM/GCL
GCL Side
Facing
Upward
Bentonite
Woven GT
Woven GT
Nonwoven GT
GM
Bentonite
Woven GT
Woven GT
Nonwoven GT
Woven GT
Woven GT
Nonwoven GT
No GCL
Nonwoven GT
Bentonite
GCL Side
Facing
Downward
GM
Nonwoven GT
Woven GT
Woven GT
Bentonite
GM
Nonwoven GT
Woven GT
Nonwoven GT
Nonwoven GT
Woven Gt
Nonwoven GT
No GCL
Woven GT
GM
Key:  GDL = geocomposite (geotextile/geonet/geotextile) drainage layer; GM = textured GM; GT = geotextile; GCL = geosynthetic clay liner.

-------
Claymax 500SP consists of bentonite mixed with an adhesive and encased between
two woven slit-film GTs that are stitched together. Lines of stitching are spaced 100
mm apart. The two GT components are identical in this type of GCL.

Gundseal is an unreinforced GCL consisting of bentonite mixed with an adhesive and
bonded to a GM. The GM component was a textured, 0.8-mm thick, HOPE material.
With this product, the exposed surface of the bentonite component was covered with a
thin GT called a "spider net," which is used to help prevent loss or dislodgment of loose
particles of bentonite from the GCL during storage, transportation, and installation.  The
spider net was incorporated in all plots with Gundseal, except for plot P, which did not
contain the spider net.  Figure 3-10 shows the two uses of this type of GCL, with the
bentonite component either facing upward or downward.  When the bentonite was
facing upward (as in plots A, F, and P), the bentonite was encased between two GMs.
In this condition, the bentonite was expected to remain essentially dry,  except for spot
hydration along the overlap or near any imperfections in the overlying GM or GM
seams.  When the bentonite faced downward (as in plot E), the bentonite was expected
to hydrate by absorbing moisture from the subsoil.

3.3.3 Other Materials
A 1.5-mm-thick textured HOPE GM was used for the GM component that was placed on
top of the GCL in all test plots except J, K, and L (which contained no GM).  The GC
drainage layer consisted of two nonwoven GTs heat-bonded to both sides of a GN. The
cover soil was a silty, clayey sand  obtained from an on-site borrow source.

3.3.4 Construction
Construction of the test plots began on November 15, 1994 and was completed on
November 23, 1994. However, plot P was constructed on June 15, 1995.

The test plots were first graded to provide a smooth subgrade, as shown in Figure 3-11.
Next, geosynthetics were installed by pulling them down from the crest of the slope
(Figure 3-12). Cover soil was placed by starting at the bottom of  the slope and
working upslope (Figure 3-13). All test plots, except plot  M, were then  covered with an
erosion control material.  The 3H:1V test plots are shown in Figure 3-14, and the 2H:1V
test plots are shown in Figure 3-15.

In plots incorporating a GC drainage layer, the GM and GC were  extended beyond the
GCL at the toe of the slope and 1.5 m past the end of the cover soil (Figure 3-8).  For
plots constructed with a sand drainage layer (plots J, K, and  L), a piece of geosynthetic
drainage material was embedded in the sand at the toe of the slope and then extended
1.5 m beyond the end of the cover soil.
                                     3-15

-------
     (A) Woven Geotextile Interfacing with Geomembrane
                     (B)  Nonwoven Geotextile Interfacing with Geomembrane
CO

en
          Cover Soil
 Geocomposite
Drainage Layer

Textured HOPE
Geomembrane
        GCL
            Subsoil
Woven
Geotextile
                                            Bentonite
                                           Nonwoven
                                           Geotextile
                            Cover Soil
                                                                    Geocomposite
                                                                    Drainage Layer
                              Subsoil
Nonwoven
Geotextile
                                                              • Bentonite
                                                             Woven
                                                             Geotextile
    Figure 3-9.  Orientation of GCL with either woven or nonwoven GT facing upward.

-------
             (A) Plots with Bentonite Component Facing Upward
                                                         (B) Plots with Bentonite Component Facing Downward
CO
                  Cover Soil
   Geocomposite
   Drainage Layer

   Textured HOPE
   Geomembrane

GCL (Bentonite Up) {
                                                  -Bentonite
                    Subsoil
                                                 Textured HOPE
                                                 Geomembrane
                                                               Cover Soil
                                                                         Geocomposite
                                                                         Drainage Layer
                                                                    GCL (Bentonite Down) {
                                                                              Subsoil
Textured HOPE
Geomembrane

 Bentonite
   Figure 3-10.  Placement of Gundseal with bentonite facing upward or downward.

-------
Figure 3-11. Prepared surface on which 21-1:1 V test plots were constructed.
Figure 3-12. Typical installation of geosynthetics.
                                     3-18

-------
Figure 3-13. Placement of cover soil near top of slope.
Figure 3-14. View of 3H:1V test plots.
                                    3-19

-------
Figure 3-15. View of 2H:1V test plots.
All of the geosynthetic materials in each test plot were brought into their respective
anchor trenches (Figure 3-3), which were then backfilled. Because the purpose of each
test plot was to test the internal shear strength of a particular GCL, the toe of each test
plot was excavated at the completion of construction to the shape shown in Figure 3-8
so that no buttressing (i.e., passive) force could be mobilized at the toe of the slope.
Similarly, any tension in the geosynthetic components located above the GCL would
reduce the shearing stress to be carried by the internal structure of the GCL.

To prevent the development of tension  in the geosynthetic components above the mid-
plane of the GCL, components above the mid-plane of the GCL were cut as shown in
Figure 3-16.  Cutting occurred in the spring of 1995, about 5 months after construction
of the test plots.  However, the geosynthetics were not cut in plot P, which was
constructed later in the program for  the sole purpose of evaluating hydration of
bentonite encased between two GMs (cutting the geosynthetics would have provided a
pathway for water to enter the bentonite near the crest of Plot P).
                                      3-20

-------
               Geomembrane
                  Cap Strip
                                     All Geosynthetics above the
                                     Mid-Plane of the GCL Were Cut,
                                     Including the Upper Geotextile or
                                     Geomembrane Component of the
                                     GCL (If Present)
;0.5 m
                       Anchor:
                        Trench
                       Backfill ;
                                                Subsoil
                                      Geomembrane
             Geocomposite-
             Drainage Layer
Figure 3-16. Cut in anchor trench above mid-plane of GCL
3.4 Instrumentation
The field test plots were instrumented, but limited funds available for instrumentation
dictated that a simple, robust, cost-effective instrumentation program be implemented
quickly and easily. The instrumentation program had two objectives: (1) provide an
indication of the moisture conditions in the bentonite component of the GCL to verify
that hydration had occurred; and (2) provide data on the downslope displacements
occurring within the GCL.  With respect to displacement, the objective was to monitor
the shearing displacement, defined as the difference in displacement between the top
and bottom surfaces of the GCL.

3.4.1  Moisture Sensors
It was expected that GCLs placed in contact with subgrade soils would hydrate by
absorbing moisture from the subgrade. Project-specific testing indicated that
substantial hydration of GCLs occurred within 10 to 20 days for GCLs placed in contact
with the subgrade soils from the test site, even for subgrade soils compacted at a
moisture content 4 percentage points dry of the standard Proctor optimum  moisture
content. Moisture sensors were installed to verify that the bentonite did indeed become
hydrated. However, in the case of plots A,  F, and P, the bentonite component of
Gundseal was sandwiched between two GMs (see Figure 3-1OA) with the expectation
                                      3-21

-------
that the bentonite would remain dry.  For plots A, F, and P, moisture sensors were
installed to verify that the bentonite remained unhydrated.

Gypsum blocks and fiberglass moisture sensors, shown schematically in Figure 3-17,
were used to monitor in-situ moisture contents. These instruments were selected based
on low cost and the ability to use multiple instruments to provide redundancy. Gypsum
blocks were placed in subgrade soils about 50 mm below the GCL.  Fiberglass sensors
were placed in contact with the GCL, either at the interface between the GCL and
subsoil or (for bentonite sandwiched between two GMs) between the GCL and GM.
Typically, three fiberglass sensors were deployed at each test plot, near the crest,
middle, and toe of the slope, and monitored every 1 to 4 weeks. However,  16 sensors
were installed in Plot P.
        Gypsum Block
Fiberglass Moisture Sensor
                       40 mm     40 mm
            25 mm
    25 mm
Figure 3-17. Schematic diagram of moisture sensors.
3.4.2  Displacement Gauges
Displacement gauges (extensometers) were installed in each test plot (except plot P,
which was constructed only to monitor the moisture content of the bentonite) to
measure total and differential displacements in the GCL at multiple locations along the
slope. Pairs of stainless steel fish hooks were embedded into either the upper or lower
geosynthetic component of a GCL and then glued with epoxy as shown in Figure 3-18.
                                     3-22

-------
A stainless steel wire was attached to the fish hooks and threaded through 6-mm
outside-diameter plastic tubing, which protected the wire and minimized friction between
the wire and overlying soil. Each wire extended from the fish hook to a monitoring point
about 1.5 m  beyond the crest of the slope.  During construction, the monitoring point
consisted of wooden stakes driven into the soil  above the crest of the slope.  After
construction, the extensometer wires were  connected to a more permanent table
(Figure 3-19) at the crest of the slope, where the displacements were monitored for 3-
1/2 years.

Fish hooks were  attached to the upper and lower surfaces of each GCL panel at five
equally-spaced locations along the length of the slope (Figure 3-19), resulting in 20
extensometer monitoring points per test section. The accuracy of the extensometers
was estimated to be approximately 10 mm, based on experience with their use in the
field.  Displacements were typically measured every 1 to 4 weeks.

3.5 Laboratory Direct Shear Tests
The project schedule did not permit performing  laboratory shear tests prior to
construction. Instead, internal shear strength data from Shan and Daniel (1991), Daniel
et al. (1993), and Shan (1993) were used for design of the test plots. Additional
information on internal shear strength of GCLs is provided by Gilbert et al. (1996) and
Well (1997). As initial data on the performance of the test plots became available, it
became apparent that certain GM/GCL interfaces were more critical with  respect to
slope  stability than the internal shear strength of the GCLs.  Thus, the laboratory testing
program focused on interfaces.

Interface direct shear tests were conducted to evaluate textured GM/GCL interfaces and
the sand/GCL interface for one GCL.  The tests were performed using 300 mm by 300
mm specimens per ASTM D5321, with samples taken from the same lots of materials
deployed in the field. The GCLs were subjected to a normal stress  of 17  kPa
(equivalent to the field value in the test plots) and then hydrated for 10 days.  However,
Gundseal with bentonite encased between two  GMs was  not hydrated because the
bentonite in the field was not expected to become hydrated. The rate of shear was 1
mm/min per ASTM  D5321. All tests were single-point tests  (i.e., one normal  load of 17
kPa).  For simplicity of presentation, the test results were  interpreted in terms of a
secant friction angle (§s).  Peak and large-displacement (50 mm) secant friction  angles
are summarized in Table 3-2.  The term "large displacement" is used rather than
"residual" because the tests were carried out to displacements of about 25 mm and may
not have reached true residual conditions.
                                      3-23

-------
Expanded View of   ^
Upper Extensiometer
                                                 6-mm-OD
                                                 Plastic Tubing
                                                           Stainless
                                                           Steel Cable


                                                               Epoxy
Bentonite
Cable for Lower
 Extensiometer
               Fishhook
                                                  Upper Extensiometer
                                               L = Differential Displacement

                                                   Lower Extensiometer
                                               Time
Figure 3-18. Displacement sensors attached to GCL.
                                    3-24

-------
      Wooden Platform

      Scale
             Weight
Displacement
  Indicator
 Attached to
 Steel Cable
Displacement-Measurement
Platform at Top of Slope
           Crest of
           Test Plot
                                                   Layout of Displacement
                                                   Gauges Attached to
                                                   GCL Panels
             Fishhoks Attached
             to Upper and Lower
             Side of GCL at Five
             Locations Along
             Each GCL Panel
                                           Toe of Test Plot
Figure 3-19.  Displacement monitoring system.
                                       3-25

-------
Table 3-2.  Summary of Results of Interface Direct Shear Tests.


Test Plot

A,E,F,
&P
B&G

C&H

1

K

D&N



Type of
GCL
Gundseal

Bentomat
ST
Claymax
500SP
Bentofix
NW
Claymax
500SP
Bentofix
NS



GCL Interface
Dry Bentonite
(Internal Shear)
Woven Slit-Film GT

Woven Slit-Film GT

Nonwoven
Needlepunched GT
Woven Slit-Film GT

Nonwoven
Needlepunched GT


Opposing
Interface
Textured HOPE
GM
Textured HOPE
GM
Textured HOPE
GM
Textured HOPE
GM
Drainage Sand

Textured HOPE
GM

Peak Secant
Friction
Angle (°)
37

23

20

37

31

29

Large-
Displacement
Secant Friction
Angle (°)
35

21

20

24

31

22

Note:   Plots J and L (plots with drainage sand and no GM) were not specifically evaluated because a
       relatively high friction angle (31°) was measured for plot K, which like plots J and L also had
       drainage sand and no GM. It was assumed that the friction angle between the drainage sand
       and either Bentomat ST (plot J) or Bentofix NW (plot L) was no less than the 31° value
       measured for Claymax 500SP.
3.6 Performance of Test Plots

3.6.1  Construction Displacements
The displacements that occurred in the GCLs were divided into construction and post-
construction displacements.  Displacements were usually largest for the displacement
sensors located closest to the toes of the slopes (gauges 5-left and 5-right in Figure
3-19) and least for monitoring points located closest to the crests of the slopes (gauges
1-left and 1-right in Figure 3-19), indicating that the GCL panels were stretching.

Maximum downslope displacements measured during construction are summarized in
Figure 3-20. The measurements shown in the plot represent the average of the
maximum downslope movement of the left and right displacement  gauges above and
below the GCL (i.e., gauges 5-left above, 5-left below 5-right above, and 5-right below in
Figure 3-19). Maximum displacements were generally 10 to 40 mm for the 3H:1V
slopes and 40 to 200 mm for the 2H:1V slopes. Differential  displacements between the
upper and lower surfaces of the GCLs were  less than the resolution of the
extensometers (i.e., <  10  mm).
                                      3-26

-------
^.uu
240
'•§ -s E* 200
1 "I
£ § CD 160
o E Q-
0 CD o
o er\ A o/-\
E JP. i. I^U
||8 80-
CD H
^ 40
n
Error bars depict range of measured maximum displacements

[

D 2H:1V Slope
: T a 3H:1V Slope
[]
"I
"t1 "t i
T a


Q []
HKLF
                                       J   ICGDBE
                                         Test Plot
A
Figure 3-20. Maximum construction displacements measured at toe of slope.
Construction displacements were caused by mobilization of shear resistance at various
interfaces within the system and development of tension in the geosynthetic
components.  The test plots with the largest movements during construction were plot C
(3H:1V slope) and plots H and K (2H:1V slope). All three of these plots contained
Claymax 500SP which, with woven slit-film GTs on both surfaces, had the lowest
interface shear resistances of the GCLs and interfaces tested.

Test plot G (Bentomat ST) had the lowest displacement of the 21-1:1 V test plots,
probably because the soil at this test plot was less clayey than some of the others and
because the component of the GCL in contact with the subsoil was a nonwoven GT
(nonwoven GTs generally have better interface shear resistances with soils than do
woven slit-film GTs).  Plot A had the smallest movement of the 31-1:1 V plots, probably
because the textured  GM component of Gundseal interfaced with the subgrade soil.  In
general, a textured GM also has comparatively good interface shear resistance (more
than a woven slit-film  GT).
                                     3-27

-------
3.6.2  Post-Construction Performance of 3H: 1V Slopes
Post-construction displacements are summarized in Table 3-3.  All 3H:1V slopes have
remained stable. Total downslope displacements have been less than 50 mm, and
differential displacements have been less than 40 mm.  There has been no visual
evidence of movement or surface cracking.

3.6.2.1  Test Plot A (Bentonite Between Two GMs)
The bentonite component of Gundseal was expected to remain dry because the
bentonite was encased between two GMs.  As indicated in Table 3-2, measured peak
and large-displacement interface secant friction angles between dry bentonite and
textured HOPE were 37°  and 35°, respectively.  Since the slope angle was 16.9°, the
slope should be stable so long as the bentonite remains dry.

Fiberglass moisture sensors in plot A have provided variable results: two of the three
moisture sensors have indicated that the bentonite is dry, but one sensor near the crest
of the slope indicated some hydration (Appendix D). Two borings were drilled by hand
near the crest and toe of  the test plot in March 1995, and 100-mm-diameter samples of
the GCL were removed.  The water contents of the bentonite in the GCL at the crest
and toe were 27% and 24%, respectively.  These values are essentially identical to the
water content at the time of installation, confirming that the bentonite had not hydrated.
Individual fiberglass moisture sensors have been found during calibration to have
relatively large scatter (Appendix D); however, the general trend indicated by the
majority of sensors has proven to be correct in all test plots.

3.6.2.2  Test Plots B, C, and D (GT-Encased GCLs)
Test plots B, C, and D contain GT-encased GCLs.  The bentonite in the GCL was
expected to hydrate by absorbing moisture from subgrade soils. Most of the fiberglass
moisture sensors have indicated that the bentonite has hydrated, although less than
expected.  One factor inhibiting hydration may have been the relatively dry, sandy
subsoils on the 3H:1V test plots, compared to the 2H:1V test plots, which had more
clayey, wetter subsoils.

Experience has shown that GCL interface shear strengths are typically less than
internal shear strengths for internally-reinforced GCLs such as those used in test plots
B, C, and D, when tested at low normal stress (Gilbert et al., 1996).  Peak interface
secant friction angles between the upward-facing GT component of the GCLs and the
textured HOPE GM are 20° to 29°, and large-displacement friction angles are 20° to 22°
degrees for essentially full hydration of the GCLs (Table 3-2).
                                      3-28

-------
Table 3-3. Summary of Post-Construction Performance of Test Plots.


Plot
A
B
C
D
E
F


G



H



1


J


K


L


N
P


Slope
3H:1V
3H:1V
3H:1V
3H:1V
3H:1V
2H:1V


2H:1V



2H:1V



2H:1V


2H:1V


2H:1V


2H:1V


2H:1V
2H:1V


Type of GCL
Gundseal
Bentomat ST
Claymax 500SP
Bentofix NS
Gundseal
Gundseal


Bentomat ST



Claymax 500 SP



Bentofix NW


Bentomat ST


Claymax 500SP


Bentofix NW


Bentofix NS
Gundseal

Stability of Test Plot As of
June, 1997
Stable
Stable
Stable
Stable
Stable
Internal Slide within the GCL
Occurred 495 Days after
Construction of Test Plot
Interface Slide between Lower
Side of GM and Upper Woven GT
of GCL 20 days after Construction
of Test Plot
Interface Slide between Lower
Side of GM and Upper Woven GT
of GCL 50 days after Construction
of Test Plot
Slumps and Surface Cracks
Developed about 900 Days after
Construction of Test Plot
Slumps and Surface Cracks
Developed about 900 Days after
Construction of Test Plot
Slumps and Surface Cracks
Developed about 900 Days after
Construction of Test Plot
Slumps and Surface Cracks
Developed about 900 Days after
Construction of Test Plot
Stable
Stable
Total
Displacement
(mm)
20
30
25
50
30
-


-



-



500


800


1200


500


30
NA
Differential
Displacement
(mm)
10
40
30
25
30
750


25



130



25


75


900


180


10
NA
Note: Total displacement is the total amount of downslope movement measured after construction was
     complete; differential displacement is the difference between downslope movement of the upper
     and lower surfaces of the GCL that occurred after construction.
                                         3-29

-------
3.6.2.3  Test Plot E (Unreinforced GCL)
Test plot E was constructed with the bentonite portion of Gundseal facing downward.
Interface shear tests were not performed on the hydrated GCL because the internal
shear strength under consolidated-drained conditions had already been studied.
Consolidated-drained conditions were used because the GCLs in the test plots were
installed dry and hydrated slowly under conditions most appropriately simulated
consolidated-drained conditions. Although consolidated-drained conditions were
thought to be most appropriate for these test plots, it is conceivable that unconsolidated-
undrained or consolidated-undrained conditions may be more critical for other
conditions, e.g., seismic loading. The designer should consider the most appropriate
and critical condition for any particular application. The consolidated-drained direct
shear tests performed previously employed fully hydrated samples that had water
contents of approximately 150%.   Resulting interface shear strengths were found to
vary with normal stress  (Figure 3-21).  For the normal stress acting on the GCL in plot E
(17 kPa), the drained angle of internal friction for fully hydrated bentonite is about 20°.
The slope angle at plot  E was 17.7°; thus, the test plot is expected to be stable if the
bentonite is hydrated, but only with FS = tan  (20°)/tan (17.7°) = 1.14 for an infinite slope.

As with most of the other test plots,  the fiberglass moisture sensors for test plot E have
yielded variable results, with some sensors indicating that the bentonite has become
hydrated and others indicating that it has not become hydrated.  Experience showed
that when the sensors indicated that the GCLs were dry, the GCLs were indeed  dry,
and that when  the sensors indicated that the GCLs had become  very wet, they were
indeed nearly fully hydrated.  However, for the broad range of moisture sensor readings
between dry and fully hydrated, the  moisture sensors were not found to be particularly
useful for indicating the  degree of partial hydration.

To verify actual moisture conditions, a boring was drilled and a sample was taken near
the crest of the slope (the driest area) in March 1995, and the water content of the
bentonite was found to be 46%.  Eight more borings were drilled in April 1996, at
various locations along  the full length of the slope. The water content varied between
54% and 79%, and averaged 60%.  Daniel et al. (1993) previously measured the shear
strength of the bentonite component of Gundseal as a function of water content using a
direct shear apparatus and a slow rate of shear that allowed excess pore water
pressures to fully dissipate.  The range of normal stress used in the testing program
was 27 to 139  kPa. The highest water content (^145%) was achieved by fully hydrating
the bentonite.  Results,  plotted in Figure 3-22, show that once the water content  of the
bentonite reaches 50%  or more, the shear strength declines to a value approximately
equal to the strength of  fully hydrated bentonite. In other words,  the bentonite does not
have to be fully hydrated for its strength to be greatly reduced. This phenomenon is
observed from handling a GCL; at 50% water content, the bentonite feels hydrated and
                                      3-30

-------
very slick.  Thus, the average water content of 60% in test plot E should be sufficiently
large to replicate the strength reduction associated with full hydration of the bentonite.
              CD
              CD
              D)
              CD
              c
              CO
              c
              D
              D)
Tilt Table
Direct Shear
                                                   100
                                 Normal  Stress (kPa)
              1000
Figure 3-21. Influence of normal stress on secant angle of internal friction for
            internal shear of an unreinforced GCL (after Shan, 1993).
3.6.3  Post-Construction Performance of 2H:1V Plots
Slides have occurred at many of the 21-1:1 V test plots for different reasons. Two slides
occurred at plots G and H a few weeks after construction was complete. These two
slides are shown in Figure 3-23. Both involved slippage at the interface between the
upper surface of the GCL (a woven GT in both cases) and the lower surface of the
textured HOPE GM. The next slide occurred in plot F about a year and a half after
construction.  In this case, the bentonite (which was encased between two GMs) in this
GCL unexpectedly became hydrated, and a slide resulted. The other test plots
remained stable for the next two years, but then several slides occurred in the subsoils
beneath other test plots. The subsoils were plastic clays, and the subsoil slides (which
occurred at the end of a wet spring season) were presumed to be the result of hydration
of the subsoil clays and possibly the buildup of excess pore water pressure in the
subsoils, as well.
                                     3-31

-------
      (U
      2
      O)
      (U
           25
     20
O
           10
                               50                100

                          Water Content of Bentonite (%
                                                            150
Figure 3-22.  Effect of water content on the secant friction angle of an
            unreinforced GCL sheared internally (Daniel et al., 1993).
3.6.3.1  Test Plots G and H
Test plots G and H consisted of Bentomat ST and Claymax 500SP, respectively. Both
plots slid at the interface between the upper GT (a woven, slit-film GT in both cases)
and the lower surface of the overlying textured  HOPE GM.  Plot H slid 20 days after
construction, and plot G slid 50 days after construction.  Pre-slide displacements were
small (< 25 to 130 mm).  There was no warning of either slide.  Both slides occurred at
night, and the slides apparently occurred quickly.

Test Plot H, which incorporated Claymax  500SP,  was constructed on a 24.7° slope, but
the measured peak and large-displacement interface friction angles for the relevant
materials under hydrated conditions were only 20° (Table 3-2). Test plot H did not  slide
immediately because the interfacial shear strength of the dry GCL was sufficient to
maintain a stable slope.  The slope slid when the  bentonite hydrated. Tests
summarized in Appendix D showed that bentonite in the GCLs hydrated in a period of
10 to 20 days when placed in contact with the subgrade soils from the test plots. Tests
reported by Daniel et al. (1993) showed similar results for other soils. Thus,  the sliding
time of 20 days after construction is consistent with  the expected period to achieve
nearly full hydration of the bentonite.
                                      3-32

-------
Figure 3-23. Photograph of slides in plot G (left) and H (right) taken
            approximately 2 months after construction and several days after the
            slide in plot G.
When GCLs containing a woven GT component become hydrated, bentonite can
extrude through the openings of the GT and lubricate the GCL/GM interface (Gilbert et
al., 1996).  After the slide, the surface of the GCL was very slick. The tendency of
bentonite to lubricate the GM/GCL interface may be related to the thinness of the woven
slit-film GT and to differences in apparent opening size between woven and nonwoven
GTs.

Test plot G, which was constructed using  Bentomat ST, was slower to slide, but the
slope angle (23.5° peak) was 1.2° flatter than for plot H, and the interface shear strength
between the GCL and overlying GM (23° peak and 21° large-displacement) was 1° to 3°
higher. Also, a nonwoven GT faced downward in plot G, but a woven slit-film GT faced
downward  in Plot H.  GCLs are expected to absorb water more slowly from subgrade
soils when the GT separating the bentonite from the subsoil is a thicker nonwoven GT.
Thus, the reason why plot G slid 30 days later than plot H appears to be that the
                                     3-33

-------
bentonite in the GCL at plot G was separated from wet subgrade soils by a thicker,
nonwoven GT, which slowed  hydration.

3.6.3.2  Test Plots F and P (Bentonite Encased Between Two GMs)
Plots F and P (both 2H:1V slopes), like plot A (3H:1V slope), contained bentonite
sandwiched  between two GMs. The bentonite in these test plots was expected to
remain dry.  However, within three months after plot F was constructed, two of the three
moisture sensors indicated that the bentonite had become hydrated.

To evaluate the condition of the bentonite, 17 borings were drilled into Plot F in March
1995, and 100-mm-diameter samples of the GCL were recovered.  The water content of
the bentonite samples varied  from 10% to 188%, and the data showed that the right
panel was much more hydrated than the left panel.  In contrast to this field data,
Estornell and Daniel (1992) reported laboratory test results for Gundseal in which water
migrated laterally through the GM-encased bentonite less than 100 mm over a test
duration of 6 months.

Water may have entered the bentonite at plot F through cuts made in the GM liner
overlying the GCL to allow insertion of the extensometer cables. Plot F was  located at a
point where surface water at the crest of the slope was channeled directly to  the anchor
trench area where the penetrations were made. The mechanism for lateral movement
of water is probably waves in  the overlying GM, which would allow water to spread.
Alternatively, the source of water could have come from the V-shaped trough between
plots F and G, and spread through waves in the GM. Unfortunately, the plot  slid before
a complete forensic study could be performed.

Displacement sensors showed large movements in the right panel of plot F throughout
the first year of observation, but significant movement was not initiated in the left panel
until later (Figure 3-24). Starting on about day 275 (August 1995), the left panel began
to move downslope, suggesting that the bentonite in the  left panel was finally becoming
hydrated over a significant percentage of the total area of the panel.

Plot F slid  on March 24, 1996, 495 days after construction.  The  cause of the slide is
hydration of the bentonite; the peak angle of internal friction for hydrated bentonite at
the normal stress existing in the field was 20°, but the slope angle was 23.6°. In
contrast, the peak interface friction angle for dry bentonite was 37°.  Had the  bentonite
not hydrated, the slope should have remained stable.

In response to the unexpected hydration, plot P was constructed on June 15, 1995.
The extensometers were not  installed in plot P to eliminate all  penetrations in the
overlying GM.  The number of fiberglass moisture sensors in the bentonite was
                                      3-34

-------
                 Differential Displacement Vs. Time - Left Panel
                      100
                                   Time (days)

                                200         J300
     100 -•

"CD

'c E
E E  200 - •

igr


° E  300 i
CC 0
C o
   CD JP.
       400 -
   E CD 500 -
   CD CD
   O..C
   Q-CO

   ^   600 i
       700 -1
                         Geosynthetics Cut

                       Next to Anchor Trench
                Key to Extensometers:
                    No. 1 (Crest of Slope)

                    No. 2


                    No. 3

                    No. 4


                    No. 5 (Toe of Slope)
                       Differential Displacement Vs. Time - Right Panel
                        100
                                     Time (days)

                                 200         300
400

—l—
    ||  200



    £¥
    5 I  300

    TO CD


    ^-§^  400



    •^ Q
    g j-  500
    i- CD
    CD CD


    I"  600
         700 -1
500
              Key to Extensometers:



                —•—  No. 1 (Crest of Slope)

                -Q-  No. 2


                -•-  No. 3

                ^0^  No. 4


                —A—  No. 5 (Toe of Slope)
Figure 3-24. Differential displacement versus time for left and right panels of plot

             F.
                                        3-35

-------
increased from 3 in the other test plots to 16 in plot P to provide additional
documentation of moisture conditions. All but one of the 16 moisture sensors have
indicated that the bentonite has remained dry in the 18 months of monitoring plot P.
There is, however, some indication of slight hydration near the toe of plot P, perhaps
due to edge effects near the toe.

There has been no indication of any displacement in plot P, although displacements are
not being monitored (other than to observe for gross and obvious movements, of which
there have been none).  It is apparent from the moisture sensors and stability of plot P
over a period of more than four years that the bentonite has not become hydrated over
a significant area of plot P. This plot confirms that if moisture is not allowed access to
bentonite sandwiched between GCLs through penetrations or other sources, the
bentonite can remain dry for at least several years.

All 31-1:1 V slopes with unreinforced bentonite (including plot A with bentonite encased
between two GMs and plot E with bentonite in contact with moist subgrade) have
remained stable with minimal displacement.

3.6.3.3  Plots I and N with Nonwoven GT Component Facing Upward
Plots I and N are similar to plot G, except that the GCL contained either one nonwoven
GT with the nonwoven GT facing upward (plot N) or two nonwoven GTs (plot I).  The
slope angles at plots I and N were similar to the other 21-1:1 V plots.  However, the
interface friction angle between the nonwoven GT component of Bentofix and the
textured HOPE (37° peak and 24° large-displacement) was much greater than for the
woven slit-film GT component of the GCLs that slid. The geosynthetic components of
plots I and N have remained stable because of the better interface shear resistance
between a nonwoven GT component of a GCL compared to a woven GT component.
The greater interface shear resistance from the nonwoven GT is attributed  to: (1) larger
shear resistance developed between nonwoven GTs and textured GMs in general; and
(2) less hydrated bentonite extrusion to the interface for the thicker nonwoven GT.

Large displacements began to develop in plot I and the  adjacent test plots J, K, and  L
about 3-1/2 years (900 days) after construction.  Several small slumps with downward
displacement of up to about  100 mm along scarps, and  associated surface cracking,
were observed. The slumps and cracks appearing in the lower half of the test plot.
The subsoils in the area of the slides are CL and CH clays, with the liquid limit and
plasticity index of the subsoil next to plot I averaging 48% and 24%,  respectively.  The
average moisture content of the clay at the time of sliding was approximately 40%, or
just slightly below the liquid limit. The displacements occurred at the end of the wet
spring season in 1997.  Examination of plot I and adjacent test plots,  coupled with
excavation  into the subsoils, showed that sliding  was occurring 0.5 to 1 m beneath the
                                      3-36

-------
GCL, in the clay subsoil. It is assumed that the buildup of pore water pressure behind
the test plots helped to trigger the slides in the subsoils. There was no indication of
movement within the GCL or either interface with the GCL. Plot N showed no signs of
slumping or cracking, but plot N was at the end of the 21-1:1 V test plots and likely was at
a location where excess pore water pressures were not as likely to develop.

3.6.3.4  Plots J, K, and L with No GM
These test plots were constructed by placing drainage sand directly above the GCL.  All
three test plots remained stable for about 900 days after construction, and then all three
underwent significant downslope displacement (0.5 to 1.2 m, as shown in Table 3-3).
All three exhibited slumping in the lower half to two-thirds of the test plots.  Scarps could
be observed at several  locations within each test plot.  Observation of the depth of
slumping clearly showed that displacement was occurring entirely beneath the GCLs.
Excavation into the subsoils showed that a layer of more plastic clay was located just
beneath the bottom of the GCLs.  The sliding mechanism was related to  the subsoils
and not to the GCLs or  GCL interfaces.  Buildup of pore water pressure in the clays
following the wet spring season was assumed to be the triggering mechanism.  The
large differential displacements indicated for plots K and L in Table 3-3 are not
representative of actual shearing of the GCLs (physical examination of the GCLs
showed that they were intact and not sheared) - the large displacements apparently
rendered the differential displacement between top and bottom sensors meaningless.

The peak secant interface friction angle between the sand drainage material and GCL
was 31° for a woven-slit film component (Table 3-2) and, although not measured,
presumably more for a nonwoven component.  An interface friction angle of 31° is
significantly greater than the slope angle (~ 25°), which explains the stability of the test
plots up until the point of sliding in the subsoil.

3.6.4 Comments on Adequacy of Current Engineering Practice
Current engineering practice for evaluation of the stability of slopes such as those
constructed for the test  plots involves three steps: (1) measurement of the internal and
interfacial shear strength, typically using direct shear apparatus to shear 300 mm by
300 mm test specimens and interfaces; (2) calculation of the factor of safety using limit
equilibrium analysis; and (3) reconfiguration of the slope or use of different materials if
the factor of safety is not found to be adequate.  In this project, the shearing tests were
performed after the slopes were constructed (due to time constraints). A critical
question is: would the usual testing/design process have correctly predicted which
slopes would be stable  and which would undergo sliding?

The calculated FS based on both peak and large-displacement shear strength
measurements is tabulated in Table 3-4. The column labeled "Peak FS" indicates factor
                                      3-37

-------
of safety based on peak shear strength. The column labeled "Large Displ. FS" refers to
the factor of safety calculated from the shear strength measured at large displacement
(50 mm).

Table 3-4.  Summary of Calculated Factor of Safety (FS) and Actual Slope
           Stability.
Test Plot
A
B
C
D
E
F
G
H
1
J
K
L
N
Slope
Angle (°)
16.9
17.8
17.6
17.5
17.7
23.6
23.5
24.7
24.8
24.8
25.5
24.9
22.9
Peak
Friction
Angle (°)
372(D)
231
201
291
202(H)
202(H)
231
201
371
-31 1
313
-31 1
~371
Large-
Displ.
Friction
Angle (°)
352(D)
211
201
221
202(H)
202(H)
211
201
241
-31 1
313
-31 1
-241
Peak FS
2.52(D)
1.31
1.11
1.81
1 ^(H)
0.82(H)
1.01
0.81
1.61
1.31
1.31
1.31
1.81
Large
Displ. FS
232(D)
1.21
1.11
1.31
1 ^(H)
Q.82(H)
0.91
0.81
1.01
1.31
1.31
1.31
1.11
Test Plot
Performance
Stable
Stable
Stable
Stable
Stable
Internal
Slide
Interface
Slide
Interface
Slide
Stable4
Stable4
Stable4
Stable4
Stable
 GCL/GM interface
' Internal GCL strength for dry (D) or hydrated (H) bentonite
!GCL/drainage sand interface
1 Large displacement occurred in subsoil below GCL, but not in or at the interface with GCL
From these data, the following conclusions are drawn: (1) all test plots with a factor of
safety > 1.0 based on peak shear strengths remained stable with respect to the critical
material or interface tested; (2) all test plots with a factor of safety > 1.0 with respect to
the large-displacement shear strength remained stable with respect to the critical
                                        3-38

-------
material or interface tested; (3) all test plots with a factor of safety < 1.0 based on peak
shear strength underwent a slope failure; and (4) all test plots with a factor of safety <
1.0 based on residual shear strength underwent a failure.  Although there are no firm
rules on what designers should and should not assume for the relationship between
factor of safety and stability of a slope, the following comments are offered:

   1.   All designers would likely assume that a slope with F < 1  based on peak
       strengths would be at great risk of failure and would not be acceptable (and,
       indeed,  all such test plots did undergo failure).
   2.   Many designers assume that slopes with a factor of safety of about 1.5 or better,
       based on peak shear strengths, will remain stable  under static loading (and all
       plots with F > 1.5 were stable).
   3.   Many designers assume that slopes with a factor of safety of about 1.2 to 1.5
       calculated from peak shear strengths may remain stable, particularly in the short
       term, under static loading conditions but might not  feel comfortable with factors
       of safety between 1.2 and 1.5 for long-term stability of critical slopes (all test
       plots with F between 1.2 and 1.5 remained stable in the GCL or at a critical
       interface for the 4-1/2 years in which the test plots  have been observed).
   4.   Many designers consider the residual shear strength of a material or interface,
       and  a common assumption is that if the factor of safety based on peak strength
       is acceptable and the factor of safety based on large-displacement shear
       strength is at least 1.0, then the design is acceptable (all test plots with F > 1.0
       based on large-displacement shear strength were  indeed stable).

It appears that the observations from this test plot program are entirely consistent with
current design practice. Had current design practices been employed for  the materials
and slopes used in these test plots, stable slopes would have resulted for the 4-1/2 year
period of this test program. The observations from the test plots are consistent with and
serve to validate current design  methodology.

3.7 Erosion Control Materials
The GEC materials that were employed for the test plots are summarized  in Table 3-5.
The erosion  control materials were installed in an overlapping manner and stapled
together. Ground anchors per the manufacturer's installation recommendations were
used throughout. Some plots were seeded prior to placement of the erosion control
material,  and others were seeded after the placement of the erosion control  material
(depending on the manufacturer's recommendation). The  owner of the landfill site
provided the seeding in December 1994.

The materials were all fully effective in preventing the development of erosion over the
4-1/2 years in which the condition of the test plots was observed. In contrast, erosion
gullies and rills formed  in the control plot M, which did not contain any erosion control
material.
                                       3-39

-------
Table 3-5. Geosynthetic Erosion Control Products.
Plot
A
B
C
D
E
F
G
H
1
J
K
L
M
N
P
Manufacturer
Tensar
Synthetic Industries
Synthetic Industries
Akzo
Akzo
Tensar
Tensar
Tensar
Synthetic Industries
Synthetic Industries
Akzo
Akzo
None
Akzo
Akzo
Product
TB 1000
Polyjute
Polyjute
Enkamat7010
Enkamat7010
TM 3000
TM 3000
TM 3000
Landlok450
Landlok450
Enkamat7010
Enkamat7010
Control Plot
Enkamat7010
Enkamat 7220
Color
Green
Beige
Beige
Black
Black
Black
Black
Black
Green
Green
Black
Black
-
Black
Black
Material
Polyolefin
Degradable Polypropylene
Degradable Polypropylene
Nylon
Nylon (with Excelsior)
Polyethylene
Polyethylene
Polyethylene
Polyolefin
Polyolefin
Nylon (with Excelsior)
Nylon
-
Nylon (with Excelsior)
Nylon
3.8 Summary and Conclusions
Fourteen test plots, designed to replicate typical final cover systems for solid waste
landfills, were constructed to evaluate the internal and interface shear strength of GCLs
under full-scale field conditions on 2H:1V and 3H:1V slopes. Five different types of
GCLs were evaluated. The test plots have been observed for 4-1/2 years.  All test plots
were initially stable, but over time as the bentonite in the GCLs became hydrated, three
slides (all on 2H:1V slopes) that involved the GCLs have occurred. One slide involved
an unreinforced GCL  in which bentonite that was encased between two GMs
unexpectedly became hydrated. The other two slides occurred at the interface between
the woven GTs of the GCLs and the overlying textured HOPE GM. Several slides (none
involving GCLs) occurred in the subsoils of 2H: 1V test plots following a wet spring about
3-1/2 years into the project.

Conclusions from the project may be summarized as follows: (1) at the low normal
stresses associated with landfill cover systems, the interface shear strength is generally
                                      3-40

-------
lower than the internal shear strength of internally-reinforced GCLs; (2) interfaces
between a woven GT component of the GCL and the adjacent material (e.g., textured or
smooth HOPE GM) should always be evaluated for stability; these interfaces may often
be critical; (3) significantly higher interface shear strengths were observed when the GT
component of a GCL in contact with a textured HOPE GM was a nonwoven GT, rather
than a woven GT; (4)  if bentonite sandwiched between two GMs has access to water
(e.g., via penetrations  or at exposed edges), water may spread laterally through waves
or wrinkles in the GM and hydrate the bentonite over a large area; (5) if the bentonite
sandwiched between two GMs does not have access to water, it was found that the
bentonite did not hydrate over a large area; (6) current engineering procedures for
evaluating the stability of GCLs on slopes (based on laboratory direct shear tests and
limit-equilibrium methods of slope stability analysis) correctly predicted  which test plots
would remain stable and which would undergo sliding, thus validating current design
practices; and  (7) based on the experiences of this study, 21-1:1 V slopes involving landfill
cover situations may be too steep to be stable with the desirable factor of safety due to
limitations with respect to the interface shear strengths of the currently  available
geosynthetic products.

The results from the test plot were consistent with the current design practice involving
measuring the shear strength of critical materials and interfaces in laboratory direct
shear tests, calculating the factor of safety against a slope failure using limit equilibrium
analysis, and adjusting the slope cross section or materials until a satisfactory factor of
safety is achieved. The test plots served to validate the current design methodology.
All test plots with a factor of safety > 1 based on peak shear strength were found to be
stable in the critical material or along the  critical interface during the 4-1/2 years that the
test plots have been observed. All test plots with a factor of safety > 1.0 based on
large-displacement shear strengths were stable.  Slides that did occur along GCL/GM
interfaces occurred at  plots where the factor of safety using peak shear strength was
< 1.0 (and in such cases, the factor of safety based on large-displacement shear
strengths was  < 1.0).

3.9  References
Boardman, B.T. and D.E. Daniel (1996), "Hydraulic Conductivity of Desiccated
  Geosynthetic Clay Liners," Journal of Geotechnical Engineering, 122(3): 204-208.
Daniel, D.E., and Boardman, B.T. (1993), Report of Workshop on Geosynthetic Clay
  Liners, U.S. Environmental Protection  Agency, Cincinnati, Ohio, EPA/600/R-93/171,
  106  p.
Daniel, D.E., Shan, H.Y., and Anderson,  J.D. (1993), "Effects of Partial Wetting on the
  Performance of the  Bentonite Component of a Geosynthetic Clay Liner,"
  Geosynthetics '93, Industrial Fabrics Association International, St. Paul, Minnesota,
  3: 1483-1496.
                                      3-41

-------
Daniel, D.E., and Gilbert, R.B. (1996), "Practical Methods for Managing Uncertainties for
  Geosynthetic Clay Liners," Uncertainty in the Geologic Environment: From Theory to
  Practice, American Society of Civil Engineers, New York.
Dobras, T.N., and Elzea, J.M. (1993), "In-Situ Soda Ash Treatment for Contaminated
  Geosynthetic Clay Liners," Geosynthetics '93, Industrial Fabrics Association
  International, 3: 1145-1160.
Estornell, P. M. and Daniel, D.E. (1992), "Hydraulic Conductivity of Three Geosynthetic
  Clay Liners," Journal of Geotechnical Engineering, 118(10): 1592-1606.
Foose, G.J., Benson, C.H., and Edil, T.B. (1996), "Evaluating  the Effectiveness of
  Landfill Liners," Environmental Geotechnics, M.  Kamon (Ed.), Balkema, Rotterdam,
  1:217-221.
Foose, G.J., Benson, C.H., and Edil, T.B. (1999), "Equivalency of Composite
  Geosynthetic Clay Liners as a Barrier to Volatile Organic Compounds,"
  Geosynthetics '99, 1: 335-344.
Fox, P.J., De Battista, D.J., and Chen, S.J. (1996), "Bearing Capacity of Geosynthetic
  Clay Liners for Cover Soils of Varying Particle Size," Geosynthetics International,
  3(4): 447-461.
Fox, P.J., Rowland, M.G., and Scheithe, J.R. (1998), "Internal Shear Strength of Three
  Geosynthetic Clay Liners," Journal of Geotechnical and Geoenvironmental
  Engineering, 124(10): 933-944.
Gilbert, R.B.,  Fernandez, F., and Horsfield, D.W. (1996), "Shear Strength of Reinforced
  Geosynthetic Clay Liner," Journal of Geotechnical Engineering, 122(4): 259-266.
Hewitt, R.D., and D.E. Daniel (1997), "Hydraulic Conductivity  of Geosynthetic Clay
  Liners after Freeze-Thaw," Journal of Geotechnical and Geoenvironmental
  Engineering, 123(4):  305-313.
James, A.M.,  Fullerton,  D., and Drake, R. (1997), "Field Performance of GCL under Ion
  Exchange Conditions," Journal of Geotechnical  and Geoenvironmental Engineering,
  123(10):  897-902.
Koerner, R.M., and Narejo, D. (1995),  "Bearing Capacity of Hydrated Geosynthetic Clay
  Liners," Journal of Geotechnical Engineering, 121(1): 82-85.
Koerner, R.J., Koerner,  G.R., and Eberle, M.A.  (1996), "Out-of-Plane Tensile Behavior
  of Geosynthetic Clay Liners," Geosynthetics International, 3(2): 277-296.
Kraus, J.F., Benson, C.H., Erickson, A.E., and Chamberlain, E.J. (1997),  "Freeze-Thaw
  Cycling and Hydraulic Conductivity of Bentonitic Barriers, " Journal of Geotechnical
  and Geoenvironmental Engineering, 123(3):  229-238.
Lagatta, M.D. Boardman, B.T., Cooley, B.H., and D.E. Daniel  (1997), "Geosynthetic
  Clay Liners Subjected to Differential Settlement," Journal of Geotechnical and
  Geoenvironmental Engineering,  123(5): 402-411.
Lai, J., Daniel, D.E., and Wright, S.G. (1998), "Effect of Cyclic Loading on Shear
  Strength  of Unreinforced Geosynthetic Clay Liner," Journal of Geotechnical and
  Geoenvironmental Engineering,  124(1): 45-52.
                                      3-42

-------
Mesri, G., and Olson, R.E. (1970), "Shear Strength of Montmorillonite," Geotechnique,
  20(3): 261-270.
Olson,  R.E. (1974), "Shearing Strengths of Kaolinite, Illite, and Montmorillonite," Journal
  of the Geotechnical Engineering Division, ASCE, 100(GT11): 1215-1229.
Ruhl, J.L., and D.E. Daniel (1997), "Geosynthetic Clay Liners Permeated with Chemical
  Solutions and Leachates," Journal of Geotechnical and Geoenvironmental
  Engineering,  123(4): 369-381.
Schubert, W.R.  (1987), "Bentonite Matting in Composite Lining Systems," Geotechnical
  Practice for Waste Disposal '87, American Society of Civil Engineers, New York, 784-
  796.
Shan, H.Y. (1993), "Stability of Final Covers Placed on Slopes Containing Geosynthetic
  Clay Liners," Ph.D. Dissertation, Univ.  of Texas, Austin, TX, 296 p.
Shan, H.Y., and D.E.  Daniel (1991), "Results of Laboratory Tests on a
  Geotextile/Bentonite Liner Material," Geosynthetics 91, Industrial Fabrics Association
  International, St. Paul, MN, 2: 517-535.
Stark, T.D., and Eid, H.T.  (1997), "Shear  Behavior of Geosynthetic Clay Liners,"
  Geosynthetics International, 3(6): 771-786.
Stark, T.D., Arellano,  D., Evans, W.D., Wilson,  V.L., and Gonda, J.M. (1998),
  "Unreinforced Geosynthetic Clay Liner Case History," Geosynthetics International,
  5(5):521-544.
Well, L.W. (1997), Testing and Acceptance Criteria for Geosynthetic Clay Liners, ASTM
  STP 1308, American Society for Testing and Materials, Philadelphia, 268 p.
                                      3-43

-------
                                  Chapter 4
                   Summary of Natural Materials Tasks

The natural materials tasks were designed to collect information on field performance of
natural materials in landfills and focused on documenting the performance of CCLs.
Although drainage materials were included, it was recognized that this could be better
evaluated from case histories, discussed later in this report. The work on CCLs
included: (1) documenting the field hydraulic conductivity (kfieid) of CCLs, including CCLs
constructed from natural soil materials and those constructed from soil-bentonite
mixtures; and (2) documenting the performance of CCLs used in landfill covers.

The rationale for focusing attention on these two topics is as follows.  CCLs are an
important component of many liner systems for waste containment facilities.  The key
performance parameter (embodied in regulatory compliance) is the hydraulic
conductivity of the CCL. Although hydraulic conductivity is routinely measured on small-
sized laboratory samples, it is not commonly measured at a large scale on field test
pads. To provide an adequate database, this task focused on the large-scale hydraulic
conductivity of CCLs in the field to provide information on how CCLs are meeting the
regulatory requirement and design objective and to relate the performance of CCLs with
critical design and construction variables.

Although CCLs are commonly used  in landfill cover systems,  data published in recent
years have cast doubt on how survivable CCLs are in landfill cover systems.
Desiccation, for example, can lead to cracking in CCLs and to permanent and
significant increases in hydraulic conductivity.  In addition, MSW landfills are known to
undergo significant amounts of settlement upon closure.  Such settlement can adversely
affect a CCL's performance by inducing stress-related fractures in the CCL barrier that,
under low compressive stress,  can increase the CCL's hydraulic conductivity. The
available information is summarized  in order to provide information on the field
performance of CCLs in landfill covers.

4.1  CCLs Constructed from  Natural Soil Liner Material

4.1.1  Introduction
One of the most important components of many liner and cover systems for landfills is a
low-permeability, CCL.  In the past decade, dozens of papers have been written on the
factors that  influence the hydraulic conductivity (k) of compacted clay, the compatibility
of clay liners with chemicals and leachates, laboratory and field hydraulic conductivity
testing methods, the correlation of laboratory and field hydraulic conductivity, and
methods of construction and CQA.
                                      4-1

-------
Despite the wealth of information that has been recently developed, the performance of
CCLs in the field is largely undocumented. There are two principal reasons for the lack
of performance data.  First, compacted clays are usually used in combination with GM
liners, and in such applications it is impossible to separate the performance of the
compacted clay component from that of the GM component. Second, few landfills have
lysimeters or other instrumentation that would enable documentation of field
performance installed directly beneath the lining system.  Although lysimeters have
been used to document the field performance of several CCLs in actual landfills
(Gordon et al., 1990; Reades et al., 1990), only a handful of CCLs are documented in
this manner. Future opportunities to document field performance of CCLs with
lysimeters will be very limited because CCLs are rarely used in lining systems without a
GM.

The purpose of this task was to collect as much information as possible on the field
performance of CCLs. In view of the dearth of information on actual performance of in-
service CCLs, the next best source of data was used: large-scale field hydraulic
conductivity tests on full-scale field test pads. The test pads were constructed with
materials, methods of construction, and quality assurance (QA) procedures that vary
from project to project but are typical of current industry practices.  Any test pads that
were constructed for research purposes were excluded from the database. Only those
test pads that were constructed for the purpose of verifying that the field-constructed
CCL had a hydraulic conductivity of 1 x 10"7 cm/s or less were included in the database
presented and analyzed here.

4.1.2  Database

4.1.2.1  Source of Data
The database described herein was assembled partially from data in the literature but
primarily from unpublished  data contained in various engineering reports. Information on
more than 120 sites was collected and screened. The data collection process is thought
to have captured the results of perhaps 50% to 75%  of all the CCL test pads that have
been constructed in North America for the purpose of demonstrating compliance with
the hydraulic conductivity requirement of k < 1 x 10~7 cm/s as measured by large-scale
field testing equipment.

Although data on more than 120 CCLs were obtained, some of the sites were not
included in the final database. The requirements for inclusion of a full-scale CCL or a
CCL test pad in the database were: (1) construction in general accord with industry
practices for full-sized liners; (2) CQA in general accord with industry practices; (3)
construction with the objective of demonstrating that  large-scale kfieid did not exceed 1 x
10~7 m/s; (4) reasonably complete documentation of test results; and (5) availability of
results from large-scale kfieid tests such as the sealed double-ring infiltrometer, or SDRI
                                      4-2

-------
(Daniel, 1989; Trautwein and Boutwell, 1994). Of the CCLs eliminated from the
database, the main reason for doing so was the construction of a test pad to meet a
hydraulic conductivity objective other than 1 x 10~7 cm/s. Approximately 20 test pads
constructed in California with a hydraulic conductivity objective of 1 x 10~6 cm/s or less
were eliminated for this reason. Several test pads were constructed for research
purposes using construction practices that are not consistent with  industry practice for
full-sized liners,  and were not included in the database for this reason.

4.12.2  The Database
The database consists of 89 CCLs.  Of the 89 CCLs in the database, 8 are actual in-
service liners for landfills and 81 are test pads. The geographic distribution of CCLs in
the database is shown in Figure 4-1.  Data for the 89 CCLs are compiled in four tables
presented in Appendix C. Specific site locations are not provided due to potential
sensitivities for some sites.

The database is summarized in Tables 4-1 and 4-2.  Many of the numbers in Tables 4-1
and 4-2 are averages of multiple measurements (geometric mean for k, arithmetic mean
for all others). The statistics on the mean and standard deviation are summarized in
Tables C-1 through C-4 in Appendix C.
Figure 4-1.  Locations of sites in database for clay liners constructed of natural
            clay material.
                                      4-3

-------
Table 4-1. Field Hydraulic Conductivity, Soil Characteristics, and Compaction Data for Natural Clay Liner Materials.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Field
Hydraulic
Conductivity
(cm/s)
Value
2.8E-07
1.5E-07
9.0E-09
1.1E-08
9.0E-08
2.7E-07
5.8E-08
1.2E-07
7.0E-09
3.0E-08
3.0E-09
2.0E-09
1.3E-08
2.0E-08
3.3E-09
3.0E-08
6.0E-09
9.8E-09
4.4E-08
8.0E-07
2.5E-07
Method
SDRI
SDRI
LYS
SDRI
SDRI
SDRI
SDRI
SDRI
LYS
LYS
LYS
LYS
SDRI
SDRI
SDRI
SDRI
LYS
SDRI
LYS
SDRI
SDRI
Liquid
Limit
(%)
24
58
25
50
43
32
33
35
55
43
57
55
37
40
85
41
50
30
32
49
51
Plasticity
Index
(%)
10
29
10
34
26
19
13
22
31
21
30
28
15
20
58
22
34
18
14
23
26
Percent
Fines
65

85
95
87
88
77
75




78
70
99
77
95
52
85
94
90
Percent
Clay
37

22
47
32
35
27
45
45
29
39
33
37
25
57
38
47
16
44
43
36
Standard Proctor
wopt
(%)
10.2

12.3
17.9


14.1





18
16.2
25.8
15.8
20.3
13

18.5
18
(Yd) max
kN/m3
20.1

19
16.8


18.6





17
16.7
14.6
17
16.4
18.7

17.2
17
Modified Proctor
wopt
(%)
9



14.3
13.5

14.5
12.7
16.6
21.7
23






10.5

11.8
(Yd) max
kN/m3
21.3



18.6
19.5

18.8
18.6
18.7
17.3
16.6






20.1

18.5
Compaction Criterion
w > wopt; yd > 90% MP

w > wopt; yd > 95% SP
wopt+2%90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt+2%; Yd > 90% MP
wopr2%90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
wopt+2%90% MP
w > wopt+4%; Yd > 98% SP
w > wopt; Yd > 1 00% SP
wopt+2%90% MP
wopt+2%90% SP
Si> 78.5%; Yd > 90% MP
w > wopt; Yd > 90% MP
Si > 82%
w > wopt; Yd > 95% SP

-------
Table 4-1. Field Hydraulic Conductivity, Soil Characteristics, and Compaction Data for Natural Clay ... (Cont.).
Site
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Field
Hydraulic
Conductivity
(cm/s)
Value
2.0E-08
1.4E-08
1.5E-08
8.0E-09
2.0E-07
1.8E-07
9.0E-08
1.7E-08
1.1E-07
6.0E-08
3.9E-08
3.9E-08
4.0E-07
3.7E-08
3.0E-08
1.3E-08
3.6E-08
3.5E-09
2.2E-08
1.0E-07
8.0E-08
Method
SDRI
LYS
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
Liquid
Limit
(%)
63
39
67
53
33
31
35
27
32
40
45
29
44
39
36
36
21
21
101
47
69
Plasticity
Index
(%)
42
18
46
41
19
18
19
10
19
24
27
15
16
19
17
17
7
7
71
30
45
Percent
Fines
96
73
94
88
85
74
89
76

58
99
87
96
97
74
48
60
60
98
66
79
Percent
Clay

30
53
36
37
26
41
28

23
42
40


30
16


49

49
Standard Proctor
wopt
(%)
20.5
20
21.5
16.1
17.5
16.5
16.6
13

12.4


17.3
22.2
13.2
12.4
10.3
10.3
31.6
19.5
23.4
(Yd) max
kN/m3
16.3
16.5
16.3
18
17.7
17.8
17.5
19.1

19.3


17.1
16.4
18.3
19
20.4
20.4
13.4
16.3
15.1
Modified Proctor
wopt
(%)


16
11.5
12.2
12.5
11.5
9
14

11
13.3









(Yd) max
kN/m3


18.4
19.8
19.3
19.4
19.4
20.5
18.6

19.9
18.9









Compaction Criterion
Si > 85%
w > wopt; yd > 90% MP
w > wopt; yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 90% MP
wopr2%90% MP

w > wopt; Yd > 95% SP
wopt90% MP
w > wopt; Yd > 90% MP
w > wopt; Yd > 95% SP
w > wopt; Yd > 95% SP
wopt95% SP
w > wopt; Yd > 95% SP
w > wopt; Yd > 90% MP
1 1 % 90% SP
wopt+1%92% SP
Yd > 95% SP


-------
Table 4-1. Field Hydraulic Conductivity, Soil Characteristics, and Compaction Data for Natural Clay ...  (Cont.).
Site
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
Field
Hydraulic
Conductivity
(cm/s)
Value
7.0E-08
2.0E-07
3.7E-08
2.0E-08
5.0E-08
4.0E-08
5.0E-08
2.6E-07
3.0E-07
1.1E-07
2.2E-08
7.0E-08
1.3E-07
2.4E-08
5.6E-08
5.0E-08
9.4E-08
1.2E-07
3.7E-08
3.1E-07
3.9E-07
Method
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
Liquid
Limit
(%)
62
62
44
35
39
41
42
43
40
37
54

66
66
69
69
69
68
68
51
51
Plasticity
Index
(%)
42
42
28
16
24
23
22
24
22
18
31

35
35
38
38
38
35
35
20
20
Percent
Fines
86
86
70
98
69
86
86
86
86
73


93
93
98
98
98
95
95
73
73
Percent
Clay



22





38
40










Standard Proctor
wopt
(%)
22.4
22.4
19.5
23.3
14.6
18
18
18
18
19.9
19.9

27.4

26.8
26.8
26.8
26.6
26.6
20.2
20.2
(Yd) max
kN/m3
15.4
15.4
16.4
15.4
17.7
16.7
16.7
16.7
16.7
16.5
16.4

14.5

14.6
14.6
14.6
14.6
14.6
15.9
15.9
Modified Proctor
wopt
(%)





13.3
13.3
13.3
13.3




27.4







(Yd) max
kN/m3





18.7
18.7
18.7
18.7




14.5







Compaction Criterion
w > w0pt+1%; yd > 95% SP
w > wopt+1%; yd > 90% SP
wopt95% SP
wopt+1%95% SP
wopt+1%90% SP
wopt95% SP
wopt95% SP
wopt 95% SP
w > wopt+3%; Yd > 95% SP
w > wopt+3%; Yd > 95% SP












-------
Table 4-1. Field Hydraulic Conductivity, Soil Characteristics, and Compaction Data for Natural Clay ...  (Cont.).
Site
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
Field
Hydraulic
Conductivity
(cm/s)
Value
2.3E-07
1.8E-07
1.2E-08
8.3E-08
2.3E-08
1.3E-08
4.0E-08
8.3E-08
2.0E-08
8.0E-08
1.0E-09
5.0E-08
2.0E-08
2.0E-08
2.0E-08
4.5E-08
4.0E-08
1.5E-07
3.0E-08
4.5E-08
Method
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
Liquid
Limit
(%)
47
47
50
49
35
22
42
29
36
76
56

37
32
32
62
52
47

39
Plasticity
Index
(%)
30
31
29
27
17
9
26
19
20
53
40

17
13
16
41
35
22

16
Percent
Fines
66
66
75
62
67
50
88
83
85

64

92


82
84

84
81
Percent
Clay




22
16
45
34
35




19
25



54
48
Standard Proctor
wopt
(%)
19.5

19
19.3
14.8
10


18
21
18
21
19.2
9.9
11.5
25
19.6
25

18.2
(Yd) max
kN/m3
16.3

16.1
16.1
17.7
19.9


16.5
15.5
16.9
15.6
16.6
19.7
19.6
14.9
15.9
15.3

17.6
Modified Proctor
wopt
(%)

13.5


11.5
8.5
14.9
12.2







17.8
14.4



(Yd) max
kN/m3

19.2


19
21.4
18.7
19.6







16.5
18



Compaction Criterion
Yd > 95% SP
Yd>91% MP
w0pt+1%95% SP
w0pt+1%95% SP
w0pt+1%95% SP
wopt+1%95% SP
wopr2%90% MP
wopt_2%90% MP
wopt+1%95% SP
w > wopt; Yd > 95% SP
w > wopt; Yd > 95% SP
w0pt+3%95% SP
w0pt+3%95% SP
w > wopt; Yd > 98% SP
w > Wopt+1 .5%; Yd>94% SP
w > wopt; Yd > 95% SP
w > wopt; Yd > 95% SP
w > Wopt+4%
Yd > 96% SP
Yd > 96% SP

-------
        Table 4-1. Field Hydraulic Conductivity, Soil Characteristics, and Compaction Data for Natural Clay ...  (Cont.).
Site
84
85
86
87
88
89
Field
Hydraulic
Conductivity
(cm/s)
Value
1.3E-07
2.8E-08
1.5E-08
1.4E-08
2.3E-08
2.1E-08
Method
SDRI
SDRI
SDRI
SDRI
SDRI
SDRI
Liquid
Limit
(%)


43
43
25
25
Plasticity
Index
(%)


24
24
14
14
Percent
Fines


84
84
70
70
Percent
Clay


37
37
29
29
Standard Proctor
wopt
(%)


17.7
17.7
11.6
11.6
(Yd) max
kN/m3


17.1
17.1
19.1
19.1
Modified Proctor
wopt
(%)






(Yd) max
kN/m3






Compaction Criterion
Yd > 90% MP
Yd > 90% MP
wopt95% SP
wopt95% SP
wopt95% SP
wopt95% SP
        Notes:
CO
SDRI = Sealed Double Ring Infiltrometer
LYS = Lysimeter (Underdrain beneath Liner)
wopt = Optimum Water Content
(Yd)max = Maximum Dry Unit Weight
w = Water Content
Yd = Dry Unit Weight
SP = Standard Proctor
MP = Modified Proctor
Percent Fines = percent passing the No. 200 (0.075 mm) sieve
Percent Clay = percent finer than 0.002 mm

-------
      Table 4-2.  Summary of Construction and Additional Hydraulic Conductivity Data.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Compactor
Mass
(kg)
32,400

30,000
19,800
36,000
32,400
32,400
12,600




32,400


39,000
19,800
25,000
32,400
32,400
32,400
Number of
Passes
Per Lift
6

4

6
5
4
4




4



6
4
5
8
6
Lift
Thickness
(mm)
150

150
150
150
150
150
150
150
150
150
150
150
170
200
170
150
150
150
150
150
Number
of
Lifts
6

8
4
10
6
8
6
10
10
10
10
8
6
7
5
4
5
10
6
6
Avg. Water
Content
(%)
10.3

13.8
21.3
17.3
13.8
17.2
15.3
19.6
17.8
25.4
26
20.7
17
30.8
19.8
23.3
16.6
13.6
17.6
19.5
Avg. Dry
Density
(kN/m3)
19.8

19.4
16
17.3
19
17.7
17.7
17
16.9
16
16.1
16.7
16.8
14.1
16.1
15.7
17.4
19
16.9
16.9
Po
(%)
44

98
80
95
32
88
8
90
50
75
78
100
78
98
91
100
85
81
8
80
Lab. Hydraulic
Conductivity
(cm/s)
3.2E-09

8.0E-09
5.0E-09
8.8E-09
2.4E-08
8.4E-08
9.0E-09
1.0E-08
8.0E-09
2.0E-09
3.0E-09
1.3E-08
4.8E-08
4.4E-09
3.7E-08
3.0E-09
1.5E-08
1.9E-08
3.0E-08
3.1E-07
Field Hyd. Cond.
from TSB
Test (cm/s)






4.3E-08










9.2E-09



Lab. Hyd. Cond. from
300 mm Diameter
Samples (cm/s)
2.6E-7



4E-08







1.4E-08

1.6E-08

5.0E-09
1.4E-08


2.2E-07
CD

-------
Table 4-2.  Summary of Construction and Additional Hydraulic Conductivity Data (Cont.).
Site
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Compactor
Mass
(kg)


18,900
18,900
18,900
18,900
18,900
32,400
27,000

32,400
19,800
19,800
32,400
17,100


19,800
59 kg/lin.
cm
19,800
Number of
Passes
Per Lift







6




4
6
8
4
12
12

40
Lift
Thickness
(mm)

150
150
150
150
150
150
170
150
300
150
150
150
150
150
150
150
130
150
150
Number
of
Lifts

10
5
5
6
6
6
9
6
2
8
8
3
6
6
5
6
6
6
4
Avg. Water
Content
(%)

22
23.6
18.9
15.5
13.5
16.2
13.9
16.2
13.1
13.9
13.4
17.8
20.7
15.5
14.1
11.5
11.6
35.5
21.9
Avg. Dry
Density
(kN/m3)

16.4
15.8
16.9
17.6
18
17.7
18.8
18.6
19.1
19.2
18.7
17.1
16.8
17.6
18.2
20.4
17.9
12.8
16
Po
(%)

89
81
71
17
6
57
84
65
75
92
80
45
78
77
45

10
100
92
Lab. Hydraulic
Conductivity
(cm/s)
2.4E-08
1.5E-08
9.0E-09
2.3E-09
2.9E-09
3.0E-08
1.9E-08
2.2E-08
3.0E-08
1.6E-08
3.0E-08
1.3E-08
1.5E-08
3.0E-08
9.1E-09
4.9E-08

2.6E-08
3.5E-09
5.5E-09
Field Hyd. Cond.
from TSB
Test (cm/s)









4.7E-08








1.6E-08

Lab. Hyd. Cond. from
300 mm Diameter
Samples (cm/s)


1.1E-08
6.0E-09
1.8E-07
1.5E-07
1.7E-07
1.7E-08




3.5E-07






4.1E-09

-------
Table 4-2.  Summary of Construction and Additional Hydraulic Conductivity Data (Cont.).
Site
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
Compactor
Mass
(kg)

7,200
7,200
19,800
10,900
19,800
19,800
19,800
19,800
19,800



19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
Number of
Passes
Per Lift

16
8
42
8


3
3


4

12
10
6
12
12
12
12
12
Lift
Thickness
(mm)
150
150
150
150
225
85
150
150
150
150


150
160
130
140
150
170
170
190
150
Number
of
Lifts
4
5
5
4
5
10
4
4
4
4

4
4
4
4
4
4
4
4
4
4
Avg. Water
Content
(%)
25
23.4
24.2
19.8
27.3
16.5
17.8
18.9
18.6
17.8
21.2
21.6

27
30.6
29.6
30.7
29.4
26.8
29.8
24.6
Avg. Dry
Density
(kN/m3)
15.1
15.4
15
16.3
15.4
17.7
17
16.7
16.9
17
16.1
15.5

15
14.2
14
14.3
14.4
15.1
14.4
15.4
Po
(%)
81
63
47
71
100
100
75
86
84
73
67


100
100
100
100
100
100
100
100
Lab. Hydraulic
Conductivity
(cm/s)

2.4E-09
2.4E-09
5.8E-09
1.5E-08

1.1E-08
5.1E-08
7.4E-08
4.1E-08

1.7E-08

8.1E-08
2.8E-08
3.4E-08
2.5E-08
2.7E-08
3.4E-08
4.3E-08
1.6E-07
Field Hyd. Cond.
from TSB
Test (cm/s)






2.1E-08
3.2E-07
7.5E-08
1.1E-09

1.2E-08









Lab. Hyd. Cond. from
300 mm Diameter
Samples (cm/s)






4.8E-08
7.7E-08
3.1E-06
5.3E-07












-------
        Table 4-2.  Summary of Construction and Additional Hydraulic Conductivity Data (Cont.).
Site
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
Compactor
Mass
(kg)
19,800
19,800
19,800
19,800
19,800
19,800
19,800


32,400
19,800
19,800
19,800
19,800
10,200
10,200
19,800
19,800
16,200
32,400-
Numberof
Passes
Per Lift
12
40
80
2
2
6
6
7
7
8
22
22

6
10
4
8
8
4
62
Lift
Thickness
(mm)
230
150
150
100
100
150
150
150
150
150
100
100
150
150
100
100
150
150
150
150
Number
of
Lifts
4
4
4
10
11
4
4
6
6
4
7
7
4
4
9
9
8
8
4
4
Avg. Water
Content
(%)
22.7
21.6
17.2
21.7
21.4
17.6
11.5
20.6
14.3
23.7
25.2
19.6
25.4
21.8
11
12.4
28.2
23.1
28
17.8
Avg. Dry
Density
(kN/m3)
15.4
16
17.4
17.2
17.2
18
19.4
16.1
18
15.5
14.8
16.1
15.2
15.9
19.2
18.8
16.2
14.9
14.2
17.1
Po
(%)
100
88
0
95
98
100
75
60
64
100
97
47
100
86
37
2




Lab. Hydraulic
Conductivity
(cm/s)
1.7E-07
5.5E-09

3.7E-08
3.0E-08
7.8E-09
2.1E-08
2.0E-08
2.0E-08
1.4E-08



4.7E-08


3.3E-09
1.8E-09
4.2E-08
1.5E-08
Field Hyd. Cond.
from TSB
Test (cm/s)



1.1E-08
8.5E-08
2.6E-08
5.6E-08













Lab. Hyd. Cond. from
300 mm Diameter
Samples (cm/s)

4.1E-09


















IV)

-------
Table 4-2. Summary of Construction and Additional Hydraulic Conductivity Data (Cont.).
Site
83
84
85
86
87
88
89
Compactor
Mass
(kg)
32,400-
19,800
19,800
19,800
19,800
19,800
19,800
Number of
Passes
Per Lift
62


4
7
4
6
Lift
Thickness
(mm)
150
200
200
150
150
150
150
Number
of
Lifts
4
20
20
6
6
6
6
Avg. Water
Content
(%)
19.3


20.5
20.4
13.2
13.2
Avg. Dry
Density
(kN/m3)
17.3


16.6
16.6
19.2
19.1
Po
(%)



100
95
100
100
Lab. Hydraulic
Conductivity
(cm/s)
1.7E-08


2.2E-08
2.6E-08
3.9E-08
3.1E-08
Field Hyd. Cond.
from TSB
Test (cm/s)







Lab. Hyd. Cond. from
300 mm Diameter
Samples (cm/s)







Notes:
   P0 = Percent of Moisture-Density Points on or Above the Line of Optimums
   TSB = Two-Stage Borehole Permeability Test

-------
The soils included in the database covered a broad spectrum of material types (Fig.
4-2). Nearly all the soils used were classified as either CL or CH soils.
         80
         60
X
CD
_c
>,
-I—'
;o
'-4—•

-------
Work by Benson et al. (1994) indicates that an area of approximately 0.1 m2 is the
minimum required to obtain a representative hydraulic conductivity measurement with
field infiltrometers or laboratory tests on "undisturbed" samples from the field. For
practically all of the CCL test pads in the database, only the SDRI test met this criterion.
Thus, the SDRI test was selected as the test that would be used to define field k for the
81 CCL test pads that were included in the database.  Most SDRIs permeated an area
of 2.3 m2, but a few were as small as 1.4 m2.

For the 8 actual in-service CCLs beneath landfills, lysimeters were used to determine
kpieid- One lysimeter covered an area of 0.37 m2, but the rest covered 64 to 225 m2.

Hydraulic conductivity was calculated from SDRI tests using the wetting-front method
with a wetting-front suction of zero (Trautwein and Boutwell, 1994).  No correction was
made for swelling of the soil. These assumptions theoretically result in a computed k
that is slightly larger than the actual k.

It is important to note that hydraulic conductivity is sensitive to the effective vertical
stress acting on the  CCL at the time the hydraulic conductivity is measured.  For test
pads, there is very little effective overburden pressure acting on the CCL.  For actual in-
service liners overlain by solid waste, the effective stress is relatively large. There is a
tendency for decreasing hydraulic conductivity with increasing effective compressive
stress because compression causes a reduction in porosity and, hence, a reduction in
hydraulic conductivity.

Because increasing  overburden stresses tends to reduce hydraulic conductivity, it would
be expected that, on average, kFieid determined from lysimeters would tend to be lower
than values determined from SDRI tests on test pads.  It should be noted, however, that
even though a CCL  in a landfill liner system will be subjected to significant overburden
stress as waste is placed, the overburden  stress is low during the period of initial waste
placement. Also, the effective vertical stress is low permanently in liquid waste
impoundments and final cover systems.

4.1.3 Hydraulic Conductivity Results

4.1.3.1 Field Hydraulic Conductivity
A histogram of the logarithms of hydraulic  conductivity is plotted in Fig.  4-3.  Hydraulic
conductivity is log normally distributed.

For the 89 CCLs in the database, the average  kFieid is 4.0 x 10~8 cm/s. The highest  kFieid
is 8 x 10~7 cm/s and  the lowest is 1.0 x 10~9 cm/s. All averages for hydraulic conductivity
are the geometric mean (i.e., based on log average).  For the 81 CCL test  pads  whose
kpieid was determined by SDRI testing, the  average kFieid was 5 x 10~8 cm/s.  For the 8 in-
place liners monitored with lysimeters, the average kFieid was 9 x 10~9 cm/s.  The lower
                                      4-15

-------
     for the in-place liners was probably affected by three factors: (1) all but one of the
in-place liners had a thickness > 1 m, which is thicker than most test pads; (2) the
compressive stress was much higher on the in-place liners compared to test pads; and
(3) the in-place  liners were all from the northeastern section of the U.S. or Canada,
where the clays tend to be wet.
         15
         10
     c
     CD
             Geometric Mean = -9.40
             (k = 4.0 x 10~8 cm/s)
                      n
 Common
 Regulatory
"Criterion of
 k < 1 x 10'7 cm/s
  n
          -10        -9         -8         -7        -6         -5

                   Log Field Hydraulic Conductivity (cm/s)
Figure 4-3.   Histogram of field hydraulic conductivities for CCLs constructed
             from natural clay materials.
All of the full-scale clay liners and test pads in the database were constructed with the
expectation that the kFieid would be no greater than 1 x 10~7 cm/s. However, while 67 of
the CCLs in the database (or 75% of the total) did achieve the objective of k < 1 x 10~7
cm/s, 22 test pads (or 25% of the total) did not. A histogram of kFieid for the failing test
pads is presented in Figure 4-4.  Most of the test pads that failed just barely failed. Of
the 22 test pads that failed to achieve field k < 1 x 10~7 cm/s, 12  had a k < 2 x 10~7 cm/s
and 18 had k < 3 x 10~7 cm/s.  Nevertheless, a hydraulic conductivity of 1 x 10~7 cm/s is
the maximum typically allowed by regulation, and failure to meet this requirement is
usually grounds for denying a construction permit.  The causes for so many of the clay
                                       4-16

-------
liners failing to meet the 1 x 10~7 cm/s design objective are discussed later, but to
summarize, the soils were unsuitable in a few cases, but in most cases, the compaction
conditions were inadequate.
 o
       0
       1E-07
 3E-07          5E-07          7E-07

Field Hydraulic Conductivity (cm/s)
9E-07
Figure 4-4.  Histogram of field hydraulic conductivities that failed to achieve the
            desired k < 1 x 10~7 cm/s for CCLs constructed from natural clay
            materials.
The database includes information from 8 lysimeters that were located beneath actual
landfill liners.  The average kFieid from the 8 lysimeters is 9 x 10~9 cm/s. The average
kpieid from the 81 test pads is significantly larger (5 x 10"8 cm/s). Of the factors that
might tend to cause kFieid from an SDRI test on a test pad to be larger than kFieid of an in-
service liner at a landfill, the larger compressive stress acting on the CCL in an actual
landfill liner is probably the most important one. However, CCLs in the field are
                                      4-17

-------
challenged to perform well at low compressive stress in final cover applications and in
liner systems during the initial placement of waste in a  landfill cell.

4.1.3.2  Hydraulic Conductivity from Small-Diameter Samples
One of the most contentious issues in clay liner testing is the degree to which laboratory
hydraulic conductivity tests performed on relatively undisturbed, 75-mm-diameter
samples obtained from the field provide an accurate indication of the large-scale kFieid.
For poorly built liners, the laboratory hydraulic conductivity (kLab) can be orders of
magnitude lower than kFieid. Figure 4-5 shows the relationship between kLab and kFieid,
and Figure 4-6 relates the hydraulic conductivity ratio kLab/kFieid to kFieid- The linear
regression in Figure 4-6 and some subsequent figures  is shown for informational
purposes only and does not imply that there is or should be a linear correlation.  In
general, kLab tends to be less than kFieid, The average kLab/kFieid is 0.4, but the ratio is as
low as 0.0038 and as high as 7.8.

The reasons for differences between kLab and kFieid are  many, as discussed in the
literature (e.g., Benson and Boutwell, 1992; Benson et  al., 1994).  It is generally
believed that small-scale measurements will yield the same k as large-scale
measurements if the clay liner is well constructed and reasonably homogeneous. On
the other hand, a poorly built liner will be heterogeneous and will contain hydraulic
defects on a scale that is too large to be properly represented in a 75-mm-diameter
sample. As indicated in Figure 4-6, there is a tendency for better correlation between
kLab and kFieid (i.e., kLab/kFieid ~ 1)for small values of kFieid.  Small values of kFieid are
presumably achieved when good construction specifications, materials, construction
procedures, and QA practices are used.  In other words, there is good correlation
between kLab and kFieid for good-quality liners with a  very low kFieid, and a tendency for
the kLab to be significantly smaller than kFieid as kFieid increases.  Not  surprisingly, the
worst correlation between  kLab and kFieid (i.e., lowest value of kLab/kFieid) is for the highest
value of kpieid in the database (Figure 4-6).

Laboratory k tests on 75-mm-diameter samples obtained from the field-compacted clay
liner are a routine part of many CQA programs, and the results often form the principal
basis for pass/fail decisions.  Of the 89 CCLs in the database,  laboratory-measured
hydraulic conductivity's were reported for 75 CCLs.  Twenty-two CCLs in the database
failed to achieve a kFieid of 1 x 10~7 cm/s or less.  Of the 18 failing CCLs for which kLab
data are available, laboratory k tests incorrectly indicated that 15 of the CCLs passed
the 1 x 10~7 cm/s requirement.  The laboratory measurements of hydraulic conductivity
provided a false positive (pass) in 15 of 18 (83%) of the failing test pads.  The kLab
values correctly predicted failing kpieid in just 3 of the 18 (17%) test pads. The database
clearly shows that kLab tests have a strong tendency to  yield  unconservative (passing)
values for failing CCLs.  On the other hand, of the 67 instances in which kpieid was < 1 x
10~7 cm/s, all  the laboratory-measured hydraulic conductivity's were < 1 x 10~7 cm/s.
                                       4-18

-------
          S
               1E-6
         '^   1E-7
        -.
        :> E
        o ro
        D
        •D-D
        5 <»
        o-
        ^ zs
        :i w
        3 T3
        2 =
        •D — '
        cc
        -
        cc
1E-8
               1E-9
                                                                     o
                  1E-9
                    1E-8             1E-07             1E-06

                Field Hydraulic Conductivity (cm/s)
Figure 4-5.   Relationship between field hydraulic conductivity (kFjeid) and the
             hydraulic conductivity measured in the laboratory (kLab
Thus, the database shows:

   •  If ki_ab is < 1  x 10"7 cm/s, the kFieid may or may not be < 1 x 10"7 cm/s. Of the 72
      CCLs with ki_ab < 1 x 10"7 cm/s, 57 had kFieid values that achieved the hydraulic
      conductivity criterion of 1 x 10~7 cm/s or less.  Based on this database,  if kLab ^
      1 x 10"7 cm/s, there is a 57/72 = 79% probability that kFieid will also be < 1 x 10"7
      cm/s.
   •  If the ki_ab is > 1 x 10"7 cm/s, this database shows that 100% of the time kFieid was
      also >  1 x 10"7 cm/s.
   •  Laboratory hydraulic conductivity tests are not that useful for discriminating
      between well and poorly built liners.
                                       4-19

-------
         10
   CD
   L_
  ^
   .Q
   CO
         0.1
        0.01
       0.001
Linear Regression
ft = o.41
                          o       o
                                o
                                                     8
          1E-9
            1E-8
1E-07
1E-06
                       Field Hydraulic Conductivity (cm/s)

Figure 4-6.   Ratio of field hydraulic conductivity to laboratory-measured
             hydraulic conductivity (kFjeid/kLab) versus field hydraulic conductivity
             for clay liners constructed from natural clay material.
4.1.4 Soil Characteristics
Construction of a CCL with  a hydraulic conductivity < 1 x 10~7 cm/s requires the use of
suitable soils.  The engineer usually specifies a minimum plasticity index (PI) for the soil
(PI typically > 12 to 15%) and a minimum percentage of fines (typically > 50% passing
the No. 200 sieve). A minimum percentage of clay (fraction finer than 2 |j,m)  is also
sometimes required (e.g., > 20 to 25% clay).  Values are also sometimes specified for
the liquid limit (LL) of the soil.  The succeeding subsections present an assessment of
the significance of LL, PI, percentage of fine material, and percentage of clay.

4.1.4.1  Liquid Limit (LL)
The relationship between kFieid and LL is shown in  Figure 4-7. Linear regression
indicates an essentially flat  best-fit curve, and r2 =  0.00 confirms the obvious
observation that there is no correlation between kFieid and the LL of the soil. All linear
                                       4-20

-------
regression analyses presented herein were calculated using the software package
Cricket Graph.
       1E-06
       1E-07
c
O
O
o
"5
03
•o
X

"o3
       1 E-08
       1E-09
                                  0   0
                            o     0
                             o             o
                            °
                                  0
                                  °

                                           o
                                           o
                                              8
                       o
                              o  oo
                              coo 8
                                     -e—•-
                                                      Linear Regression,
                                                      ^ = 0.00
             0
                      25          50          75          100

                              Liquid  Limit of Soil (%)
125
Figure 4-7.  Field-measured hydraulic conductivity versus liquid limit of the soil
            for CCLs constructed from natural clay material.
4.1.4.2  Plasticity index (Pi)
The relationship between kFieid and PI is shown in Figure 4-8. There is a slight tendency
for decreasing kFieid with  increasing PI, but the correlation is very weak.

Because the PI of the soil is perhaps the single most frequently used indicator of the
quality of a natural soil material for use in a CCL, it is useful to consider whether some
minimum value of PI provides a good indication of the suitability of a soil for use in a
CCL.  Figure 4-9 shows the percentage of CCLs  that attained a kFieid < 1 x 10~7 cm/s as
a function of PI. Soils with Pi's greater than 30 to 40% have a higher probability of
                                      4-21

-------
achieving kFieid < 1 x 10"7cm/s than soils of lower PI. However, of the 11 CCLs for which
the PI was lower than 15%, 8 had kFieid < 1 x 10~7 cm/s, confirming that clays of low
plasticity can be used successfully.  The database does not support the concept that
there is a minimum PI that should be specified for clay liner materials, although no liners
with Pi's < 7 are included in the database.
          1E-06
          1E-09
                             20           40            60

                               Plasticity Index (%) of Soil
80
Figure 4-8.  Field-measured hydraulic conductivity versus plasticity index of the
            soil for CCLs constructed from natural clay material.
There have been correlations published between kLab and PI of the soil (e.g., Benson et
al., 1992). Figure 4-10 presents the relationship between kLab and PI for this database.
There is a much stronger correlation between kLab and PI, compared to kFieid and PI.
Because it is kFieid that counts, and because kFieid does not correlate well with PI (Figure
4-8), the value of plots such as the one shown in Figure 4-10, which shows a slight
trend of decreasing kLab with increasing PI, is questionable.
                                      4-22

-------
4.1.4.3  Percent Fines
As shown in Figure 4-11, no correlation was observed between kFieid and the
percentage of fines in the soil (defined as the percent on a dry weight basis passing
through a No. 200 sieve, which has openings of 0.075 mm). However, all but one soil in
the database had at least 50% fines, which is a minimum value that is commonly
specified or required.  The data support the current approach of specifying a minimum
percentage fines but providing no preference over the amount of fines in the soil, other
than to ensure that the specified minimum is met.
              100
          CO
       i!
       £ VI
       c £?
               80
               60
       1=
       o5  ™
       CL -a
          >>
         I
40
               20
                      7 to 14
                 15 to 20    21 to 30    31 to 40   41 to 71
                                Plasticity Index of Soil
Figure 4-9.  Percent of clay liners with a field hydraulic conductivity less than or
            equal to 1 x 10~7 cm/s as a function of the plasticity index of the soil.
                                     4-23

-------
                 1E-06
           P
             "-
                 1E-07
           o o
           .  TO
           "5W
           £Z
             w
           £•-5
                 1 E-08
           .0-
           CO
                 1E-09
                                                     Linear Regression,
                                                      = 0.14
20         40         60

  Plasticity Index (%) of Soil
                                                                 80
Figure 4-10.  Laboratory-measured hydraulic conductivity versus plasticity index
              of the soil for CCLs constructed from natural clay material.
                  1E-06
              't/)s
              E
              ~   1E-07
              "o
              T3
              O
              O
              O

              TO
              •^   1 E-08
              I
              p
              
-------
4.1.4.4  Clay Fraction
As shown in Figure 4-12, there is no relationship between kFieid and the percentage of
clay (defined as the fraction finer than 2 |j,m) in the soil.  Because clay content is
relatively difficult and expensive to determine, the value of clay content tests for CQA
programs is questioned.  Other factors overshadow the clay content in terms of impact
on
             1E-06
         V)

         o
!>   1E-07
"o
T3
c
o
o
o
"5
CO
•o   1E-08
         CD
             1E-09
                    Linear Regression,
                    r2 = 0.00
                                    O  o
                                                  oo°o
                                         oo      o
                                 080
                               o
                               o
                         10      20      30      40      50

                           Clay Fraction (% Finer than 2
                                                        60
Figure 4-12.  Field-measured hydraulic conductivity versus percentage of clay in
              the soil for CCLs constructed from natural clay material.
Figure 4-13 shows the relationship between kLab and clay content.  A better correlation
was found for kLab than for kFieid.  The kLab values may be more dependent on clay
content than kFieid values because construction variables play a more important role for
kpieid than kLab- As will be seen later, moisture-density conditions have a dominant effect
on the kpieid of CCLs and probably mask the comparatively subtle effects of soil
characteristics, such as clay fraction.
                                       4-25

-------
               1E-06
        P
        b1^
        --'CD
               1E-07
          -
        °E
        «ro
        "5 CO
        03 r
               1E-08
        .a -
        03
               1E-09
                                                       Linear Regression,
                                                       r2 = 0.10
                                      ,0 O
                                         O   o     O   O
                                            O
                   10
20       30        40        50        60

 Clay Fraction (% Finer than 2
Figure 4-13.   Laboratory-measured hydraulic conductivity versus percentage of
              clay in the soil for CCLs constructed from natural clay material.
4.1.5 Compaction Conditions
The results of standard Proctor and/or modified Proctor compaction tests were reported
for most test pads.  The maximum dry unit weights are plotted as a function of optimum
water content in Fig. 4-14.  The standard Proctor optimum water content averaged 4.0
percentage points higher than the modified Proctor optimum, and the maximum dry unit
weight averaged 1.7 kN/m3 more for modified Proctor.

Construction specifications for CCLs typically require that the water content fall within a
specified range and that the percent compaction (defined as the dry unit weight divided
by the maximum dry unit weight from a specified compaction test, times 100%) equal or
exceed a specified minimum value.  The concept is illustrated in Fig. 4-15.  However,
the recommended procedure (e.g., Daniel and Koerner, 1995) involves testing the soil
to determine the appropriate range of water content and dry unit weight to achieve the
desired hydraulic conductivity (Fig. 4-16). However, when the data are plotted as
recommended in Fig. 4-16C, no single acceptable zone of water content-density is
observed (Fig. 4-17).  It appears that each material should be evaluated on its own
merits.
                                      4-26

-------
           CO
            E
            CT
            x
            03
                 25
                 20
                 15
                 10
                                                 • Standard Proctor
                                                 n Modified Proctor
                                                  Zero Air Voids Curve
                       Notes:
                       1. S = Degree of Saturation
                       2. Assumed Specific Gravity
                          of Solids = 2.70
                               10
20
30
                  S=100% '
                  S=90%
                  S=80%

                  S=70%
40
                                Optimum Water Content (%)

Figure 4-14.  Results of laboratory compaction tests.
             D)
                    I
                     d,max
                 PY,
                    d,max
                                          Zero Air Voids Curve
                                                      Acceptable Zone
                                         I Specified
                                         I   Range
                                        wopt

                                 Molding Water Content (w)


Figure 4-15.  Typical water content-density specification relying on water
              content and percent compaction (P).
                                       4-27

-------
  D)
         (A)
>,
-I—'
'>
'•g
T3
O
O
O
"5
ro
                   Maximum
                   Allowable
                   Value
                                                          o  o
(B)
          Molding Water Content
Molding Water Content
   D)
                       Acceptable
                       Zone
          (C)
                                      D)


                                     -i—*

                                     Z)


                                     Q
               Modified
               Acceptable
               Zone
           Molding Water Content
Molding Water Content
Figure 4-16. Recommended procedure for determining acceptable compaction
            zone for low hydraulic conductivity: (A) compact soil over range of
            compactive energy; (B) permeate compacted specimens; (C)
            determine acceptable water content-density zone; and (D) modify
            acceptable zone to account for other factors such as shear strength.
                                    4-28

-------
           22.5
             20
       CO
       —   175
            ''-°
             15
            12.5
                                           o  kFie|d > 1 x 10-7 cm/s

                                           •  kFie|d £ 1 x 10~7 cm/s
                          Zero Air Voids Curve .
                          for Gs = 2.70
               10
15
20
25
30
35
40
                                 Water Content
Figure 4-17.  Dry unit weight vs. water content for CCLs with hydraulic
             conductivity's that did or did not meet the hydraulic conductivity
             objective of 1 x 10"7 cm/s.
Hydraulic conductivity was found to be a function of the amount by which the water
content of the soil exceeds optimum.  Engineers have known for decades that clays
must be compacted wet of optimum in the laboratory to achieve low k. Figure 4-18
quantifies the trend for decreasing kFieid with increasing  amount wet of optimum.  Even
more significant is the relationship between how wet the soil is at the time of compaction
and the probability of achieving k < 1 x 10~7 cm/s.  This relationship, shown in Fig. 4-19,
shows a very low probability (^40%) of success when the soil is compacted dry or at
optimum, increasing  to a 100% probability of success for soils compacted 6 to 8
percentage points wet of optimum.  It should be noted that many soils cannot be
compacted so wet of optimum with heavy field equipment, and that for soils compacted
substantially wet of optimum the low shear strength of the soil should be considered.
Indeed, achieving a sufficiently wet soil to achieve low k, but not  too wet (to preserve
shear strength), is an essential trade-off in clay liner design.
                                      4-29

-------
               1E-06
            E
            o
               1 E-07
            o
            o
            O
               1E-09
            I
            p
            CD
               1E-09
                                   o Standard Proctor Compaction Used for CQA
                                   • Modified Proctor Compaction Used for CQA
                                 Linear Regression,
                                 ft = 0.26 for
                             *   Modified Proctor
Linear Regression,
ft = 0.08 for
Standard Proctor
                   -202468
                           Number of Percentage Points Average
                            Water Content Was Wet of Optimum


Figure 4-18.  Field hydraulic conductivity vs. number of percentage points above
              the optimum water content.
               ,1
              1,°
                   100
                    80
                    60
              £ 2   4°
              0. -a
                I
                    20
                         -2 toO    0 to +2    +2 to +4   +4 to +6    +6 to +8
                                   Percent Wet of Optimum


Figure 4-19.  Percent of CCLs with field hydraulic conductivity <1 x 10~7 cm/s as a
              function of the percent wet of optimum water content.
                                         4-30

-------
Percent compaction is a critical field construction control parameter for structural fills.
The percent compaction is often specified in clay liner construction, as well (e.g., Table
4-1).  As shown in Fig. 4-20, there was no significant correlation between kFieid and
percent compaction.  The  data actually show a counter-intuitive trend of increasing kFieid
with increasing percent compaction.  This trend is probably just a reflection of the high
degree of scatter (r2 = 0.03) and is not a real trend. Compaction is important (it reduces
the porosity of the soil and kneads the soil into a homogeneous mass), but percent
compaction is a poor indicator of kFieid for CCLs.
                  1E-06
             I
             T3
              O
             o
              O
             T3
              >>
                  1E-07
                  1E-08
                  1E-09
                                     o Standard Proctor Compaction Used forCQA
                                     • Modified Proctor Compaction Used for CQA
                                                o o   Linear Regression,
                                                     r2 = 0.03
                                        o 9
                     85       90       95       100      105

                                   Percent Compaction
                                                              110
Figure 4-20.  Field hydraulic conductivity vs. percent compaction.
Some engineers prefer to use degree of saturation as the specification requirement
rather than water content and dry density (e.g., Benson and Boutwell, 1992). There is a
tendency for kFieid to decrease with increasing degree of saturation (Fig. 4-21), but the
trend is not as strong as that shown in Fig. 4-18 for water content relative to optimum.
Values of degrees of saturation greater than 100% are the result of normal variability in
water content and density measurements, and (probably more significantly)
uncertainties in the specific gravity of the soil solids, which was assumed to be 2.7 but
which probably varied from soil to soil within the normal range of 2.65 to 2.8. The
relationship between the probability of achieving kFieid < 1  x 10~7 cm/s and the average
degree of saturation immediately after compaction is shown in  Fig. 4-22. While there is
a trend, it is not an especially strong one.
                                       4-31

-------
                 1E-06
 o
 3
 T3
 C
 O
 O
 o

 "5
 as



 I


 ID
                 1E-07
                 1E-08
                 1E-09
                       Linear Regression,

                       r2 = 0.03
                     70
                                80         90         100


                                   Degree of Saturation (%)
                                                                 110
Figure 4-21.   Field hydraulic conductivity vs. degree of saturation at the time of

              compaction of a CCL.
»=i
E =
CD 2
Q- -a
                   100
                    80
                    60
                    40
                    20
                        <80   80 to 85  85 to 90  90 to 95  95 to 100  > 100


                                Average Degree of Saturation
                                Immediately After Compaction



Figure 4-22. Percent of CCLs with field hydraulic conductivity <1 x 10~7 cm/s as a

             function of the degree of saturation at the time of compaction.
                                       4-32

-------
One reason why the correlation in Fig. 4-21 is not stronger is that the average degree of
saturation fails to capture the degree of scatter in the compaction data. Benson and
Boutwell (1992) found that an important factor in CCL performance is the percent of
water content-dry density (w^) points (determined from the CQA program) that lie on
or above the line of optimums. This percentage is termed "P0." The "line of optimums"
is a line (the "line" is actually slightly curved in most cases) connecting the peaks of
compaction curves developed using a range of compactive energy.  The concept is
illustrated  in Fig. 4-23. The value of P0 is calculated as illustrated in Fig. 4-24.
Hydraulic conductivity should decrease as the percentage of the measured (w-yd)
points that line on or above the line of optimums increases.  Indeed, a very strong
correlation between hydraulic conductivity and P0 was  observed, as shown in Fig.  4-25.
Of all the variables that were examined, P0 is the single best indicator of the probability
of achieving kFieid < 1  x 10~7 cm/s. The obvious implication is that specification writers
should consider requiring a minimum P0.  A suggested minimum P0 is 70% to 80%, but
the engineer should recognize that as P0  increases, shear strength (including interface
shear strength with geosynthetics) often decreases. Thus, it is critical that the P0 not be
arbitrarily raised without careful consideration given to  the effects of reduced internal
shear strength of the  CCL or interface strength with adjacent components. Of the  liners
that failed to achieve  kFieid < 1 x 10~7 cm/s, half were specified using standard Proctor
compaction (ASTM D698) as the reference, and half were specified using modified
Proctor compaction (ASTM D1557) as the reference.  Thus, there was no relationship
between the tendency of a liner to achieve the desired low hydraulic conductivity and
the method of specification. Instead, the  key factor appears to be the percentage  of
(w,  yd) points that lie on or above the line  of optimums, i.e.,  P0.

4.1.6 Construction Parameters
The construction parameters that were evaluated were the thickness of lifts, mass of the
compactor, and number of passes of the compactor. The relationship between kFieid
and thickness of individual lifts is shown in Fig. 4-26. Although there is a trend for
increasing kFieid with increasing lift thickness, the scatter is so large (r2 = 0.02) that the
trend cannot be viewed as statistically significant.  Further,  the vast majority of lifts had
a nominal compacted thickness of 150 mm, and there are relatively few data points for
other lift thicknesses. Nevertheless, for a given mass of compactor and number of
passes of the compactor over the soil, the compactive  energy  per unit volume of
compacted soil decreases as the thickness of the lift increases.  Thus, it is to be
expected that hydraulic conductivity would tend  to increase with  increasing lift thickness.
However, the data do not suggest any particular minimum lift thickness.  Other factors
(rather than lift thickness) appear to be dominant in terms of determining the field
hydraulic conductivity.
                                      4-33

-------
               Z)
               ^
               Q
                                               Zero Air Voids Curve
                                                Line of Optimums
                      Compaction
                      Curves  ~~
                     Energy A > Energy B > Energy C
                             Molding Water Content (w)

Figure 4-23.  Determination of the line of optimums.
               CD
                                  _ No. of (w,yd) Points Above Line of Optimums   lnno/
                                0     Total No. of (w,yd) Measurements
                                                     Zero Air Voids Curve
P0=(8/10)x100%
   = 80% in this
          example
                      Line of Optimums
                               Result of Field (w,yd)
                               Measurement
                                   Molding Water Content (w)
Figure 4-24.  Definition of percent of points wet of optimums (P0).
                                         4-34

-------
              OK
              ® -^
              .E.~
              _i >
              £ =>
              o> 2
              Q- -a
                    100
                     80
                     60
                     40
                          Oto20
                                  21 to 40
                                           41 to 60
                                                   61 to 80
                                                           81 to 100
                                  Percent of (w,-/^ Points Wet
                                    of the Line of Optimums

Figure 4-25. Percent of CCLs with field hydraulic conductivity 
-------
The relationship between kFieid and mass of compactor is shown in Fig. 4-27. Linear
regression showed that the data are very scattered (r2 = 0.00), and that the regression
line is nearly horizontal. Thus, there is no relationship between mass of compactor and
kpieid-  This observation comes as a surprise because laboratory tests have consistently
shown that the energy of compaction of the soil is a very significant variable, with a
strong trend of decreasing hydraulic conductivity with increasing compactive energy
(e.g., as discussed in Daniel and Koerner, 1995). However, the compactive energy is a
function of the mass of the compactor, the number of passes, and thickness of lifts. The
relationship between number of passes of the compactor per lift and  the mass of the
compactor for the CCLs in the database is shown in Fig. 4-28.  One would expect that
perhaps the heavier the compactor, the fewer the number of passes  of the compactor
on the assumption that fewer passes would be required to achieve the  required dry
density for a heavier compactor. However, Fig. 4-28 shows large scatter and practically
no trend - the data do not suggest that heavier compactors are typically associated with
fewer passes of the compactor over a given lift of soil.
                 1E-06
-0)
E


^   1E-07
"-^
o

T3

O
O
O
                 1 E-08
             I
             ^
             0
                 1E-09
                                           p
                                           00
                                            R

             o

             o
                         Linear Regression,
                         r2 = 0.00
                                                           o
                                                           0
                                                           o
                                                           o
                                                           o
10000       20000        30000

   Mass of Compactor (kg)
                                                                  40000
Figure 4-27. Field hydraulic conductivity versus mass of compactor.

The relationship between kFieid and number of passes of the compactor per lift is shown
in Fig. 4-29.  The expected trend is decreasing hydraulic conductivity with increasing
number of passes, but the data do not support this expectation.  The data are very
scattered, and there are very few data points for more than 20 passes of the compactor
                                      4-36

-------
per lift. As discussed in 4.1.5, the data do indicate that percentage of (w, yd) points lying
on or above the line of optimums (P0) is a critical parameter. The relationship between
P0 and mass of compactor and number of passes of the compactor is shown in Figs.
4-30 and 4-31, respectively. There is a trend for increasing P0 with increasing mass of
the compactor (Fig. 4-30), but the scatter is large.  Similarly, the relationship between P0
and number of passes of the compactor is weak (Fig. 4-31). The trend line in Fig. 4-31
is significantly influenced by a single point at a very large number of passes (80) for Site
65. Despite the very large number of passes, P0 was zero, probably as a result of the
low water content of the soil. Compactor mass and number of passes alone do not
capture all of the critical factors that influence  P0.
            o
            "o
            CD
            Q.
            E
            o
            O
            CD
            w
            V)
            CD
            CL
            0
            .0
            E
               100
                80
60
40
                20
O       Linear Regression,
        r2 = 0.00
                  0         10000       20000       30000       40000

                                Mass of Compactor (kg)

Figure 4-28.  Number of passes of compactor per lift of soil versus mass of
             compactor.
                                      4-37

-------
              1E-06
              1E-07
          "o
          3
          T3
          c
          o
          O
          .o
          T3
          >s
          X

          ^

          
-------
                  100
                             20        40       60       80

                           Number of Passes of Compactor per Lift
100
Figure 4-31.   Percent of (w, yd) points lying on or above the line of optimums (P0)
              versus number of passes of compactor per lift.
4.1.7 Thickness of Liner
The relationship between kFieid and the thickness of the clay liner is shown in Fig. 4-32.
Although the scatter is large (r2=0.13), there appears to be a trend for decreasing
hydraulic conductivity with increasing thickness.  As shown in Fig. 4-33, there is a
relationship between the probability that the kFieid will be < 1  x 10~7 cm/s and the
thickness of the liner.  It appears from the data that liners with a thickness of 1  m or
more have a significantly better chance of attaining a kFieid < 1 x 10~7 cm/s.
There are 21 CCLs in the database with thicknesses > 1 m. These 21 sites are not
uniformly distributed in terms of geography or data type. The thickest liners in the U.S.
tend to be constructed in Wisconsin, whose state regulations require CCLs with
thicknesses of typically > 1.2 m.  Clays in Wisconsin tend to be very wet.  Also, of the
89 CCLs in the database, kFieid was obtained by lysimeters at 8 sites  and by SDRIs at
81 sites.  Of the 8 in-place liners where kpieid was obtained by lysimeter,  7 of these
CCLs had thicknesses > 1 m.
                                      4-39

-------
         1E-06
         1E-07
    o
    O
    .o
    "5
    SE
    T3
    >,
    X
    ^
    CD
1E-08
         1E-09
                                                   Linear Regression,
    0.25      0.5      0.75       1       1.25      1.5

                 Thickness of Clay Liner (m)
                                                                 1.75
Figure 4-32. Field hydraulic conductivity versus thickness of clay liner.
Thus, the tendency seen in Fig. 4-33 for lower kFieid for liners with thicknesses > 1 m
may simply reflect the fact that the thick liners were wetter than average or were
subjected to greater effective overburden stress.  Construction with a high percentage
of data points on or above the line of optimums appears to be more important than
thickness, at least for liners with thicknesses in the range of 0.6 to 1.5 m.

Benson and Daniel (1994) found that kFieid decreases with increasing thickness of the
liner but that little benefit is gained by increasing the thickness beyond 0.6 to 0.9 m. In
the U.S., the minimum thickness of CCLs is typically 0.6 to 0.9 m, although, as
mentioned earlier, some states require thicker liners. The database does indicate a
trend of decreasing kFieid with increasing thickness (Fig. 4-33), but for liners with
thickness of 0.6 to 0.9 m  or more, construction variables (especially P0) appear to be
more important than thickness.  There may, however, be justification for increasing the
thickness to 1 m or more for critical facilities or for situations in which added redundancy
is desired.
                                       4-40

-------
             100
              80
        VI     60
              40
     CL -o
        >,
        X
              20
                   <0.50   0.51 to 0.75 0.76 to 1.00 1.01 to 1.25  1.25 to 1.53

                             Thickness of Clay Liner (m)

Figure 4-33.  Percentage of clay liners with a hydraulic conductivity < 1 x 10~7
             cm/s versus thickness of the CCL.
4.1.8 Field Hydraulic Conductivity Testing Method
Information was collected on the hydraulic conductivity measured with the TSB test,
which is experiencing expanded use for kFieid measurements on test pads because of
the lower cost, faster testing time, and greater information on variability of kFieid provided
by multiple TSB tests compared to a single SDRI test.  As shown in Fig. 4-34, the kFieid
from the TSB test correlates well with  that of the SDRI test. However, for the four test
pads in Fig. 4-34 for which the SDRI test  indicated that the test pad failed to meet the k
< 1  x 10~7 cm/s criterion, the TSB test  indicated that the test pads had met the hydraulic
conductivity criterion, thus providing four false positives (passing hydraulic conductivity's
when, based on SDRI tests, the hydraulic conductivity was not satisfactory).  Also, in
one case,  the TSB test indicated that kFieid was > 1 x 10~7 cm/s, but the SDRI test
indicated it was < 1 x 10~7 cm/s, thus providing a false negative.  It should be
emphasized that both the SDRI test and two-stage borehole test have limitations, and
the natural range of scatter in hydraulic conductivity and the testing methods may
account for much of the differences noted.
                                      4-41

-------
                1E-06

          o .c
          Z!   1 x 10~7 cm/s) in 7  of the 8 test pads.  Thus, the database
supports the use of laboratory tests on 300-mm-diameter samples from the field as an
alternative to SDRI tests.

4.1.9 Case Histories
This section focuses on an analysis of test pads that failed to achieve a hydraulic
conductivity of 1  x 10~7 cm/s or less.  The reasons why the test pads failed to meet
design objectives are identified and compared.
                                       4-42

-------
             1E-05
             1E-06
      31
      -26
      03 O
      .
      05
             1E-08
             1E-09
                1E-09        1E-08        1E-07        1E-06        1E-05

                    Field Hydraulic Conductivity (cm/s) from SDRI Test

Figure 4-35.   Laboratory-measured hydraulic conductivity on 300-mm-diameter
              samples versus field hydraulic conductivity from the sealed double
              ring infiltrometer (SDRI) test.
4.1.9.1  Test Pads at Sites 26 and 27
These test pads were constructed for the purpose of verifying that the hydraulic
conductivity of two compacted glacial tills would be < 1 x 10~7 cm/s.  The specifications
for both test pads were the same:  (1) compact wet-of-optimum, as determined by
modified Proctor, and (2) compact to a minimum dry density equal to 90% of the
maximum dry density based on modified Proctor.

For test pad 26, 75% of field water content and 100% of dry density measurements
were met the specifications.  However, only 17% of the field water content and dry
density measurements fell on or above the line of optimums,  i.e. P0 = 17%.  Thus, the
low P0 value indicates that a large portion of the test pad was compacted dry of the line
of optimums despite the fact that most of the water content-density data met the project
specifications.
                                     4-43

-------
The test pad at Site 27 had a similar problem.  At Site 27, 100% of the field data met the
density criterion and 90% of the field data met the water content criterion. The P0 value
was only 6%, however.  Compaction below the line of optimums was the problem  at
both of these sites.

Analysis of the case histories in the database shows that compaction dry-of-optimum is
the most common cause for field hydraulic conductivity values > 1 x 10~7 cm/s. The
fundamental flaw of many compaction criteria is that they allow compaction to occur dry
of the line of optimums.  However, the recommended procedure, as described by  Daniel
and Benson  (1990) and as illustrated in Fig. 4-16, will prevent the development of a
specification that allows excessive compaction dry of the line of optimums.

4.19.2  Test Pad at Site 21
The SDRI test performed at this site achieved a kFieid of 2.5 x 10~7 cm/s, whereas the
ki_ab was 3.0  x 10~7 cm/s.  The soil was a western loess that contained 90% fines and
36% clay.  The compaction criteria for the test pad were as follows:  the as-compacted
water content must be wet of standard  Proctor optimum and the dry density must  be
greater than  95% of the standard Proctor maximum. The field water content and dry
density measurements indicate that P0 at this site was 80%.

Laboratory compaction curves  for a reduced Proctor sample indicate a four-order-of-
magnitude decrease in hydraulic conductivity between water contents of 15% and 21%.
The reduced Proctor compaction test uses the same equipment as the standard Proctor
compaction test but only fifteen blows from  the hammer are applied to each lift. The
reduced Proctor test was not performed until after the SDRI test failed to assist with
identification of the causes for the high SDRI result.

Maximum reduced Proctor density was approximately 16.5 kN/m3 and maximum
standard Proctor density was 17.2 kN/m3. The density specification allowed compaction
to occur within the reduced Proctor density  range (i.e. 95% of 17.2 = 16.3 kN/m3). Field
density measurements indicate that although all of the field-measured dry density
values met the density specification, approximately 40% of the dry density values  were
less than or equal to the maximum density from reduced Proctor tests.  Therefore, an
inadequate density specification was the major reason this site had a kFieid greater than
1 x 10~7cm/s.

From  the investigation of this test pad, the usefulness of the reduced Proctor
compaction test is evident.  It is especially important to use the reduced Proctor test as
a lower bound when standard Proctor density is a compaction  criterion. In this case, a
density lower than standard Proctor was allowed, but the hydraulic conductivity of the
soil was never determined at that density. The soil at this site was very sensitive to
changes in compactive energy. If the soil had been tested at a density below standard
Proctor, the compaction criteria most likely would have been changed.
                                     4-44

-------
Another problem with this test pad, not evident in the database information, is the lower
boundary condition. The SDRI test requires a known lower boundary condition in order
to calculate kFieid accurately if the wetting front from the infiltrating water penetrates to
the bottom of the test pad. Sometimes a sand drainage layer is placed beneath the
compacted clay test pad to provide free drainage at the lower boundary of the test pad.
At this site, no sand drainage layer was installed and the subbase material had a
hydraulic conductivity of approximately 1x10" cm/s. As a result, after thirty days of
testing, the hydraulic conductivity was 2.5 x 10  cm/s, until it abruptly decreased to 1 x
10" cm/s a short period  later. It is believed that after thirty days the wetting front
reached the subbase, and then the hydraulic conductivity was misinterpreted as 1 x 10"
cm/s. Therefore, the drainage layer can provide a critical boundary condition for the
SDRI test.

4.1.9.3  Test Pads at Sites 55-63
Nine test pads were constructed at one site using four different clays and various water
contents. Three of the four soils were suitable for use as a  liner with field hydraulic
conductivity < 1 x 10"7 cm/s.  For these sites, the test pads  consistently showed that
whether or not the soil met the hydraulic conductivity criterion of 1 x 10~7 cm/s or less
was dependent almost entirely upon  the water content of the soil. When the soil was
compacted near optimum water content, the soil failed to meet the hydraulic
conductivity criterion. When the soil was typically 2 to 4% wet of optimum, the objective
was met.

One of the soils in this series of field  tests, however, was simply unsuitable.  The soil
was compacted 2% to 6% wet of optimum at Sites 61 and 62, and despite the high
water content the soil still failed to achieve a hydraulic conductivity < 1 x 10~7 cm/s.  All
of the soils were kaolin clays obtained from commercial clay pits.  The clay that was not
suitable was from a different geologic formation than the other kaolin clays that were
suitable.

One lesson learned from these tests is that, in retrospect, a comprehensive program of
laboratory hydraulic conductivity testing prior to field testing would have been desirable
for two reasons:  (1) to define the appropriate range of water content for field
compaction,  and (2) to verify that the soil was suitable.  The soil at Sites 61 and 62
probably would have been found to be unsuitable had it first been tested in the
laboratory.

4.1.9.4  Test Pads at Sites 64 and 65
Sites 64 and 65 provide  an excellent example of the futility of conventional compaction
specifications. Test pads were constructed of the same material but with a minimum
dry unit weight of 95% of standard Proctor (Site 64) or 91 % of modified Proctor (Site
65). Both test pads were compacted several percentage points wet of optimum.
                                      4-45

-------
Neither achieved the desired kFieid of 1 x 10~7 cm/s. Site 65 is particularly interesting
because it was compacted 3.7% wet of optimum to a density greater than 91% of the
maximum from modified Proctor using an astonishing 80 passes per lift of a heavy,
footed compactor. Despite all this, the liner failed to achieve the desired kFieid
apparently because  not a single (w, yd) point was on or above the line of optimums.

4.1.9.5 Test Pads at Sites 43 and 44
Although there was no statistically significant correlation between field hydraulic
conductivity and either weight of compactor or number of passes of the compactor,
there were examples in which the number of passes of the compactor was the only
variable between two test pads.  For example, Sites 43 and 44 represent two test pads
that were constructed from the same soil at essentially the same water content, and
compacted with the same compactor.  The only variable was number of passes of the
compactor: 16 passes for Site 43 and 8 passes for Site 44. The kFieid for Sites 43 and
44 were 7 x 10~8 and 2 x 10~7 cm/s, respectively, illustrating that increasing the
compactive effort can have a significant impact on field hydraulic conductivity. The soil
compacted with 16 passes of the compactor achieved 63% of the (w, yd) points on or
above the  line of optimums, whereas P0 was only 47% when 8 passes were used.

4.1.10 Practical Findings from Database
CCLs are almost always constructed with the objective of achieving a hydraulic
conductivity of 1 x 10~7 cm/s or less. All of the 89 liners and test pads in this database
were constructed for the purpose of demonstrating that the kFieid would meet this
requirement.  Despite all that has been written and learned about CCLs in the past 15
years, the  hydraulic  conductivity objective of kFieid < 1 x 10~7 cm/s was not met at  more
than one-fourth (26%) of the sites in the database. Why such poor success?

The soil was unsuitable at a few sites. It appears that in all cases involving unsuitable
soils, test pads were built without the benefit of a comprehensive laboratory testing
program prior to construction. No simple way of identifying unsuitable soils (based on
plasticity information or other index properties) was identified. The database reinforces
the recommendation that kLab tests be performed on representative samples of soil prior
to construction, e.g., following the procedures recommended by Daniel and Benson
(1990).

Perhaps the  single biggest problem identified from the database is failure to recognize
that conventional specification of water content and dry unit weight based on a minimum
percent compaction  often leads to difficulty. The problem is that this procedure does
not guarantee that any of the (w, yd) points will lie on or above the line of optimums.
Despite widespread  publication of procedures that will avoid this problem (e.g., Daniel
and Benson, 1990; Daniel and Koerner,  1995), many design professionals and
specification writers  continue to repeat the mistake. The type of specification that is not
                                      4-46

-------
generally recommended (but which is still commonly used) is shown in Fig. 4-36A; the
more appropriate and recommended approach is shown in Fig. 4-36B.

A conclusion from analysis of the database is that P0 is the single most important CQA
parameter. The definition of P0 is summarized in Fig. 4-24: P0 is the percent of water
content-dry density (w, yd) points lying on or above the line of optimums.  The line of
optimums is the locus of peaks of compaction curves developed  employing different
compaction energies (Fig. 4-23). The key to success in achieving a kFieid < 1 x 10~7 cm/s
is to ensure that a high percentage (the data indicate a minimum of 70% to 80%) of the
field-measured (w, yd) points lie on or above the line of optimums, i.e., in the shaded
zone shown in Fig. 4-36B.  It should be emphasized that it is practical to compact soil in
the shaded zone in Fig. 4-36B, particularly in the shaded area  above the line of
optimums but below the compaction curve (compaction in  this  zone could be achieved
with a smaller compactive effort than used to develop the compaction curve indicated in
the figure). The data in Fig. 4-25 provides confirmation that it is possible and practical
to achieve a large value for P0, i.e., to compact within the shaded zone shown in Fig. 4-
36B.

Although this analysis has focused on hydraulic conductivity, the authors emphasize
that other factors (such as bearing capacity, internal shear strength of a clay liner, and
interfacial shear strength with geosynthetics) are equally important considerations. The
engineer must give proper consideration to these factors and avoid the temptation to
add too much water to the clay (in order to drive down hydraulic conductivity), at the
expense of compromising other critical engineering properties  of the clay liner. As the
value of P0 is increased,  the shear strength of the soil tends to decrease.  The designer
must ensure that all  criteria, and not just hydraulic conductivity, are satisfied.

4.2  Soil-Bentonite Mixtures

Even though the most common type of CCL is one that is  made from natural soils that
contain a significant quantity of clay,  if the soils found near the waste disposal facility
are not sufficiently clayey to be suitable for direct use as a liner material, a common
alternative is to blend natural soils available on or near a site with sodium  bentonite.
Soil-bentonite liners are discussed by Daniel and Koerner (1995).

4.2.1  Database
A database of 12 test pads that had bentonite added to the natural clay soils was
developed. The database is summarized in Tables C-5 through C-8 in Appendix C. All
test pads were constructed for the purpose of demonstrating that kFieid < 1 x 10~7  cm/s.
                                      4-47

-------
D)
(U

•^

D


Q
                                        Specified
                                      Water Content
                                         Range
                      Line of
                      Optimums
                                 Zero Air
                                 Voids Curve
                           Laboraotry   Minimum yd Based
                           Compaction     ...  .
                           o..~.~       on Minimum
                                      Percent Compaction
     Curve
                                    Water Content

                        (A) Typical Type of Specification Currently Used
               D)
Line of
Optimums
                      Minimum Water
                      Content Determined
                      from Lab. Hydraulic
                      Conductivity Tests
                                  Zero Air
                                  Voids Curve
                                   Minimum Yd Based
                                   on Minimum
                                   Percent Compaction
                                    Water Content

                         (B) Recommended Compaction Specification


Figure 4-36.  Water content-density specifications indicating: (A) the traditional
             (but not recommended) type of specification, and (B) the
             recommended type of specification emphasizing compaction to
             water content-density values on or above the line of optimums.
                                      4-48

-------
In all cases, the large-scale hydraulic conductivity tests were the SDRI test. The SDRI
test results are referred to as kFieid- Additionally, the results of kLab tests obtained on thin-
walled sampling tubes were documented at several sites. Finally, one large block
sample test was performed for kLab-

There is a basic limitation of the database associated with that fact that only 12 test
pads comprise the database.  Further, the information is not complete for all 12 tests
pads. Although this is the most complete database of its type assembled to date, there
remain relatively few well-documented cases in which the large-scale performance of
soil-bentonite clay liners is documented.

4.2.2  Hydraulic  Conductivity Results
kpieid covered a comparatively narrow range of 2 x 10~9 cm/s to 1 x 10~7 cm/s.  All liners
in the database met the objective of kFieid < 1 x 10~7 cm/s. The geometric mean kFieid
was 1.9x 10~8cm/s.

The most important parameter in soil-bentonite mixtures is the amount of bentonite
added to the mixture.  kFieid is plotted as a function of percent bentonite in Fig. 4-37. As
expected, there is a tendency for kFieid to decrease when the amount of bentonite is
increased.  However, the scatter in the data is significant (r2 = 0.28), and there are
relatively few data points in the database.
                1E-06
>
"o
T3
O
O
O
"5

-I
X

15
                1 E-°7
                1E-08
                1E-09
                                          Linear Regression,
                                          ft = 0.28
                             8        10        12        14

                                    Percent Bentonite
                                                     16
Figure 4-37.  Field hydraulic conductivity versus percent bentonite in soil.
                                       4-49

-------
The results of kLab tests are compared to the results of kFieid tests in Fig. 4-38. There is
good agreement between kLab and kFieid- This suggests that all liners were reasonably
homogeneous and free of large-scale features that would tend to lead to large
differences between the results of kLab and kFieid tests.

It was found that for natural soil liner materials, the percent wet of optimum was a
significant variable.  For soil-bentonite liners, this was not found to be the case (Fig.
4-39).  Instead, percent compaction (Fig. 4-40) was found to be more significant.
Experience in the laboratory has shown that water content is much less important for
soil-bentonite admixtures than for natural soil materials, perhaps because of the high
swelling that occurs in bentonite regardless of the initial water  content.  It does appear,
however, that the more compact the soil (i.e., the higher the percent compaction), the
lower the kFieid.

The masses of the compactors used to compact the liners in the database were nearly
all identical. However, the number of passes of the compactor did vary.  kFieid was
found to slightly increase with number of passes within the very small range of number
of passes used to construct the liners that comprise the database (Fig. 4-41).
Unfortunately,  the projects that comprise the database were not sufficiently well
documented to permit determination of P0.
                 1E-06
              >
             "§   1E-07
             T3
              C
              O
             O
              o
             "5
              ro
             T3
             Is  1E-08
             o
             .Q
             ro
                 1E-09
                    1E-09          1E-08          1E-07

                              Field Hydraulic Conductivity (cm/s)
1E-06
Figure 4-38.  Laboratory-measured hydraulic conductivity vs. field-measured
             hydraulic conductivity for soil-bentonite liners.
                                      4-50

-------
i ^-uu
E
!> 1 E-°7
"o
T3
O
O
O
^
CD
•£ 1 E-08
I
T3
CD
-ip.nQ

o


o o o
° 0
i o
f
-
I •
Linear Regression,
r2 = 0.00
O
                     -1012345
                              Percentage Points Wet of Optimum

Figure 4-39.   Field hydraulic conductivity versus percentage points wet of
              optimum for soil-bentonite liners.
                  1E-06
~    1E-07
"G
T3
O
O
O
"5
.§    1 E-08
I
"CD
                  1E-09
                                            Linear Regression,
                     94         96         98         100        102
                                    Percent Compaction

Figure 4-40.  Field hydraulic conductivity vs. percent compaction for soil-
             bentonite liners.
                                       4-51

-------
          E
          o
          O
          O
          o
          
-------
 4.3.1  Omega Hills Final Cover Test Plots
 The first detailed information to be presented on the performance of CCLs in landfill
 covers was described by Montgomery and Parsons (1989).  The study involved the
 construction of three large test pads on the top of a closed municipal solid waste landfill,
 the Omega Hills landfill, which is located approximately 30 km northwest of Milwaukee,
 Wisconsin. The test plots were constructed to evaluate the performance of alternative
 final cover designs.

 The cross sections of the three test plots are shown in Fig. 4-42.  Test plot 1, consisting
 of 150 mm of topsoil overlying 1.2 m of CCL, represented the existing final cover system
 design at the time that the study was initiated. Test plot 2 involved the same thickness
 of CCL, but a thicker topsoil layer that was intended to promote better vegetative growth
 and thereby enhance evapotranspiration. Test plot 3 involved the use of a layer of
 coarse-grained  soil (sand) sandwiched between two CCLs.  The idea for the third plot
 was to take advantage of the so-called capillary barrier effect in which the coarse-
 grained soil (sand) remains unsaturated and thereby serves as a  barrier to downward
 infiltration of water.  With test  plot 3, the intention was for the sand layer to promote
 retention of water in the upper CCL, where the water could be returned to the
 atmosphere via evapotranspiration. The use of this alternative design is consistent with
 the research nature of these test plots. All test plots were constructed on 31-1:1 V
 sideslopes of the actual landfill surface.
         Test Plot 1
Test Plot 2
Test Plot 3
150 mm
       i
                      i
  1.2m
                     450 mm
                       1.2m
L
F
L
F


150 mm
j
600 mm
i
300 mm ;
j
600 mm
i







Topsoil
Compacted
Clay Liner
Sand
Compacted
Clay Liner
 Figure 4-42.  Cross-sectional view of test plot arrangement at Omega Hills
               landfill (after Montgomery and Parsons, 1989).
 The CCL material consisted of CL soil with a high silt content.  The soil was placed and
 compacted in 150-mm-thick lifts to a hydraulic conductivity < 1 x 10~7 cm/s, based on
 laboratory hydraulic conductivity tests on "undisturbed" samples of the compacted soil.
 The topsoil was an uncompacted clay loam to silty clay loam.  The intermediate sand in
                                       4-53

-------
test plot 3 was a clean, washed, medium sand. The topsoil was seeded with a mixture
of grasses.

The test plots contain two principal data collection systems. The first was a lysimeter
located beneath the test plot to collect water that percolated through the cover soils and
permit quantification of the rate of percolation. Figure 4-43 shows the plan location of
the lysimeter system, and Fig. 4-44 shows the location in profile.  The lysimeter
consisted, from top to bottom, of a GT filter, a GC drainage layer, and a GM. The
second data collection system was designed  to collect and measure surface runoff
(Figs. 4-43 and 4-44).

The test plots were constructed in 1986.  Data collection and analysis started  in
September, 1986. Measurements were obtained of precipitation, runoff, percolation,
and other parameters such as temperature.  Soil moisture content was monitored with
neutron access probes.

The 12-month period September 1986 through August 1987 was near normal. The
period of September 1987 through August 1988 was dominated by a severe drought in
1988. The summer months in 1988 were characterized by substantially below-average
rainfall and temperatures that averaged 6°C above normal.  The drought reduced the
cover vegetation to a dry, dormant state, and cracking of the surface of the cover soils
was obvious.  The third and final year of data collection saw a return to moist conditions.

At the end of three years, test pits were excavated in each test plot, outside the area of
the lysimeters. The test pits measured 3 m in length, 1.2 m in width, and 2 m  in depth.
A summary of data collected  is presented in Table 4-3.  The key parameter is  the
quantity of percolation, i.e., rate of flow of water into the lysimeter. In test plots 1  and 2,
the percolation in the first year was 2 to 7 mm/year (6 x 10"9 to 2 x 10"8 cm/s,
respectively).  However, by the third year, these values had increased to a range of 56
to 98 mm/year (2 x 10~7 and 3 x 10~7 cm/s, respectively). The test pits showed that the
CCLs in test plots 1 and 2 were in a similar condition after three years:

  • the upper 200 to 250 mm of CCL was weathered and blocky (probably from
    desiccation and/or freeze-thaw;
  • cracks 6 to 12 mm wide extended 0.9 to 1 m into the CCL;
  • roots penetrated 200 to 250 mm into the CCL in a continuous mat, and some roots
    extended into crack planes as deep as 750 mm into the CCL; and
  • the base of the CCL appeared to be undamaged.
                                      4-54

-------
                                15m
           18m
                              6 m
                  12m
                         :::::^f:^:]
Percolation
Collection
Lysi meter
(Beneath
Test Plot)
                                                         ,Test Plot
    Surface Runoff
   ^Isolation Barrier
   "To Quantify
    Surface Runoff
                                                                3H:1V
                                                                Slope
                                     Collection Pipe for Percolation
                                     Collection Lysimeter (Located
                                     Beneath Test Plot)
                                        Percolation Collection Tank
            Runoff
            Collection
            Tanks
Figure 4-43.   Plan view of test plot arrangement at Omega Hills landfill (after
              Montgomery and Parsons, 1989).
The drought conditions in the second year of the study period apparently caused severe
desiccation of the CCL, which led to the significantly increased hydraulic conductivity in
subsequent years. Although the CCL may have initially had a hydraulic conductivity of
1 x 10~7 cm/s or less, after three years, the desiccation damage caused the CCLs in test
plots to no longer have this low level of hydraulic conductivity.

Test plot 3 was designed with the intention of maintaining moisture in the upper CCL.
The percolation rate through test pad 3 remained more consistent and was found to
range from  22 to 41 mm/year (7 x 10"8  to 1.3 x 10"7 cm/s).  At the end of the three-year
study period, the upper 200 to 250 mm of the uppermost CCL was weathered and
blocky, and cracks extended through the entire thickness of the uppermost CCL.
Cracking of the uppermost CCL allowed significant amounts of water to enter the sand
drainage layer.  Discharge of water from the sand layer was found to occur within hours
of the start of precipitation events, suggesting rapid transmission of water through the
                                       4-55

-------
upper CCL due to flow through cracks. Moisture in the sand drainage layer probably
helped to protect the underlying CCL from damage. The multi-component cap in test
plot 3 did not function as anticipated.  It was expected that the sand drainage layer
would help the overlying CCL retain moisture, but the uppermost CCL quickly cracked.
Runoff Collection
Pipe
                                                    Geomembrane
                                      Percolation Collection
                                      Lysimeter
                  Percolation Collection Pipe
Figure 4-44.
Cross section of underdrain system at Omega Hills landfill (after
Montgomery and Parsons, 1989).
Table 4-3.     Summary of information concerning performance of field test plots
              at Omega Hills landfill (data from Corser and Cranston, 1991).
Test Plot
1
2
3
Year
1986-87
1987-88
1988-89
1986-87
1987-88
1988-89
1986-87
1987-88
1988-89
Precipitation
(mm)
896
579
823
896
579
823
896
579
823
Runoff
(mm)
180
38
56
109
38
51
97
38
66
Percolation
(mm)
2
5
56
7
30
98
40
22
41
                                     4-56

-------
The principal lesson learned from the Omega Hills study was that in a fairly short period
of time (3 years), CCLs overlain by 150 to 450 mm of topsoil are subject to major
desiccation, cracking, and increases in hydraulic conductivity.  The CCL was not
"survivable" with a hydraulic conductivity of 1 x 10~7 cm/s or less under these conditions.

4.3.2  Test Plots in Kettleman City, California
Corser and Cranston (1991) and Corser et al. (1992) describe test plots constructed at
a waste disposal facility located in Kettleman City, California. Three test plots were
constructed as shown in Fig. 4-45.  Test plot 1 consisted of a 900-mm-thick CCL
overlain by an exposed HOPE GM. Test plot 2 consisted of the same profile as test plot
1, except that 600 mm of topsoil covered the GM. Test plot 3 contained 450 mm of
topsoil covering the CCL, with no GM covering the CCL. A portion of the test plots was
flat, and a portion that sloped at 3H:1V. The test plots were constructed to study the
factors that influence desiccation in a CCL placed in a final cover system profile.

The CCL was a high-plasticity clay that was expected to be used to construct final cover
systems for approximately 30 ha of landfill at the site.  The clay had an  average liquid
limit of 66% and plasticity index of 48%.  The instrumentation consisted of thermistors to
monitor temperature in the soil and CCL, and tensiometers to measure  soil suction.
Corser and Cranston (1991) summarize the first 6 months of data collection.  At the end
of the six month period, the surfaces of the CCLs were exposed over an area of 1.5 m
by 1.5 m to observe and document cracking patterns.

Test plot 1 did not represent a final cover situation but is representative of a bottom liner
with an exposed HOPE GM during the construction or operations phase. The clay
exhibited some drying and cracking in areas where the HOPE was not in contact with
the CCL. In other areas where the HOPE was in contact with the CCL,  the moisture
content of the CCL at the surface increased.  It appears that the high temperature of the
exposed HOPE GM  caused  heating and drying of the underlying CCL.  In some areas
(e.g.,  around wrinkles in the  GM), moisture could migrate away via vapor transport.  In
other  areas, the moisture could condense during cooler periods, causing moistening of
the soil. In any case, there clearly was desiccation of the CCL beneath some portions
of the exposed GM.

Test plot 3 did not perform well during the summer season.  The CCL dried, and
cracking was observed at the surface when a test pit was excavated in  the fall.  In
contrast, was no evidence of drying or cracking of the CCL at test plot 2.

Although the test plots were  observed for only six months, significant deterioration of the
CCL was observed in test plots 1 and 3.  Only test plot 2, in which the CCL was covered
with a GM and  450 mm of top soil, performed well. The observations from Kettleman
City are consistent with those of Omega Hills and suggest that perhaps the only
practical way to protect a CCL from desiccation damage in typical final cover system
                                      4-57

-------
cross sections is to incorporate a GM and sufficiently thick cover soil over the GM/CCL
composite barrier.
                        Test Plot 1
          1.5 mm HOPE
          Geomembrane"
              ICompacted Clay
  900 mm
                        Test Plot 2
   1.5mm HOPE
   Geomembrane
              ICompacted Clay
                      Topsoil
  600 mm

  900 mm
                        Test Plot 3
            Topsoil
              ICompacted Ciay|
i
  450 mm

  900 mm
Figure 4-45.  Cross sections of test plots at Kettleman City facility (after Corser
            and Cranston, 1991).
                                 4-58

-------
4.3.3  Test Plots in Hamburg, Germany
Melchior et al. (1994) describe what may be the most extensive test plot program of any
constructed to date involving CCLs. Three test plots were constructed, as shown in Fig.
4-46.  All test plots were constructed on top of an existing MSW landfill.  There were two
sections for each test plot.  The upper section was located on the relatively flat portion
near the top of the landfill and sloped at 4%.  The lower half sloped more steeply at an
inclination of 51-1:1 V (20%).  The test plots were underlain with a lysimeter, much like the
Omega Hills facility (Fig. 4-44).

The CCLs at the Hamburg site were constructed in three lifts, each 200 mm thick.  The
material consisted of 17% clay, 26% silt, 52% sand, and 5% gravel. The principal clay
minerals in the clay fraction were (in decreasing abundance) illite, smectite, and
kaolinite. The liquid limit was 20%, and the plasticity index was 9%. The soil was
compacted 2% wet of optimum at an average degree of compaction of 96%. The CCL
at the Hamburg  site was significantly different from that at Omega  Hills and Kettleman
City. At Omega Hills, the low-plasticity (CL) clay contained a large amount of silt, which
can make the material vulnerable to shrinkage cracking.  The Kettleman City clay was a
highly plastic (CH) clay. At Hamburg, the soil contained more than 50% sand- and
gravel-sized particles and would therefore be classified as a clayey sand (SC). Clayey
sands tend to be less vulnerable to shrinkage cracking than  clays (especially highly
plastic clays) that contain relatively little coarse-grained particles.

The percolation  rates through the CCLs and into the lysimeters are summarized in
Table 4-4. Also shown are the drainage rates in the sand drainage layer that overlies
the CCL. The last column in Table 4-4 expresses the leakage through the CCL as a
percentage of the drainage from the sand drainage layer. The leakage as a function of
drainage is plotted vs. time in Fig. 4-47.

Test plots  1 and 3, which did not have a GM overlying the CCL, underwent a very large
increase in leakage in 1992. The summer of 1992 was extremely dry  in Hamburg, and
the subsequent fall season was very wet.  Excavations made in 1993 confirmed that the
clay liner was cracked. Barely visible fissures were observed between soil aggregates
(around 50 mm in diameter). Plant roots were observed to have reached the upper
parts of the CCLs.  Under the conditions of a hydraulic conductivity of  1 x 10"7 cm/s for
the CCL and a unit hydraulic gradient, the calculated percolation rate through the CCL
is approximately 30 mm/year.  The actual leakage rates through the CCLs at test plots 1
and 3 exceeded 30 mm/year in 1992. The apparent problem was gradual deterioration
of the CCL caused by desiccation during a particularly dry summer.

Test plot 2, on the other hand,  has  maintained a very low leakage rate. This test plot
contains a GM overlying the CCL (Fig. 4-46).
                                      4-59

-------
          Test Plot 1
                      Test Plot 2
                           Test Plot 3
    750
    250
    600
    200
  Topsoil
[Drainage)
I  Sand  III
IllCompactedl
I Clay Liner ||
 ((Drainage!
 I  Sand  II
750
250
   Topsoil
 ([Drainage!
 I  Sand  III
750
250
600
                         200
IllCompactedl
I Clay Liner I
     IDrainagel
     I  Sand  I
                      Thickness of
                      Layer (mm)
                                              400
                                              600
                                     250
 Topsoil
(Drainage!
I!  Sand  I!
                          [Compacted))
                          [ Clay Liner ||
                                           Fine Sand
                           IDrainagel
                           I   Sand  I
      Note: Geotextile Separation and
            Filtration Layers Not Shown.
Figure 4-46.  Cross sections of test plots in Hamburg, Germany (after Melchior et
            a\., 1994).
                                 4-60

-------
Table 4-4.     Summary of information concerning performance of field test plots
              at Hamburg, Germany (data from Melchior et al., 1994).
Test Plot
1
2
3
Year
1988
1989
1990
1991
1992
1988
1989
1990
1991
1992
1988
1989
1990
1991
1992
Drainage
(mm)
371
181
291
184
225
296
155
269
164
311
390
233
321
198
278
CCL Leakage
(mm)
7
8
18
9
103
3
0.6
0.4
0.5
0.8
8
14
31
32
116
Leakage /
Drainage (%)
2
4
5
5
31
1
0.4
0.1
0.3
0.3
2
6
10
16
42
The results from Hamburg are consistent with those from Omega Hills and Kettleman
City, even though the CCL material was very different.  It appears that a CCL placed in
a final cover system without a GM and soil covering the GM is likely to fail to maintain a
hydraulic conductivity < 1 x 10~7 cm/s.  If the CCL is to have a chance of maintaining this
level of hydraulic conductivity for extended periods, it appears that the CCL must be
protected with both a GM and a sufficiently thick layer of cover soil above the GM.

Melchior (1997) describes additional work performed at the Hamburg site in which two
additional test covers were constructed and  monitored. The two additional test covers
both consisted of 300 mm of topsoil underlain by 150 mm of drainage sand, which in
turn was underlain by a GCL. Two different  GT-encased, needlepunched GCLs were
used for the two different test plots. As with  the first three test covers, the two additional
test covers with GCLs were underlain by drainage sand and a GM to collect any water
that percolated through the test cover. Both GCLs performed well for about a year,  with
almost no liquid appearing in the drainage layers beneath the test covers.  However,
about a year after construction (in the fall,  following a dry summer), percolation began to
occur and was closely linked with rainfall events. Peak percolation rates were on the
order of 0.4 mm per hour (about 1 x 10~5 cm/s).  Melchior (1997) states that research on
the causes for high percolation rate is on-going, but indications are that the causes for
the increase in hydraulic conductivity of the GCLs may have been related to: (1)
penetration of the GCL by plant roots; (2) desiccation of the GCL, leading to high initial
seepage rates following major rainfall events; and (3) ion exchange (calcium was
                                      4-61

-------
apparently leached from cover soils, and replacement of sodium in the bentonite with
calcium is expected to cause an increase in hydraulic conductivity).
       CD
       O)
       ^    50
Q

2
3    40
CO
_l
"5
CD
O)
       .2

       I
       CD
       Q.
       to
       CO
       _l
       O
       O
       .c
       O)
       D
       o
       CD
       O)
       CO
       _*:
       co
       CD
      30
     20
      10
                            Test Plot 1

                            Test Plot 2

                            Test Plot 3
      0
      1987
1988
1989
1990

Year
1991
1992
1993
Figure 4-47.   Leakage through CCL as a percentage of the drainage from the
              overlying sand drainage layer plotted vs. time.
4.3.4 Final Covers in Maine
The Maine Bureau of Remediation and Waste Management (1997) reported the results
of field measurements of percolation rates through four CCLs in actual municipal solid
waste landfill covers.  All liners appear to have been constructed using methods of
construction and construction quality assurance practices that are typical of the landfill
industry.

4.3.4.1  Cumberland Site
The 2-ha Cumberland Municipal Solid Waste Landfill was closed in 1992 with a cover
system that consisted of 150 mm of vegetated topsoil that was underlain by 450 mm of
compacted silty clay,  which in turn was underlain by sand-filled trenches that served to
collect gases.  kLab tests were performed during construction and in post-construction
                                      4-62

-------
investigation programs conducted in 1994 and 1996. An SDRI test was performed in
1994.

At the time of construction, the average kLab was 5 x 10~8 cm/s. In the 1994
investigation, kLab was 1 to 2 x 10~7 cm/s. kFieid, measured with the SDRI, was 6 x 10~6
cm/s. It is not certain whether the liner originally had a kFieid < 1 x 10"7 cm/s (since no
field testing was performed at the time of construction).

4.3.4.2  Vassalboro Site
The Vassalboro Municipal Solid Waste Landfill consists of one 1.6 ha site that was
closed in 1990.  The final cover consists of 150 mm of vegetated cover (sludge
amended topsoil) overlying 450 mm of compacted glacial till CCL, which in turn was
underlain by a gas collection layer. kLab tests were performed at the time of construction
(1990) and again in 1994 and 1996. An SDRI test was performed in 1994.

ki_ab at time of construction averaged 2 x 10"7 cm/s.  In 1994, kLab ranged from 9 x 10"7 to
5 x 10~6 cm/s. The kFieid measured by SDRI test in 1994 was 2 x 10~6 cm/s.  It appears
that the hydraulic conductivity of the CCL increased about an order of magnitude from
1990 to 1994.

4.3.4.3  Yarmouth Site
The Yarmouth Municipal Solid Waste Landfill covers 2.5 ha and was closed in 1990
using a cover system consisting of 150 mm of sludge-amended topsoil overlying 450
mm of compacted silty clay, which was underlain by a gas collection  layer.

ki_ab tests performed at the time of construction (1990) indicated an average kLab of 8 x
10~8 cm/s. In a 1994 investigation,  kLab was approximately 3 x 10~7 cm/s, and in 1996,
k|_ab was found to be 2 x 10"6 to 2 x 10"5 cm/s, or about 20 to 100 times larger than  in
1990. kpieid was measured with the SDRI in 1994 and again in 1996. kFieid was 2 x 10~7
cm/s in 1994 and 2 x 10~6  cm/s in 1996. There is a clear trend of increasing hydraulic
conductivity over time, with the magnitude of increase being one to two orders of
magnitude over the six-year study period.

4.3.4.4  Waldoboro Site
The Waldoboro Municipal  Solid Waste Landfill covers 1.6 ha and was closed in 1991
with a cover system consisting of 150 mm of sludge-amended topsoil overlying 450 mm
of compacted silty clay, which in turn was underlain by a gas collection layer.

kLab tests indicated that kLab increased over time from an initial average value of about 5
x 10"8 cm/s (1991) to 1 x 10"6 cm/s  (1993) and to 3 x 10"6 cm/s (1996).  kFieid measured
with SDRI tests was 1 x 10~6 cm/s in 1993 and 4 x 10~6 cm/s in 1996. Thus, the data
indicate that the hydraulic conductivity increased about two orders of magnitude over a
five year period.
                                      4-63

-------
4.3.4.5  Discussion
The observations from these four actual cover systems are consistent with those of the
other sites mentioned previously in this section of the report.  All of the available field
performance data indicate that a CCL overlain by a relatively thin layer of topsoil (150 to
450 mm thick), and without a GM above the CCL, cannot maintain a hydraulic
conductivity of 1 x 10~7 cm/s or less.  From analysis of the condition of the four CCLs at
these sites, it appeared that desiccation was the  most significant factor leading to an
increase in kFieid. Freeze/thaw may also have contributed significantly to damage.
Penetration of plant roots into the CCL was also observed.

4.3.5 Alternative Cover Demonstration at Sandia National Laboratory
A major field demonstration project, initiated  in the mid 1990s, is underway at Sandia
National Laboratories and, although only preliminary data were available at the time of
preparation of this report, the project bears mentioning here.  The project is known as
the Alternative Landfill Cover Demonstration (ALCD), and is a large-scale field test
conducted at  Sandia National Laboratories, located on Kirtland Air Force Base in
Albuquerque, New Mexico.  The climate at the test site is semi-arid. The goal of the
ALCD is to field test, compare, and document the performance of alternative landfill
cover technologies, of various complexities and costs, with emphasis on arid and semi-
arid environments (Dwyer, 1997). The purpose of the ALCD  is to provide information on
cost, construction, and performance, so that  design engineers and regulatory agency
officials will have data on alternatives to conventional cover design.

The test plots are each 13 m wide by 100 m  long (Dwyer, 1997). All covers were
constructed with a 5% slope in all layers. Slope lengths are 50m (the test covers are
crowned at the middle half of the length). The western slopes are maintained and
monitored under natural conditions while a sprinkler system was installed on the eastern
slopes to facilitate stress testing of the covers. Two conventional covers and four
alternative covers comprise the six test covers, with cross sections as follows:

   1.   Baseline Test Cover 1 is a "RCRA Subtitle D" conventional cover, consisting of
       150 mm of topsoil underlain by 450 mm of compacted "barrier layer soil" with a
       maximum hydraulic conductivity of 1  x 10~5 cm/s (actual hydraulic conductivity
       measured on laboratory samples recovered from the  constructed barrier layer
       were in the range of 5 x 10~7 cm/s to 6 x 10~6 cm/s, and an in situ hydraulic
       conductivity test yielded a hydraulic conductivity of 5 x 10~7 cm/s).
   2.   Baseline Test Cover 2 is a "RCRA Subtitle C" conventional cover, consisting
       (from  top to bottom) of 600 mm of topsoil, a GT separator/filter, 300 mm of sand
       drainage material, a 1-mm-thick linear low density polyethylene GM, and 600
       mm of compacted clay with a design hydraulic conductivity less than or equal to
       1 x 10~7 cm/s (although an in situ hydraulic conductivity test indicated a hydraulic
                                      4-64

-------
       conductivity of 8 x 10~7 cm/s, with the comparatively large hydraulic conductivity
       thought to have been caused by desiccation cracking during construction).
  3.   Alternative Test Cover 1 is essentially identical to the RCRA Subtitle C cover,
       except that it incorporates a GCL rather than CCL and (very significantly), the
       GM component was punctured with 8 holes, each measuring 1 cm2, to simulate
       defects in the GM.
  4.   Alternative Test Cover 2 is a capillary barrier, which makes use of a clean,
       granular layer below a topsoil layer to provide a capillary break between the
       topsoil and underlying soils, thus promoting moisture  retention in the topsoil
       layer.  So long as the granular layer beneath the topsoil remains relatively dry,
       the downward movement of moisture should be minimal.  The capillary barrier
       test cover consists (from top to bottom) of 300 mm of topsoil, an upper lateral
       drainage layer comprised of 80 mm of sand underlain by 220 mm of clean pea
       gravel (the sand serves as a filter that prevents the overlying topsoil from
       migrating downward into the gravel), a barrier layer consisting of 450 mm of
       compacted soil, and a lower drainage layer comprised of 300 mm of sand.  The
       "barrier layer" was compacted dry of optimum and was not intended to have a
       hydraulic conductivity comparable to a traditional CCL.
  5.   Alternative Test Cover 3 is referred to as the anisotropic barrier and attempts to
       limit downward movement of water with  a layering of capillary barriers. The
       various layers are enhanced by varying  soil properties and techniques that  lead
       to the anisotropic properties of the cover. The anisotropic barrier consists (from
       top to bottom) of 150 mm of vegetative material (a mixture of 75% topsoil and
       25% pea gravel by weight), 600 mm of native soil to allow for water storage, a
       150-mm-thick interface layer consisting of fine sand that serves as a filter
       between the overlying native soil and underlying gravel, and 150 mm of pea
       gravel.  The fine sand layer was intended to create one capillary break, and the
       gravel was intended to create a second  capillary break.
  6.   Alternative Test Cover 4 is referred to as the evapotranspiration (ET) cover.
       The ET cover consists of a single, 900-mm-thick layer of native soil.  The bottom
       750 mm of soil was placed in lifts and compacted, while the top 150 mm was not
       compacted.  The cover material was seeded with native species that contained
       a mix of cool and warm weather plants (primarily native grasses).

Preliminary results have indicated that all six test covers are performing well, although
there are significant differences in percolation rates. Dwyer (personal communication)
provided a summary of the first year of percolation, as indicated in Table 4-5.  Cost data
are also summarized in Table 4-5.  The test cover program will provide valuable insights
into conventional and alternative cover designs as data are developed, analyzed, and
published.
                                      4-65

-------
Table 4-5.  Summary of Preliminary Data from ALCD Project.
Test Cover
RCRA Subtitle D Cover
RCRA Subtitle C Cover
Alternative RCRA Subtitle C Cover
with GCL (GM with 8 Defects)
Capillary Barrier
Anisotropic Barrier
Evapotranspiration Cover
Construction Cost
($/m2)
$51
$158
$90
$93
$75
$74
Percolation (L)
after One Year
6724
46
572
804
63
80
4.3.6 Test Covers in East Wenatchee, Washington
Khire et al. (1997) describe a project at the Greater Wenatchee Regional Landfill in East
Wenatchee,  Washington (a semi-arid region), in which two test covers were constructed
and monitored.  The test covers measured 30 m by 30 and were constructed on a
2.71-1:1 V slope.  The test covers were instrumented to measure runoff and percolation,
as well as to monitor moisture conditions within the various layers of the test covers.

Test Cover 1 was referred to as a "resistive barrier" and was a RCRA Subtitle D type
cap. Test Cover 1 consisted of 150 of topsoil underlain by a 450-mm-thick barrier layer
constructed from low plasticity silty clay that was compacted to achieve a hydraulic
conductivity of 2 x 10~7 cm/s.  The low-permeability barrier layer was intended to provide
resistance to infiltration of water, and thus the use of the term "resistive barrier." Test
Cover 2 was a capillary barrier consisting of 150 mm of vegetated silt  topsoil, underlain
by a 750-mm-thick layer of medium,  uniformly graded sand that served as the capillary
break layer.

Performance of the test covers has been documented for a 3-year period (Khire et al.,
1997).  For the first three years, Test Cover 1  allowed percolation of a total of 33 mm of
water (equal to 5.1% of precipitation) through the cover, while Test Cover 2 allowed  only
5 mm (equal to 0.8% of precipitation) to percolate through the cover.  Significant
percolation through the capillary barrier occurred only during the winter of 1993, when
record snow fall occurred.  If the surface layer of the capillary barrier cover had been
increased, it is anticipated that percolation through the capillary barrier would have been
nearly zero.  In the resistive barrier cap, percolation occurred only when the wetting
front reached the base of the low-permeability barrier layer. Percolation increased
significantly in 1995. The primary reason for this increase appeared to be preferential
flow through vertical cracks in the barrier layer, which apparently formed from
desiccation during the previous summer.  Animal burrows, found during field
reconnaissance in the spring of 1995, may have also contributed to increase in
percolation.
                                      4-66

-------
The data indicating an increase in percolation through the low-permeability clay layer is
consistent with observations at Omega Hills, Kettleman City, and Hamburg.

4.3.7  Test Covers at Los Alamos National Laboratory
Nyhan et al. (1997) describe the performance of four test covers constructed at Los
Alamos  National Laboratory for the Protective Barrier Landfill Cover Demonstration.
The four test plots were each constructed on slopes of 5, 10, 15, and 25%, making a
total of 16 test plots. None of the plots was vegetated, apparently to simulate extreme
conditions in which plants provided no evapotranspiration.  Precipitation, runoff,
drainage, and percolation were measured for each plot.  The moisture content of the
soils was also monitored.  Performance for the first four years is documented by Nyhan
etal. (1997).

The four test plots contained the following cross sections:

   1.   Test Cover 1 was termed the "conventional design" for Los Alamos, and
       consisted of 150 mm of loam topsoil underlain by 760  mm of crushed Los
       Alamos tuff (an angular, silty sand), underlain by 300 mm of gravel.
   2.   Test Cover 2 was termed the "EPA design" and consisted (from top to bottom)
       of 610 mm of loam topsoil, a GT separator/filter,  300 mm of sand drainage
       material, and 610 mm of low-permeability clay-sand material. The GM
       component that usually overlies compacted clay in "EPA designs" was
       intentionally omitted because it was thought when the design was conceived in
       the late 1980s  that the GM would not have a  sufficiently long service life for
       radioactive waste disposal units.
   3.   Test Cover 3 was termed the "loam capillary  barrier design" and consisted of
       610 mm of loam topsoil  underlain by 760 mm of fine sand, which served as the
       capillary break.
   4.   Test Cover 4 was termed the "clay loam capillary barrier design" and consisted
       of 610 mm of clay loam topsoil underlain by 760  mm of fine sand.

Performance data showed that 86% to 91 %  of all precipitation that fell on the covers
was evaporated from the unvegetated test covers, which was  not unexpected in the
semi-arid climate of Los Alamos. Of the four test covers, the EPA design provided the
least amount of percolation through  the test plots (zero percolation on all four test plots
employing the EPA design). The bentonitic clay mixed in with sand to form  the barrier
layer apparently helped with the water balance at this semi-arid site.  Test Cover 1
(conventional design for Los Alamos) allowed the greatest amount of seepage, varying
from 174 mm  of percolation for the 5% slope to 31 mm for the 25% slope over a 4.5
year period.  Test Cover 3 (loam capillary barrier design) allowed 76 mm of percolation
for the 5% slope, 36 mm of percolation for the 10% slope, and no percolation for the
15% and 25% slopes over the same 4.5 year period.  Test Cover 4 (clay loam capillary
                                      4-67

-------
barrier design) allowed 48 mm of percolation for the 5% slope, and no percolation for
test plots on steeper slopes for the 4.5-year observation period.

The Los Alamos test plots appear to be the only documented cases in which a
compacted clay placed in a test cover without a protective GM worked well over a
period of several years of observation.

4.3.8  Other Studies
Other papers have been  published on the performance of CCLs in final covers, but
none is as comprehensive as the studies discussed in the preceding sections.
Questions have been raised about the long-term survivability of CCLs, even if the CCL
is covered by a GM.  Suter et al. (1993) discuss the factors that might cause long-term
degradation of CCLs. The primary mechanisms of concern are desiccation, freeze-
thaw,  thermally induced moisture movement leading to desiccation,  root penetration,
subsidence, and animal intrusion.

It appears that the best way to document the field performance of CCLs in landfill final
cover systems is with the use of lysimeters installed at the base of the cover system.
Lysimeters consist of a barrier (typically GM) overlain by a drainage material (typically
sand or gravel, but possibly GN or other geosynthetic drainage material), and drained
by gravity at the low point. Few, if any, such systems have been installed, except in test
plots.  Until performance data are  collected over a period of many years on actual
covers, the long-term performance of CCLs will remain the subject of speculation.
However, the admittedly  sparse data that are available points to the likelihood (if not
certainty) of desiccation and subsequent flow rates well in excess of those associated
with a CCL having a hydraulic conductivity of 1 x 10~7 cm/s or less for cover systems
employing a layer of topsoil overlying a CCL (with no GM separating the CCL and
topsoil).   Although the monitoring of leachate production rates can be very useful in
indicating whether or not the cover is working reasonably well, a careful analysis of
actual flow rates through the cover system may be difficult with such a global
measurement as leachate production rate. Scientists and engineers are encouraged to
collect percolation data with lysimeters whenever possible.

4.4 Summary and Conclusions
The objective of this  component of the study was to document the field performance of
CCLs, and particularly to address  the question of whether CCLs are meeting the
objective of having a hydraulic conductivity of 1 x  10~7 cm/s or less.  Field performance
data on CCLs employed  as liners for actual landfills are very limited (only 8 such cases
are documented),  and even fewer data are available on CCLs in landfill cover systems.
Therefore, the  approach  taken was: (1) document large-scale hydraulic conductivity
measurements obtained  from test pads (with the 8 cases of documented field
performance), and (2) document information available on test covers used to evaluate
CCL performance in landfill cover  systems.
                                      4-68

-------
A database consisting of 89 CCLs (81 test pads plus 8 actual bottom liners) constructed
from natural soils was assembled and analyzed for large-scale field hydraulic
conductivity. The cases covered a broad range of soil types, construction methods, and
regions. All CCLs were constructed using construction and CQA procedures that
appear to be consistent with the current state of practice, and only those CCLs that
were constructed for the explicit purpose of achieving a field hydraulic conductivity
(kpieid) < 1 x 10~7 cm/s were included in the database. Conclusions may be summarized
as follows: (1) 25% of the 89 CCLs failed to achieve the desired  large-scale hydraulic
conductivity of 1  x 10~7 cm/s or less, confirming the difficulty that is often encountered in
achieving the required low hydraulic conductivity and indicating that achieving this level
of impermeability requires careful planning, use of appropriate materials, specification of
suitable compaction requirements, and thorough CQA; (2) a few of the CCLs failed to
meet the desired hydraulic conductivity because the soil materials turned out not to be
suitable; (3) for soils that were found to be unsuitable, a comprehensive laboratory
testing program had not preceded construction of the test pad - thorough laboratory
testing is recommended for all CCLs prior to construction to verify the suitability of the
soil and the proposed compaction specification; (4) the single most common problem in
achieving the desired low level of hydraulic conductivity was failure to compact the soil
in the zone of moisture and dry density that will yield low hydraulic conductivity; (5) the
most significant control parameter was not found to be water content or density, but
rather a parameter denoted "P0", which represents the percentage of field-measured
water content-density points that lie on or above the line of optimums - when P0 was
high (80% to 100%) nearly all the CCLs achieved the desired field hydraulic
conductivity, but when P0  was low (0 to 40%), fewer than half the CCLs achieved the
desired field hydraulic conductivity; (6) practically  no correlation was found between
hydraulic conductivity and frequently measured  soil characterization parameters, such
as liquid limit, plasticity index, percentage of clay,  percentage of fine material, indicating
that natural soil CCLs can be constructed with a relatively broad range of soil materials;
(7) good agreement was obtained between kFieid and kLab on small samples for well
constructed  liners with kFieid < 1 x 10~7 cm/s, but poor agreement was found for poorly
constructed  liners with kFieid > 1 x 10"7 cm/s (with laboratory measurements often
yielding significantly lower hydraulic conductivities); and (8) hydraulic conductivity
decreased with increasing thickness of CCLs, up to a thickness of about 1 m, at which
point all CCLs in the database achieved kFieid < 1 x 10~7 cm/s.

A database on soil-bentonite CCLs was also assembled, but only contained  12 field-
measured hydraulic conductivities on test pads. Relatively little information could be
gleaned from the database for soil-bentonite liners.  The desired hydraulic conductivity
of < 1 x 10~7 cm/s was achieved in all 12 cases.  However, all the CCLs in the database
contained a relatively large amount of bentonite (more than 6%).  The data suggest that
there is justification for focusing attention on a high percent compaction for soil-
                                      4-69

-------
bentonite liners rather than a high water content. More data are needed to be able to
draw more definitive conclusions about soil-bentonite liners.

Finally, data were assembled from the literature on performance of CCLs in landfill final
cover systems. Most of the field data indicate that CCLs tend to desiccate over time,
and if not covered with a GM overlain by soil, are likely to undergo significant increases
in hydraulic conductivity within five years after construction as a result of desiccation
cracking. If the CCL is protected from desiccation by a GM and covering soil,  it appears
that water percolation through the composite barrier will be extremely small, and that
the CCL probably be protected from desiccation for at least several years, if not longer.
Of the data analyzed from final cover systems, only the Omega Hills and  Hamburg test
sites were situated on an actual landfill cover, where differential settlement was a
possibility.  MSW landfills are known to undergo significant settlement, which can
produce stress-induced cracking that increases hydraulic conductivity. Although the
discussion herein focused primarily on the desiccation issue, this was because the test
covers themselves were impacted far more by desiccation than by settlement, due to
the nature of the  test arrangements. Settlement-induced cracking may be a far more
significant effect than indicated by this collection of information from performance of
cover system test sections.

4.5 References
Benson, C.H. and Boutwell, G. (1992). "Compaction Control and Scale-Dependent
   Hydraulic Conductivity of Clay Liners," Proceedings of the 15th Annual Madison
   Waste Conference, University of Wisconsin, Madison, Wisconsin, 62-83.
Benson, C.H., Zhai, H., and Rashad, S.M. (1992), "Assessment of Construction Quality
   Control Measurements and Sampling Frequencies for Compacted Soil Liners,"
   Environmental Geotechnics Report No. 92-6, Univ. of Wisconsin, Dept. of Civil and
   Environmental Engineering Madison, Wisconsin.
Benson, C.H., Hardianto, F.S., and Motan, E.S. (1994), "Representative Specimen Size
   for Hydraulic Conductivity Assessment of Compacted Soil Liners," Hydraulic
   Conductivity and Waste  Contaminant Transport in Soils, ASTM STP 1142, D.E.
   Daniel and S.J. Trautwein (Eds.), American Society for Testing and Materials,
   Philadelphia, 3-29.
Benson, C.H., and Daniel, D.E. (1994), "Minimum Thickness of Compacted Soil Liners:
   II. Analysis and Case Histories." Journal  of Geotechnical Engineering, 120 (1): 153-
   172.
Corser, P., and Cranston, P. (1991), "Observations on Long-Term Performance of
   Composite Clay Liners and Covers," Proceedings, Geosynthetic Design and
   Performance, Vancouver Geotechnical Society, Vancouver, BC.
Corser, P., Pellicer, J., and Cranston, M. (1992), "Observation on Long-Term
   Performance of Composite Clay Liners and Covers," Geotechnical Fabrics Report,
   November, pp. 6-16.
                                      4-70

-------
Daniel, D.E. (1989), "In Situ Hydraulic Conductivity Tests for Compacted Clays," Journal
  of Geotechnical Engineering, 115(9): 1205-1227.
Daniel, D.E., and Benson, C.H. (1990), "Water Content-Density Criteria for Compacted
  Soil Liners," Journal of Geotechnical Engineering, 116(12): 1811-1830.
Daniel, D.E., and Koerner, R.M. (1995), Waste Containment Systems: Guidance for
  Construction, Quality Assurance, and Quality Control of Liner and Cover Systems,
  ASCE Press, New York, 354 p.
Dwyer, S.F. (1997), "Large-Scale Field Study of Landfill Covers at Sandia National
  Laboratories," Proceedings, Landfill Capping in the Semi-Arid West: Problems,
  Perspectives, and Solutions, Environmental Science and Research Foundation,
  Idaho, Falls, ID, ESRF-019, 87-108.
Gordon, M.E., Huebner, P.M., and Mitchell, G.R.  (1990), "Regulation, Construction and
  Performance of Clay Lined Landfills in Wisconsin," Waste Containment Systems, R.
  Bonaparte (Ed.), American Society of Civil Engineers, New York, 14-27.
Khire, M.V., Benson, C.H., and Bosscher, P.J. (1997), "Water Balance of Two Earthen
  Landfill Caps in a Semi-Arid Climate," Proceedings, International Containment
  Technology Conference, St. Petersburg, Florida, 252-261.
Maine Bureau of Remediation and Waste Management (1997), An Assessment of
  Landfill Cover System Barrier Layer Hydraulic Performance, Augusta, Maine.
Melchior, S., Berger, K., Vielhaver, B., and Miehlich, G. (1994), "Multilayered Landfill
  Covers: Field Data on  the Water Balance and  Liner Performance," In-Situ
  Remediation: Scientific Basis for Current and Future Technologies, G.W. Gee and
  N.R. Wing (Eds.), Battelle Press, Columbus, Ohio, Part 1, pp. 4111-425.
Melchior, S. (1997), "In-Situ Studies of the Performance of Landfill Caps," Proceedings,
  International Containment Technology Conference, St.  Petersburg, Florida, 365-373.
Montgomery, R.J., and Parsons, L.J. (1989), "The Omega Hills Final Cover Test Plot
  Study: Three-Year Data Summary," Presented at the Annual Meeting of the National
  Solid Waste Management Association, Washington, DC.
Nyhan, J.W., Schofield, T.G., and Salazar, J.A. (1997), "A Water Balance Study of Four
  Landfill Cover Designs Varying in Slope for Semiarid Regions," Proceedings,
  International Containment Technology Conference, St.  Petersburg, Florida, 262-269.
Reades, D.W., Lahti, L.R., Quigley, R.M., and Bacopoulos, A. (1990), "Detailed Case
  History of Clay Liner Performance,"  Waste Containment Systems:  Construction,
  Regulation, and Performance, R. Bonaparte (Ed.), American Society of Civil
  Engineers, New York,  156-174.
Sai, J.O., and Anderson,  D.C. (1990), "Field Hydraulic Conductivity Tests for
  Compacted Soil Liners," Geotechnical Testing Journal,  13(3): 215-225.
Suter, G.W., Luxmoore, R.J., and Smith, E.D. (1993), "Compacted Soil Barriers at
  Abandoned Landfill Sites Are Likely to Fail in the Long Term," Journal of
  Environmental Quality, Vol. 22, pp. 217-226.
Trautwein, S.J., and Boutwell, G.P. (1994),  "In-Situ Hydraulic Conductivity Tests for
  Compacted Soil Liners and Caps," Hydraulic Conductivity and Waste Contaminant
                                     4-71

-------
Transport in Soils, ASTM STP 1142, D.E. Daniel and S.J. Trautwein (Eds.), American
Society for Testing and Materials, Philadelphia, 184-223.
                                   4-72

-------
                                  Chapter 5
              Detailed Summary of Field Performance Tasks

5.1  Introduction

5.1.1  Scope of Work
This portion of the project involved four tasks designed to evaluate the field
performance of liner systems and final cover systems (referred to as cover systems in
this chapter) for modern landfills in the U.S.  The term "modern landfill" refers to a
landfill designed with components substantially meeting current federal regulations and
constructed and operated to the U.S. state of practice from the mid-1980's forward.  The
four tasks are:

   •    evaluation of available published information on the field  performance of modern
       landfills;
   •    collection and analysis of liquids management data for double-lined landfills;
   •    evaluation of problems that have occurred in waste containment systems (i.e.,
       liner systems and cover systems) for waste management facilities; and
   •    assessment of the adequacy of the EPA HELP computer model as a tool for
       LCRS design.

The purpose of performing these tasks is to  develop an improved understanding of the
actual field performance of modern landfill liner systems and cover systems  and, to the
extent possible, provide data that allow answers to be developed for the following
questions:

   1.   What conclusions can be drawn from available LCRS and LDS data  regarding
       leakage rates through primary liners  and hydraulic efficiencies of liners?
   2.   How much leachate is generated in modern landfills, both during  active
       operations and after closure, and what is the effect of site location (climatic
       region) and waste type on leachate generation rates?
   3.   What is the chemistry of modern  landfill leachate and what  is the effect of site
       location, waste type,  and operation conditions on leachate chemistry?
   4.   What is the effect of the federal solid waste regulations of the 1980's and early
       1990's, which limit the disposal of certain types of wastes to HW facilities and
       prohibit the disposal of certain types  of waste in any facility, on landfill leachate
       chemistry?
   5.   How do leachate generation  rates estimated using the EPA HELP computer
       model compare to actual leachate generation rates at modern operating
       facilities?
   6.   Do the HELP model simulations predict the same effects of site location and
       waste type on leachate generation rates as observed from the actual data?
   7.   What is the nature, frequency, and significance of identified problems in liner
       systems and cover systems for modern waste management facilities?
   8.   How can the identified problems be prevented in the future?

                                      5-1

-------
   9.   What overall conclusions can be drawn regarding the likely long-term
       performance of landfills?

Complete results from the first three tasks were incorporated into two appendices to this
report. These appendices are:

   •   "Appendix E:  Evaluation of Liquids Management Data for Double-Lined
       Landfills"; and
   •   "Appendix F: Waste Containment System Problems and Lessons Learned".

Summaries of Appendices E and F are presented below in Sections 5.2 and 5.3,
respectively, of this report.  Section 5.4 presents the results of the fourth task, an
assessment of the appropriateness of the HELP model as a design tool.  The HELP
model is evaluated by comparing LCRS flow rate data for six landfill cells to leachate
generation rates predicted for these cells from HELP model simulations with typical
input parameters.  Sections 5.2 through 5.4 also present the findings of the project with
respect to the questions listed  above.  References are presented in Section 5.5.

5.1.2  Terminology
Waste containment systems for landfills consist of liner systems that underlay the
wastes placed in them and cover systems constructed over the wastes (Figure 5-1). A
liner system consists of a combination of one or more drainage layers and low-
permeability barriers (liners). The functions of liners and drainage layers are
complementary. Liners impede leachate percolation and gas migration out of a landfill
and improve the collection capability of overlying drainage layers.  Drainage layers
collect and convey liquids on underlying liners to controlled collection points (sumps)
and limit the buildup of hydraulic head on the liners. Most liner systems installed
beneath modern landfills are classified as single-composite  liner systems or double-liner
systems and include the components illustrated in  Figure 5-1.  A single-composite liner
system consists of a composite liner overlain by an LCRS drainage layer.  A double-
liner system consists of a primary liner and a secondary liner with an LDS drainage
layer between the two liners and an LCRS drainage layer above the primary liner. The
LCRS and LDS may also contain networks of perforated pipes, sumps, pumps,
flowmeters, and other flow conveyance and monitoring components.  A liner system
may also include a protection layer over the LCRS drainage layer to further isolate the
liner from the environment (e.g., freezing temperature, stresses from equipment).

Once  an area of a landfill is filled to final grade, a cover system is constructed over the
area to contain the waste, minimize the infiltration of water into the waste, and control
the emissions of gases produced by waste decomposition or other mechanisms. A
cover system consists of up to six basic components, from top to bottom: (i) surface
layer;  (ii)  protection layer; (iii) drainage layer; (iv) barrier; (v) gas collection layer; and (vi)
foundation layer.  In some cases, the functions of several adjacent components can be
                                      5-2

-------
provided by one soil layer. For example, a sand gas collection layer may also serve as
a foundation layer. Many modern landfills have a cover system consisting of a soil
surface and protection layer, drainage layer, barrier, and gas collection layer.
         Cover
        System
                      J
                JL
Surface and Protection Layer

    Drainage Layer
                   Gas Collection Layer
                     GM Barrier
  Single-Composite
     Liner System
        (a)
                      Solid Waste
                    LCRS Drainage Layer
                                            Double-Liner
                                              System
                     Composite Liner
                                                            Solid Waste
                                                          LCRS Drainage Layer
                                          LDS Drainage Layer
                                       Composite Secondary Liner
                               (b)
Figure 5-1.  Typical waste containment system components for landfills: (a)
            single-composite liner system and cover system for a closed landfill;
            and (b) double-liner system for an active landfill.
In general, the materials used to construct liners and barriers in modern landfills are
GMs alone and composites consisting of GMs overlying CCLs or GCLs (i.e., GM/CCL or
GM/GCL composites).  Drainage layers and gas collection layers are typically
constructed with sand, gravel, GNs, or GCs.  Protection layers typically consist of soil or
thick GTs. The protection layer over the LCRS drainage layer sometimes consists of
select waste. Surface layers for cover systems are typically constructed with vegetated
topsoil.

Liner systems for modern MSW landfills and nonhazardous MSWcombustor ash (MSW
ash) landfills must, based on state-specific implementation of RCRA Subtitle D
requirements, meet federal  minimum design criteria or performance-based design
requirements (40 CFR 258.40) as described in Section 1.2. Federal minimum design
criteria require a single-composite liner system for new MSW landfills and MSW ash
landfills. While the federal minimum design criteria were adopted by many states, a few
states require that MSW landfills or MSW ash landfills have a double-liner system. For
RCRA HW landfills, federal  regulations (40 CFR 264.301) require a double-liner system
with at least a GM primary liner and a GM/CCL secondary liner, as described in Section
1.2.
                                       5-3

-------
Cover systems for modern lined MSW landfills and MSW ash landfills (40 CFR 258.40)
must meet federal minimum design criteria or performance-based design requirements
(40 CFR 258.60), as described in Section 1.2. The cover system meeting federal
minimum design criteria consists of a soil surface layer over a composite barrier.  Cover
systems for RCRA HW landfills must meet federal performance-based design
requirements (40 CFR 264.310).  There is not a federal minimum design criteria cover
system for HW landfills; however, EPA guidance (EPA, 1989) recommends that the
cover system for these landfills contain a soil surface and protection layer, drainage
layer, and composite barrier, as described in Section 1.2.

There are currently no federal  minimum design requirements for liner systems or cover
systems for ISW landfills.  ISW landfills contain such wastes as papermill sludge, coal
ash, and construction and demolition waste (C&DW).

5.1.3  Data Collection Methodology
The landfill performance data presented in Appendices E and F and summarized in this
chapter were obtained from the technical literature, engineering drawings, project
specifications, as-built records, and operation records, and from discussions with facility
owners, facility operators, design  engineers, and federal and state regulators throughout
the U.S. The data were collected in accordance with a quality assurance plan, which
was reviewed and approved by the EPA. Efforts were made to obtain data from a wide
variety of facilities with different waste types (i.e., MSW, MSW ash, HW,  and  ISW), site
conditions, and waste containment system components. The study focused on landfills,
and only information on landfills is summarized in this chapter. Based on the broad-
based method of data collection for this study, it is believed that the data in this report
are generally representative of landfills nationwide.

5.2 Evaluation of Liquids Management Data for Double-Lined Landfills

5.2.1  Scope of Work
The scope of work for this portion of the project consisted of the collection and analysis
of liquids management data for 187 active or closed double-lined cells at 54 modern
landfills located throughout the U.S. These data are typically  required to be collected
and reported to regulatory agencies as part of the permit conditions for a landfill. The
data were used to evaluate: (i) leakage rates and hydraulic efficiencies of landfill
primary liners; (ii) landfill  leachate generation rates (LCRS flow rates), including how
these flow rates vary with waste type, site location, and presence of cover system; and
(iii) landfill leachate chemistry (LCRS flow chemistry), including how leachate chemistry
varies with waste type, site location, and operation conditions, and whether federal solid
waste regulations promulgated in the 1980's and early 1990's have had an effect on the
quantity of potentially-toxic trace chemicals found in leachate.
                                       5-4

-------
5.2.2 Description of Database
The liquids management data and related data collected for the 187 landfill cells
include: (i) general facility information (including location, average annual rainfall,
subsurface soil types, groundwater separation distance from bottom of landfill); (ii)
general cell information (including cell area, type of waste, height of waste, dates of
construction, operation, and closure); (iii) double-liner system and cover system design
details (including type, thickness, and hydraulic conductivity of each layer); (iv) LCRS
flow rate and chemical constituent data; and (v)  LDS flow rate and chemical constituent
data.  For comparison purposes, the data are sorted according to liner system type,
waste type, and site geographic location (which  is indicative of site climate).  The full
database is presented in Appendix E. The reader is referred to the following figures and
tables in Appendix E for specific information:

   •   double-liner system types:  Table E-1.1 and Figure E-1.1;
   •   geographic regions and site locations: Figure E-1.3;
   •   general facility information: Table E-3.1;
   •   general cell information: Table E-3.2;
   •   double-liner system design details: Table E-3.3;
   •   cover system design details:  Table E-3.4;
   •   LCRS flow rate data: Table E-3.5;
   •   LDS flow rate data: Table E-3.6; and
   •   LCRS and LDS flow chemistry data: Table E-3.7.

The distributions of the landfill facilities and cells in the database by waste type and
geographic region and by primary liner and LDS types are shown in Tables 5-1 (a) and
(b), respectively. From  Table  5-1 (a), most of the landfills in the database are located  in
the northeast (NE).  This is not surprising because: (i) the NE has a relatively dense
population; and (ii) double-liner systems are required for MSW landfills in several states
in the NE. In addition, the majority of the landfills in the database are used for disposal
of MSW. Based on the extent of the database and comparisons of these data with
published data, discussed in Section E-2 of Appendix E, the database appears to
adequately characterize conditions for MSW landfills in the NE and southeast (SE), HW
landfills in the NE and SE, and MSW ash landfills in the NE. The database is quite
sparse for landfills in the west (W), coal ash landfills, and C&DW landfills. Additional
data from these facilities should be collected and evaluated.

From Table 5-1 (b), most of the cells at most of the landfills have either a GM primary
liner (37% of all cells) or GM/CCL or GM/GCL/CCL primary liner (48%).  Fewer cells
(15%) have a GM/GCL  primary liner. About 48% of the cells have a sand or gravel LDS
and 52% have a GN LDS. Based on the distribution of the data, the database appears
to be representative of typical  double-liner system designs in landfills.
                                       5-5

-------
Table 5-1 (a). Distribution of Database by Waste Type and Geographic Region.

Waste Type
MSW
HW
MSW Ash
Coal Ash
C&DW
Geographic Region
Northeast U.S.
24 landfills
71 cells
5 landfills
26 cells
5 landfills
12 cells
1 landfills
1 cell
2 landfills
4 cells
Southeast U.S.
8 landfills
26 cells
5 landfills
31 cells
2 landfills
4 cells
0 landfills
0 landfills
West U.S.
1 landfill
2 cells
3 landfills
10 cells
0 landfills
0 landfills
0 landfills
Table 5-1 (b). Distribution of Database by Primary Liner and LDS Types

Primary Liner Type
GM
GM/GCL Composite
GM/CCL or
GM/GCL/CCL
Composlte
LDS Type
Sand or Gravel
13 landfills
41 cells
3 landfills
19 cells
13 landfills
31 cells
GN
11 landfills
28 cells
4 landfills
9 cells
16 landfills
57 cells
Most of the liquids management data are for open cells; only about 23% of the cells in
the database had received a cover system.

5.2.3  Data Interpretation

5.2.3.1  Landfill Development Stages
In evaluating  LCRS and LDS flow rate and chemical constituent data for this report,
three distinct  landfill development stages were considered: (i) the "initial period of
operation"; (ii) the "active period of operation"; and (iii) the "post-closure period".  These
stages are defined by the waste filling and capping rates of a landfill cell and are
described below.  The initial period of operation occurs during the first few months after
the start of waste disposal in a cell. During this stage, there is not sufficient waste in a
cell to significantly impede the flow of rainfall into the LCRS.  To the extent rainfall
occurs during this stage, it will rapidly find its way into the LCRS. LCRS flow rates
during this stage are  usually controlled by rainfall and can be directly correlated to local
climatic conditions. LCRS flow rates are higher at landfills in wetter climates than at
those in arid climates. During the active period of operation, the cell is progressively
                                       5-6

-------
filled with waste and daily and intermediate layers of cover soil.  As waste placement
continues, more of the rainfall occurring during this stage falls onto the waste and cover
soils rather than directly onto the liner system. As a consequence, the LCRS flow rates
decrease and eventually stabilize.  LCRS flow rates during this stage are generally
dependent on rainfall, waste thickness, waste properties (i.e., initial moisture content,
field capacity, and permeability), and storm-water management practices. During the
post-closure period,  the cell has been closed with a cover system  that further reduces
infiltration  of rainfall into the waste, resulting in further reduction in LCRS flow rates.
LCRS and LDS flows associated with these three development stages are illustrated for
a MSW landfill in Pennsylvania in Figure 5-2.

5.2.3.2 Primary Liner Leakage Rates and Hydraulic Efficiencies
LCRS and LDS flow data were interpreted to assess primary liner leakage rates and/or
apparent efficiencies for the following:

   •   GM primary liners by development stage, LDS type, and use of CQA;
   •   GM/GCL primary liners by development stage and LDS type; and
   •   GM/CCL primary liners by development stage.

The data were first assessed using a methodology presented by Gross et al. (1990) for
using LCRS and LDS flow data to evaluate the performance of primary liners in terms of
primary liner leakage. The basic approach involves the evaluation of LCRS  and LDS
flow rate and chemical constituent data to quantify that portion of LDS flow that is
attributable to primary liner leakage as opposed to other sources.  Other sources of LDS
flow include: (i) water (mostly rainwater) that infiltrates the LDS during construction and
continues  to drain to the  LDS sump after the start of facility operation ("construction
water"); (ii) water that infiltrates the LDS during construction, is held in  the LDS by
capillary tension, and is expelled from the LDS during waste placement as a result of
LDS compression under  the weight of the waste ("compression water"); (iii) water
expelled into the LDS from any CCL and/or GCL components of a composite primary
liner as a result of clay consolidation  under the weight of the waste ("consolidation
water"); and (iv) water that percolates through the secondary liner and  infiltrates the
LDS ("infiltration water"). The sources of LDS flow are illustrated in Figure 5-3.
Evaluation of the potential flow rates  and times of occurrence of each of these potential
sources of flow were made using the calculations procedures contained in Gross et al.
(1990). That portion of LDS flow attributable to leakage through the composite primary
liner of a double-liner system would be leakage into the ground for a single-composite
liner system if the two composite liners have similar characteristics.  Extrapolation of
primary liner performance levels to the secondary liner of a double-liner system enables
inferences to be drawn regarding performance of the entire double-liner system.
                                      5-7

-------
                Initial Period
                of Operation
                            Active Period of Operation
   160
   140
                                       o>

                                       —>
                                            CM
                                            O>
CM
O>
CO
O)
CO
O>
                o>
                c
                ro
                o>
                "5
                -3
LO
O)
LO
O)
                                           DATE
Figure 5-2. LCRS and LDS flow rates over time at a MSW landfill in Pennsylvania.
            (Flow rates are given in liters/hectare/day (Iphd).)
                                         5-8

-------
                           GM
           '. _ .QRQUNPWATER.TABLE_ . _
        Q = TOTAL FLOW
        Q=A+B+C+D

        SOURCES:
        A = PRIMARY LINER LEAKAGE
        B = CONSTRUCTION WATER AND COMPRESSION WATER
        C = CONSOLIDATION WATER
        D = INFILTRATION WATER
    Figure 5-3. Sources of flow from LDSs (from Bonaparte and Gross, 1990).

The relative performances of the different types of primary liners were then evaluated
using the "apparent liner hydraulic efficiency" parameter, Ea, introduced by Bonaparte et
al. (1996) and defined as:
              Ea(%) = (1 - LDS Flow Rate/LCRS Flow Rate) x 100
(Eq.5-1)
The higher the value of Ea, the smaller the flow rate from an LDS compared to the flow
rate from an LCRS. The value of Ea may range from 0 to 100%, with a value of zero
corresponding to an LDS flow rate equal to the LCRS flow rate and a value of 100%
indicating no flow from the LDS.  The parameter Ea is referred to as an "apparent"
hydraulic efficiency because, as described above, flow into the LDS sump of a landfill
may be due to sources other than primary liner leakage (Figure 5-3). The value of Ea is
calculated using total flow into the LDS, regardless of source. If the only source of flow
into the LDS sump is primary liner leakage, then Equation  5-1 provides the "true" liner
hydraulic efficiency, Et. True liner efficiency provides a measure of the effectiveness of
a particular liner in limiting or preventing advective transport across the liner.  For
example, if a liner is estimated to have an Et value of 99%, the rate of leakage through
the primary liner would be assumed to be 1 % of the LCRS flow rate. The true efficiency
of a liner is not constant but rather a function of the hydraulic head in the LCRS and size
of the area over which LCRS flow is occurring (the area is  larger at high flow rates
                                      5-9

-------
compared to low flow rates). The true efficiency of a liner is also a function of design:
identical liners overlain by different LCRSs or placed on different slopes will exhibit
different Et values.  Also, the efficiency of a liner for a given set of hydraulic conditions
could change over time if the physical condition of the liner changes. For example,
time-dependent changes in GMs could result from chemical degradation or stress
cracking under certain conditions. Time-dependent changes in CCLs or GCLs could
result from chemical degradation, consolidation, or other factors.  Notwithstanding all of
these limitations, the hydraulic efficiency concept  has been found useful in
characterizing liner hydraulic performance.

The methodology described above was used to evaluate the hydraulic performance of
GM primary liners and GM/GCL composite primary liners.  Chemical constituent data
were not utilized in  the evaluation of these types of liners because the initial hydraulic
assessment (i.e., comparing LCRS and LDS flow  rates) yielded significant insight into
these liners' true hydraulic efficiencies. However,  the situation was found to be more
complicated for GM/CCL and GM/GCL/CCL composite primary liners due the
generation of consolidation water by these liners not only during the initial period of
operation, but also  during the active and post-closure periods.  The performance
evaluation of these liners included the additional step of comparing the chemistry of
LCRS and LDS liquids to assess whether the liquids had different primary sources (i.e.,
leachate for LCRS  liquids and  CCL pore water for LDS liquids). The concentrations of
five key chemical constituents  (i.e., the inorganic anions sulfate and chloride and the
aromatic hydrocarbons benzene, toluene, and xylene) in the LCRS and LDS flows were
compared in more detail to further assess whether primary liner leakage had contributed
to LDS flows.

It is noted that the presence of chemical constituents in the LDS was evaluated
empirically. Therefore, the concentrations of chemicals collected in the LDS were
directly compared to concentrations of the same chemicals collected in the LCRS. No
fate and transport analysis was performed that accounts for attenuation of the LCRS
chemicals migrating through the primary liner CCL.  However, to overcome the need to
perform such an analysis, the five key chemical constituents were  selected based on
their high solubility  in water, low octanol-water coefficient, high  resistance to
hydrolization, and high resistance to anaerobic biodegradation  in soil.

5.2.3.3 Leachate Generation Rates
LCRS flow rate  data were interpreted in terms of average and peak monthly leachate
generation rates for the following:

   •   MSW landfills by geographic region and development stage;
   •   HW landfills by geographic region and development stage;
   •   ash (i.e., coal ash and  MSW ash) landfills  by geographic region and
       development stage; and
                                      5-10

-------
   •   C&DW landfills by development stage.

The data were also used to evaluate the ratios of average LCRS flow rates to historical
average annual rainfalls by waste type, geographic region, and development stage.

5.2.3.4  Leachate Chemistry
Data on leachate chemical constituents were interpreted in terms of the average
concentrations and detection frequencies (i.e., were the chemicals detected in 50% or
less of the samples or more than 50% of the samples) of 30 representative chemical
parameters. These data were then used to assess the following:

   •   effect of waste type on leachate chemistry; and
   •   effect of federal solid waste regulations of the 1980's and early 1990's on
       leachate chemistry (i.e., has the amount of trace toxic inorganic and synthetic
       organic chemicals in leachate decreased).

The 30 representative chemical parameters consist of water quality indicator
parameters (e.g., pH, specific conductance, total dissolved solids (TDS), etc.), major
inorganic cations and anions (e.g., calcium, chloride, sulfate, etc.), trace metals (e.g.,
arsenic, chromium, lead, etc.), and volatile organic compounds (VOCs) (e.g., benzene,
methylene chloride, trichloroethylene, etc.). The specific trace metals and VOCs were
chosen for this study because these metals and VOCs are sometimes found in
leachates from MSW, HW, and ISW landfills. They were also selected based on
availability of parameters between landfills, frequency of detection, and concentration.

It is recognized that the leachate chemistry database is limited in terms of completeness
and duration of monitoring.  In addition, key MSW and HW leachate constituents, such
as alcohols and ketones, are poorly  represented in the database and, thus, could not be
included in the list of select parameters.  It is important that these additional data be
collected so that our understanding of leachate chemistry can continue to  improve. The
chemical data presented herein are  intended to be representative, not comprehensive.
The data should not be considered complete for purposes of evaluating potential human
health or ecological impacts.

5.2.4  Evaluation Results

5.2.4.1  Primary Liner Leakage Rates and Hydraulic Efficiencies

GM Liners
The performance of 31 of the 69 cells with GM primary liners was assessed. The
remaining 38 cells with GM primary liners were excluded from the assessment primarily
because they do not have continuous LCRS and LDS flow rate data available for an
individual cell from the start of operation and for a significant monitoring period. Flow
rate data are available for the considered 31 cells at 14 landfills with monitoring periods
                                      5-11

-------
of up to 114 months. Twenty-five cells have a HOPE GM primary liner, and the
remaining six cells have a chlorosulfonated polyethylene (CSPE) GM primary liner. A
formal CQA program was used in the construction of 23 of the 25 cells with an HOPE
GM primary liner. To the best of the authors' knowledge, none of the cells received
electrical leak location surveys or ponding tests as part of the CQA program.  The
remaining two cells with an HOPE GM primary liner (i.e., F1 and K1) and the six cells
with a CSPE GM primary liner (i.e., B1, B2, and E1 to E4) were all constructed without
CQA.  A summary of the LCRS and LDS flow rate data for the 31  cells is provided in
Table 5-2, and detailed results are given in Figure E-4.1 and Tables E-4.3 through E-4.5
of Appendix E. The major findings from the data evaluation of cells with GM primary
liners are given below:

   •   LDS flows during the initial period of operation are attributed primarily to
       construction water and primary liner leakage.  LDS flows during the active and
       post-closure periods are attributed primarily to primary liner leakage.
   •   Average monthly LDS flow rates for cells constructed with a formal CQA
       program ranged from about 5 to 440 Iphd during the initial period of operation, 1
       to 360 Iphd  during the active period, and 2 to 60 Iphd during the post-closure
       period. Peak monthly flow rates for these cells were typically below 500 Iphd
       and exceeded 1,000 Iphd in only two of the 23 cells.
   •   Based on an analysis  of the available data, average monthly active-period LDS
       flow rates for cells with HOPE GM primary liners constructed with CQA will often
       be less than 50 Iphd, but occasionally in excess of 200 Iphd. These flows are
       attributable  primarily to liner leakage and, for cells with sand LDSs, possible
       construction water.
   •   The eight cells constructed without a formal CQA program exhibited average
       monthly LDS flow rates about one to two orders of magnitude greater than LDS
       flow rates for cells constructed with CQA. The average flow rates from the eight
       cells ranged from 120 to 2,140 during the initial period of operation, 70 to 1,600
       Iphd during  the active  period, and, for the two cells for which post-closure data
       are  available, 210 to 240 Iphd during the post-closure period. The large
       differences  in LDS flow rates between cells constructed with CQA and cells
       constructed without CQA are partly attributed to the benefits of CQA and partly
       due to differences in the GM materials and construction (i.e., seaming) methods.
       The two cells that had HOPE GM primary liners and no formal CQA had average
       LDS flow rates that are about two to seven times greater than the mean LDS
       flow rate for all cells constructed with a formal CQA program. In contrast, the
       cells with CSPE  GM primary liners and no formal CQA exhibited average LDS
       flow rates that are about one to two orders of magnitude greater than the mean
       LDS flow rate for all cells that had CQA. There are not sufficient data in this
       appendix, however, to accurately separate the effects of CQA and GM type (i.e.,
       HOPE vs. CSPE) and construction methods on leakage rates through GM liners.
   •   Based on an analysis  of the available data, GM liners can be constructed to
       achieve very good hydraulic performance (i.e. Et values greater than 99%).
       However, even when constructed with a CQA program, GM liners sometimes
                                      5-12

-------
        Table 5-2. Summary of LCRS and LDS Flow Rate Data for Landfill Cells with GM Top Liners.
Cell
No.
B1
B2
C1
C2
C3
C4
C5
D1
D3
D4
Cell
Area
(hectare)
3.3
3.5
3.2
3.7
3.6
3.7
2.6
0.4
0.3
0.4
Start of
Waste
Place.
(month-
year)
5-84
5-84
5-90
4-91
8-91
2-92
11-92
10-85
7-87
1-89
End of
Final
Closure
(month-
year)
11-88
11-88
NA(5)
NA
NA
NA
NA
5-86
NA
NA
Initial Period of O
Time
Period
(months)
1-19
1-19
1-9
1-12
1-8
1-4
1-12
1-7
1-12
1-11
LCRS Flow<4)
Avg.
(Iphd)
ND(5)
ND
ND
1,475
3,417
14,828
6,419
ND
20,292
31,281
Peak
(Iphd)
ND
ND
ND
2,585
9,558
41 ,331
12,528
ND
51 ,265
120,527
Deration
LDS Flow
Avg.
(Iphd)
ND
ND
ND
92
63
178
23
32
12
233
Peak
(Iphd)
ND
ND
ND
398
268
265
40
80
56
801
^am
(%)



93.74
98.16
98.80
99.64

99.94
99.25
Active Period of Operation
Time
Period
(months)
20-31
32-43
44-54
20-31
32-43
44-54
10-21
22-33
34-45
46-56
13-24
25-36
37-45
9-20
21-32
33-41
5-16
17-28
29-35
13-26
NA
13-24
25-28
NA
LCRS Flow
Avg.
(Iphd)
2,245
5,223
3,975
2,732
3,740
2,337
789
259
159
103
435
300
161
311
314
268
937
438
407
2,513
NA
13,003
1,010
NA
Peak
(Iphd)
5,754
6,845
7,464
5,393
5,707
3,982
1,419
780
286
200
859
610
464
671
752
987
2,055
622
686
10,440
NA
44,895
2,413
NA
LDS Flow
Avg.
(Iphd)
266
424
892
404
996
665
123
89
27
40
9
22
7
2
33
16
70
51
26
28
NA
7
283
NA
Peak
(Iphd)
499
808
1,426
605
1,690
1,102
304
170
128
227
31
125
14
9
276
103
147
92
29
115
NA
73
341
NA
^am
(%)
88.14
91.87
77.55
85.20
73.36
71.54
84.40
65.52
83.08
61.27
98.03
92.71
95.40
99.49
89.56
94.02
92.52
88.39
93.71
98.88

99.95
71.97

/3\
Post-Closure Period^ '
Time
Period
;months)
55-66
67-78
79-90
91-102
103-114
55-66
67-78
NA
NA
NA
NA
NA
8-19
20-26
27-38
39-50
NA
NA
LCRS Flow
Avg.
(Iphd)
317
703
1,146
1,306
510
493
337
NA
NA
NA
NA
NA
ND
ND
376
715
NA
NA
Peak
(Iphd)
670
1,877
1,956
1,943
718
1,040
654
NA
NA
NA
NA
NA
ND
ND
1,455
1,352
NA
NA
LDS Flow
Avg.
(Iphd)
106
267
279
326
74
154
328
NA
NA
NA
NA
NA
102
1
5
64
NA
NA
Peak
(Iphd)
222
1,134
451
612
97
393
514
NA
NA
NA
NA
NA
886
10
70
156
NA
NA
^am
(%)
66.48
62.02
75.64
75.01
85.41
68.83
2.80





98.58
91.05


en
CO

-------
Table 5-2.  Summary of LCRS and LDS Flow Rate Data for Landfill Cells with GM Top Liners (cont).
Cell
No.



E1


E2


E3
E4
F1

G1





G2


M(6)







|2<6)






Cell
Area


(hectare)
2.4


2.4


1.2
1.2
1.8

3.0





1.6


3.2/2.7<7)







4.2/2.3<7)






Start of
Waste
Place.
(month-
year)
3-88


10-87


5-90
7-90
7-92

6-89





6-89


8-87







10-87






End of
Final
Closure
(month-
year)
NA


NA


NA
NA
NA

NA





NA


10-94







10-94






Initial Period of O
Time
Period

(months)
1-7


1-12


1-12
1-12
1-12

1-12





1-12


1-5
6-8






1-7






LCRS Flow<4)
Avg.

(Iphd)
ND


ND


9,425
20,148
14,472

22,371





22,371


ND
ND






6,627






Peak

(Iphd)
ND


ND


25,394
55,785
45,010

46,120





46,120


ND
ND






13,959






Deration
LDS Flow
Avg.

(Iphd)
2,144


483


1,595
996
124

ND





197


234
ND






31






Peak

(Iphd)
5,026


3,518


1,951
2,362
479

ND





645


508
ND






77






Earn
(%)








83.08
95.06
99.14







99.12










99.53






Active Period of Operation
Time
Period

(months)
8-19
20-31
32-40
13-24
25-36
37-45
13-14
NA
13-24
25-30
13-24
25-36
37-42
43-51
52-63
64-67
13-24
25-36
37-42
9-15
16-32
33-44
45-48
49-54
55-66
67-78
79-84
8-24
25-36
37-40
41-46
47-58
59-70
71-76
LCRS Flow
Avg.

(Iphd)
8,432
11,521
6,525
5,821
4,547
4,434
6,062
NA
9,000
7,826
12,893
3,438
8,356
ND
ND
ND
12,893
3,438
8,356
16,224
ND
7,167
231
ND
624
541
904
ND
1,030
427
ND
624
541
904
Peak

(Iphd)
19,614
36,164
13,075
10,445
11,014
6,830
9,038
NA
25,450
10,932
23,485
1 1 ,652
10,303
ND
ND
ND
23,485
1 1 ,652
10,303
48,932
ND
22,020
332
ND
1,580
752
1,827
ND
3,241
1,054
ND
1,580
752
1,827
LDS Flow
Avg.

(Iphd)
1,436
1,051
743
802
685
596
1,603
NA
66
67
ND
156
101
121
74
49
37
35
60
5
ND
10
4
ND
2
13
79
ND
5
6
ND
8
8
5
Peak

(Iphd)
3,069
1,915
1,015
2,447
1,404
999
1,758
NA
83
77
ND
238
116
384
139
64
65
42
100
18
ND
44
10
ND
5
42
157
ND
35
11
ND
37
23
6
Earn
(%)


82.97
90.88
88.61
86.22
84.93
86.56
73.56

99.27
99.15

95.47
98.79



99.71
98.98
99.28
99.97

99.86
98.49

99.68
97.60
91.26

99.52
98.67

98.67
98.54
99.49
Post-Closure Period'3'
Time
Period

'months;
NA


NA


NA
NA
NA

NA





NA


85-93







77-85






LCRS Flow
Avg.

(Iphd)
NA


NA


NA
NA
NA

NA





NA


800







800






Peak

(Iphd)
NA


NA


NA
NA
NA

NA





NA


1,794







1,794






LDS Flow
Avg.

(Iphd)
NA


NA


NA
NA
NA

NA





NA


62







2






Peak

(Iphd)
NA


NA


NA
NA
NA

NA





NA


119







4






Earn
(%)





















92.25







99.71







-------
        Table 5-2. Summary of LCRS and LDS Flow Rate Data for Landfill Cells with GM Top Liners (cont).
Cell
No.



I3(6)






K1




N2



01(8)




02(8)



S1



S2


Cell
Area


(hectare)
3.4/1 .8(7)






2.7




6.3



4.2




4.9



2.0



1.6


Start of
Waste
Place.
(month-
year)
4-88






12-89




1-92



9-88




3-89



9-90



8-90


End of
Final
Closure
(month-
year)
10-94






NA




NA



NA




NA



NA



NA


Initial Period of O
Time
Period

(months)
1-7






1-12




1-12



1-6




1-12



1-10



1-9


LCRS Flow<4)
Avg.

(Iphd)
1 1 ,559






17,808




ND



ND




4,407



2,226



2,185


Peak

(Iphd)
21,081






24,832




ND



ND




9,826



5,081



4,650


Deration
LDS Flow
Avg.

(Iphd)
37






122




ND



293




6



12



5


Peak

(Iphd)
87






163




ND



620




24



39



24


^am
(%)


99.68






99.31













99.86



99.45



99.78


Active Period of Operation
Time
Period

(months)
8-24
25-36
37-40
41-46
47-58
59-70
71-76
13-24
25-36
37-48
49-60
61-66
13-19
20-31
32-34
35-39
7-18
19-30
31-42
43-54
55-64
13-24
25-36
37-48
49-59
11-22
23-28
29-40
41-45
10-17
18-33
34-46
LCRS Flow
Avg.

(Iphd)
ND
1 1 ,684
2,464
ND
624
541
904
12,929
10,879
6,155
5,952
9,494
4,547
2,561
6,399
2,741
4,407
4,023
7,089
6,201
8,661
4,023
7,089
6,201
8,661
653
ND
1,571
1,086
654
ND
1,255
Peak

(Iphd)
ND
26,339
4,666
ND
1,580
752
1,827
27,663
17,683
11,331
8,024
12,245
5,741
3,460
7,274
3,170
9,826
13,231
16,467
12,561
15,327
13,231
16,467
12,561
15,327
1,220
ND
4,074
2,067
1,135
ND
3,638
LDS Flow
Avg.

(Iphd)
ND
7
5
ND
4
13
17
88
76
514
349
282
113
203
786
201
0
3
0
1
3
2
1
3
1
38
ND
8
4
5
ND
5
Peak

(Iphd)
ND
23
8
ND
17
55
53
180
104
892
495
378
468
669
1,058
406
3
7
5
6
9
5
4
11
5
68
ND
26
7
24
ND
8
^am
(%)



99.94
99.80

99.39
97.64
98.14
99.32
99.30
91.64
94.14
97.03
97.52
92.08
87.72
92.65
99.99
99.93
99.99
99.98
99.97
99.95
99.98
99.96
99.99
94.18

99.51
99.64
99.20

99.63
/3\
Post-Closure Period^ '
Time
Period

;months;
77-85






NA




NA



NA




NA



NA



NA


LCRS Flow
Avg.

(Iphd)
800






NA




NA



NA




NA



NA



NA


Peak

(Iphd)
1,794






NA




NA



NA




NA



NA



NA


LDS Flow
Avg.

(Iphd)
3






NA




NA



NA




NA



NA



NA


Peak

(Iphd)
12






NA




NA



NA




NA



NA



NA


^am
(%)


99.57































en
01

-------
Table 5-2. Summary of LCRS and LDS Flow Rate Data for Landfill Cells with GM Top Liners (cont).
Cell
No.
V1(8)
V2(8)
W1
W2
X1
Cell
Area
(hectare)
4.2
3.9
15.4
15.4
3.0
Start of
Waste
Place.
(month-
year)
1-90
1-90
5-92
5-92
8-92
End of
Final
Closure
(month-
year)
NA
NA
NA
NA
NA
Initial Period of O
Time
Period
(months)
1-10
1-10
1-8
9-12
1-8
1
2-7
LCRS Flow(4)
Avg.
(Iphd)
13,622
13,622
ND
7,492
ND
1 1 1 ,031
32,469
Peak
(Iphd)
49,828
49,828
ND
8,799
ND
111,031
104,645
Deration
LDS Flow
Avg.
(Iphd)
117
135
ND
439
ND
364
4
Peak
(Iphd)
153
256
ND
765
ND
364
25
^am
(%)
99.14
99.01
94.14

99.67
99.99
Active Period of Operation
Time
Period
(months)
NA
NA
13-24
25-35
9-20
21-32
33-35
8-19
20-33
LCRS Flow
Avg.
(Iphd)
NA
NA
2,693
943
4,288
4,813
719
5,926
2,188
Peak
(Iphd)
NA
NA
6,365
1,572
9,389
10,524
2,141
14,315
5,376
LDS Flow
Avg.
(Iphd)
NA
NA
34
19
594
204
32
5
0
Peak
(Iphd)
NA
NA
109
44
1,826
1,217
52
45
2
^am
(%)


98.72
97.98
86.15
95.76
95.50
99.92
99.99
/3\
Post-Closure Period^ '
Time
Period
'months)
NA
NA
NA
NA
NA
LCRS Flow
Avg.
(Iphd)
NA
NA
NA
NA
NA
Peak
(Iphd)
NA
NA
NA
NA
NA
LDS Flow
Avg.
(Iphd)
NA
NA
NA
NA
NA
Peak
(Iphd)
NA
NA
NA
NA
NA
^am
(%)





Notes:
(1) "Initial Period of Operation" represents period after waste placement has started and not more than a few lifts of waste and daily cover have been
    placed in the cell (i.e., no intermediate cover).
(2) "Active Period of Operation" represents period when waste thickness in cell is significant and/or an effective intermediate cover is placed on the waste.
(3) "Post-Closure Period" represents period after final cover system has been placed on the entire cell.
(4) Flow rates are given in liter/hectare/day.
(5) NA =  not applicable; ND = not determined.
(6) LCRS for Cells 11, 12, and I3are combined after February 1992. The measure average flow rates are assumed to represent flow rates for the three cells.
(7) Values given represent LCRS and LDS areas, respectively.
(8) LCRS flows are combined for Cells O1 and O2 and for Cells V1 and V2. Measured average flow rates are assumed to represent flows for the two cells at each landfill.

-------
       will not achieve this performance level and lower Et values, in the range of about
       90 to 99%, will occur. This relatively broad range of Et values is a consequence
       of the potential for even appropriately installed GMs to have an occasional small
       hole, typically due to an imperfect seam, but also potentially due to a
       manufacturing or construction-induced defect not identified by the CQA
       program. Leakage can occur, relatively unimpeded, through a GM hole if the
       GM is not underlain by a low-permeability material such as a CCL or GCL. If a
       hole occurs at a critical location where a hydraulic head exists, such as in a
       landfill sump, the leakage rate through the hole can be significant. In contrast,
       the GCL or CCL component of a  composite liner can impede flow through a GM
       hole, even if it occurs at a critical location.
   •   The conclusion to be drawn from the above data evaluation is that single liner
       systems with GM liners (installed on top of a relatively permeable subgrade)
       should not be used in applications where Et values as low as 90% would be
       unacceptable, even if a thorough CQA program is employed. In these cases,
       single-composite liner systems or double-liner systems should be utilized.  An
       exception to this conclusion may  be made for certain facilities, such as surface
       impoundments or small, shallow landfill cells, with GM primary liners that can be
       field tested over the GM sheet and  seams using electrical leak location  surveys,
       ponding tests, or other methods.  For these facilities, higher efficiencies (i.e.,
       greater than 99%) may be achieved with GM liners by identifying and repairing
       the GM holes during construction and, especially for surface impoundments,
       during operation.  In  all cases, GM  liners should be manufactured and installed
       using formal quality assurance programs.

GM/GCL Composite Liners
The performance of all 28 cells with GM/GCL composite primary liners was assessed.
Flow rate data are available  for the 28 cells at seven landfills with monitoring periods of
up to 83 months. All of these cells were constructed with formal CQA programs.  A
summary of the LCRS and LDS flow rate data for the cells is provided in Table  5-3 and
detailed results are given in Table E-4.10 of Appendix E. The major findings from  the
data evaluation of cells with  GM/GCL primary liners (excluding cell 14, which may have
surface-water infiltration into the LDS at the anchor trench) are given below:

   •   LDS flows during the initial  period of operation are attributed primarily to
       construction water. LDS flows during the active and post-closure periods are
       attributed primarily to primary liner leakage and compression water.
   •   Average monthly LDS flow  rates ranged from about 0 to 290 Iphd during the
       initial period of operation, 0 to 11  Iphd during the active period, and 0 to 2 Iphd
       (with many values reported as zero) during the post-closure period.  Peak
       monthly flow rates were typically below 200 Iphd and exceeded 500 Iphd in only
       four of the 28 cells.
   •   Based on the above  data, average  monthly active-period LDS flow rates
       attributable to leakage through  GM/GCL primary liners constructed with CQA will
       often be less than 2 Iphd, but occasionally in excess of 10 Iphd.
                                      5-17

-------
         Table 5-3.  Summary of LCRS and LDS Flow Rate Data for Landfill  Cells with GM/GCL Composite Primary Liners.
Cell
No.

C6
14
15
AW1
AW2
AX1
AX2
AX3
AX4
AX5
AX6
AX7
AX8
AX9
AX10
AX11
AX12
AX13
AX14
AX15
AX16
AY1
AY2
AYS
AZ1
BB1
BB2
BBS
Cell
Area
(ha)
3.6
4.7
4.7
2.4
2.4
2.0
2.0
1.7
1.7
2.8
3.9
2.6
3.8
3.3
3.9
3.0
4.0
3.0
2.8
2.8
4.5
1.3
1.0
1.0
3.8
4.0
2.4
2.8
Waste
Place.
Start
Date
Aug-93
May-92
Jul-92
May-93
Aug-93
Jul-88
Jul-88
Sep-88
Sep-88
Oct-88
Dec-88
Feb-89
Jul-89
Dec-89
Jul-90
Feb-90
Oct-90
Jan-91
Apr-91
May-92
Jan-93
Oct-94
Aug-94
Aug-94
Dec-92
Feb-91
Jan-93
Jan-93
Final
Closure
Date

NA
Jul-94
May-94
NA
NA
Feb-91
Feb-91
Apr-93
Apr-93
NA(5)
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Initial Period of Operation'1'
Time
Period(6)
(months)
1-10
1-12
1-12
1-12
1-10
1-2
1-5
1-5
1-12
1-11
1-9
1-10
1-14
1-9
1-7
1-16
1-12
1-7
1-11
1-12
1-10
1-9
1-11
1-11
2-12
1-6
1-11
1-11
LCRS Flow(4)
Avg.
(Iphd)
3,273
4,494
3,938
6,358
3,553
16,718
15,521
3,366
2,534
1,384
3,759
5,376
4,881
1,047
2,786
4,675
3,494
6,683
2,777
5,573
8,601
6,803
10,964
12,198
4,093
10,378
ND(5)
ND
Peak
(Iphd)
12,155
17,251
7,985
20,570
7,480
19,738
58,671
7,985
12,688
3,394
7,171
12,155
21,038
3,478
13,698
14,586
8,836
14,343
6,582
1 1 ,809
17,756
12,439
23,914
32,326
5,219
22,130
ND
ND
LDS Flow
Avg.
(Iphd)
178
24
2
131
290
0
15
35
101
37
53
34
48
1
0
0
0
0
0
0
0
0
3
6
0
15
1
0
Peak
(Iphd)
823
70
11
524
514
0
45
151
860
92
93
47
189
7
0
0
0
0
0
0
0
0
12
28
0
65
12
0
Ea

94.57
99.47
99.95
97.94
91.84
100
99.90
98.97
96.01
97.30
98.58
99.37
99.02
99.91
100
100
100
100
100
100
100
100
99.97
99.95
100
99.86

Active Period of Operation'2'
Time
Period*6'
(months)
11-17
13-26
13-21


3-33
6-33
6-56
13-56
12-80
10-80
11-76
15-71
10-65
8-59
17-62
13-56
8-53
12-38
13-37
11-29



13-31
7-47
12-23
12-23
LCRS Flow
Avg.
(Iphd)
393
2,041
3,108


540
281
307
75
56
168
234
439
41
374
150
803
1,408
281
299
819



3,473
2,494
5,422
2,284
Peak
(Iphd)
1,403
4,282
1 1 ,669


2,383
570
1,075
187
191
655
851
1,384
159
645
337
3,029
9,294
449
561
5,096



5,054
8,983
14,042
7,945
LDS Flow
Avg.
(Iphd)
3
26
11


0
2
4
1
2
0
2
0
0
0
0
0
0
0
0
0



2
6
0
0
Peak
(Iphd)
15
142
54


0
21
47
13
37
0
9
0
0
0
0
0
0
0
0
0



22
25
0
1
Ea

99.29
98.73
99.65


100
99.33
98.78
98.75
96.67
100
99.20
100
100
100
100
100
100
100
100
100



99.94
99.78
100
100
Post-Closure Period13'
Time
Period*6'
(months)

27-36
22-34


34-83
34-83
57-81
57-81

















LCRS Flow
Avg.
(Iphd)

567
189


66
178
206
47

















Peak
(Iphd)

1389
779


94
421
458
84

















LDS Flow
Avg.
(Iphd)

59
2


0
0
1
0

















Peak
(Iphd)

133
8


0
0
10
0

















Ea


89.59
98.78


100
100
99.55
100

















Ul
oo
         Notes:
         (1) "Initial Period of Operation" represents period after waste placement has started and only a small amount of waste has been placed in the cell.
         (2) "Active Period of Operation" represents period when waste thickness in cell is significant and/or an effective intermediate cover is placed on the waste.
         (3) "Post Closure Period" represents period after final cover system has been placed on the entire cell.
         (4) Flow rates are given in liter/hectare/day.
         (5) NA = not applicable; ND = not determined.
         (6) Breakthrough time for steady-state saturated flow through GCL component of composite liner is estimated to be 2 months based on a calculation using Darcy's equation and a saturated hydraulic conductivity of
            5 x 10~11 m/s, hydraulic gradient of 5, and effective porosity of 0.2.  For this calculation, it is assumed that flow through the GM component of the composite liner occurs through small holes and is instantaneous.

-------
The above data indicate that GM/GCL composite liners can be constructed to achieve
Et values of 99.9% or more.  However, Et values in the range of 99 to 99.9% will also
occur.  These high efficiencies demonstrate that the GCL component of a GM/GCL
composite liner is effective in impeding leakage through holes in the GM component of
the liner.

GM/CCL  and GM/GCL/CCL Composite Liners
The performance of 13 of the 88 cells with GM/CCL or GM/GCL/CCL primary liners was
assessed. The remaining 75 cells with GM/CCL or GM/GCL/CCL primary liners were
generally excluded from the assessment because: (i) they did not have continuous
LCRS and LDS flow rate data available for an individual cell from the start of operation;
or (ii) there were insufficient LCRS and LDS chemical constituent data to evaluate
whether primary liner leakage did or did not occur. Flow rate data are available for 13
cells at nine landfills with monitoring periods of up to 121 months.  All of these cells
were constructed with CQA.  A summary of the LCRS and LDS flow rate data for the
cells is provided in Table 5-4. The main findings from the evaluation of flow rate data
for cells with GM/CCL or GM/GCL/CCL primary liners are given below:

   •   LDS flows during the initial period of operation are attributed primarily to
       construction water. LDS flows during the active and post-closure periods are
       attributed primarily to consolidation water.
   •   Average monthly LDS flow rates ranged from about 10 to 1,400  Iphd during the
       initial period of operation, 0 to 370  Iphd during the active period,  and 5 to 210
       Iphd during the post-closure period.

Given the "masking" effects of consolidation water, chemical constituent data must be
used to assess the hydraulic performance of composite primary liners having a CCL or
GCL/CCL lower component,  as described  in Section 5.2.3.2. This approach was
applied to the 13 landfill cells. Concentrations of chemical constituents in LDS liquids
were compared to concentrations of the same constituents in LCRS liquids.  These
chemical  data are reported in Table E-4.9 of Appendix E. The general water quality
characteristics of LDS liquids were found to be different than the corresponding
characteristics for the LCRS  liquids.  This is due to the different origins of the primary
sources of the  two liquids: leachate for LCRS  liquids and CCL pore water for LDS
liquids. The different origins  of the two liquids are reflected in different major ion
chemistries, as well as differences in chemical oxygen demand (COD), biological
oxygen demand (BOD), and total organic carbon (TOC) concentrations.

To further evaluate whether primary liner leakage had contributed to the LDS flows, the
concentrations of the previously mentioned five key chemical constituents (i.e., sulfate,
chloride, benzene, toluene, and xylene) in LCRS and LDS liquids were investigated.
The results of the comparison of key constituents are presented in Tables 5-5 and 5-6.
Table 5-5 presents the concentrations of the five key constituents as a function of time
                                     5-19

-------
       Table 5-4. Summary of LCRS and LDS Flow Rate Data for Landfill Cells with GM/CCL or GM/GCL/CCL
                Composite Primary Liners.
Cell
No.


B3(b)







Y2



AD1







AD7




AK1
AL1

AM1




Cell
Area

(ha)
6.4







3.0



0.6







1.5




1.4
14.9

3.2/2.4(7)




Waste
Place.
Start
Date
Jul-87







Jan-91



May-85







Sep-87




Oct-93
1990

Oct-90




Final
Closure
Date

NA(6)







NA



Jul-88







Oct-93




NA
NA

NA




Initial Period of Operation01
Time
Period
(months)
1-4







1-10



1-12







1-12




1-12
1-29

1-9




LCRS Flow(4)
Avg.
(Iphd)
15,304







23,368



ND(6)







12,597




9,867
ND

ND




Peak
(Iphd)
24,858







36,791



ND







26,492




17,986
ND

ND




LDS Flow
Avg.
(Iphd)
1,394







655



ND







135




206
ND

ND




Peak
(Iphd)
4,250







1,768



ND







1,101




804
ND

ND




Active Period of Operation (2)
Time
Period
(months)
5-16
17-28
29-40
41-52
53-64
65-76
77-88
89-93
11-22
23-34
35-46
47-54
13-20
21-32






13-24
25-36
37-48
49-60
61-69

30-41
42-54
10-21
22-33
34-45
46-57
58-69
LCRS Flow
Avg.
(Iphd)
5,700
9,272
7,575
2,859
1,189
403
560
578
10,353
11,344
4,404
4,397
ND
373






2,212
1,539
1,429
249
480

934
1,349
270
236
111
20
18
Peak
(Iphd)
8,935
22,444
13,978
6,043
2,280
490
919
648
19,204
25,309
6,380
5,199
ND
892






2,857
2,755
2,813
629
614

2,085
5,885
533
329
283
77
21
LDS Flow
Avg.
(Iphd)
124
101
262
231
45
92
102
98
370
90
70
48
ND
107






71
96
17
33
64

231
103
15
10
3
1
1
Peak
(Iphd)
266
168
803
713
152
133
193
109
1,993
168
248
56
ND
603






291
393
21
74
112

367
183
64
15
14
1
1
Ea


(%)
97.8
98.9
96.5
91.9
96.2
77.3
81.8
83.0
96.4
99.2
98.4
98.9

71.4






96.8
93.8
98.8
87.0
86.6

75.3
92.4
94.4
95.8
97.3
95.0
94.4
Post-Closure Period'3'
Time
Period
(months)












33-44
45-51
52-63
64-75
76-87
88-99
100-111
112-121
70-81
82-87











LCRS Flow
Avg.
(Iphd)












145
85
3
3
3
1
1
2
375
165











Peak
(Iphd)












652
130
22
42
21
4
2
9
533
334











LDS Flow
Avg.
(Iphd)












24
26
28
42
23
8
5
6
73
105











Peak
(Iphd)












42
31
45
103
68
46
43
24
157
172











Ea


(%)












83.4
69.5
-833
-1300
-667
-700
-400
-200
80.5
36.3











CJ1

o

-------
          Table 5-4.  Summary of LCRS and  LDS Flow Rate Data for Landfill Cells with GM/CCL or GM/GCL/CCL
                         Composite Primary Liners (Continued).
Cell
No.


AM1
AM2





AO1


AO2

AQ1



AQ10


AR1

Cell
Area

(ha)

4.8/2.4(7)





1.8


1.8

0.6



0.9


9.7

Waste
Place.
Start
Date

Oct-90





Jan-92


Jul-92

Mar-86



Jan-89


Mar-92

Final
Closure
Date


NA





NA


NA

early 90



mid 91


NA

Initial Period of Operation (1)
Time
Period
(months)

1-9





1-5


1-5

1-6



1-9


1-11

LCRS Flow(4)
Avg.
(Iphd)

ND





ND


15,881

10,203



ND


27,042

Peak
(Iphd)

ND





ND


24,541

18,944



ND


65,871

LDS Flow
Avg.
(Iphd)

ND





ND


149

352



14


292

Peak
(Iphd)

ND





ND


191

569



32


705

Active Period of Operation (2)
Time
Period
(months)
70-81
10-21
22-33
34-45
46-57
58-69
70-81
6-17
18-29
30-37
6-17
18-31
7-25
26-34
35-46
47-58
10-14
15-26

12-23
24-36
LCRS Flow
Avg.
(Iphd)
11
32
35
17
67
64
112
1,984
1,299
1,144
3,027
1,688
ND
ND
ND
4,530
ND
15,933

11,251
9,668
Peak
(Iphd)
18
154
51
45
274
181
136
4,130
1,577
1,371
5,266
2,383
ND
ND
ND
10,531
ND
38,751

23,384
26,274
LDS Flow
Avg.
(Iphd)
5
9
9
3
0
8
9
184
96
60
110
33
255
ND
197
116
26
48

181
155
Peak
(Iphd)
8
42
29
26
0
13
13
353
126
102
158
64
1239
ND
435
143
32
250

470
442
Ea


(%)
54.4
71.9
74.3
82.4
100
87.5
92.0
90.7
92.6
94.8
96.4
98.1



97.4

99.7

98.4
98.4
Post-Closure Period13'
Time
Period
(months)












59-65
66-77
78-89
90-97
27-38
39-50
51-63


LCRS Flow
Avg.
(Iphd)












5,835
644
1,367
1,615
682
300
852


Peak
(Iphd)












1 1 ,244
1,011
3,264
3,575
2,251
1,709
1,588


LDS Flow
Avg.
(Iphd)












215
117
98
51
29
18
24


Peak
(Iphd)












246
165
132
118
48
63
75


Ea


(%)












96.3
81.8
92.8
96.8
95.7
94.0
97.2


en
ro
          Notes:
          (1) "Initial Period of Operation" represents period after waste placement has started and only a small amount of waste has been placed in the cell.
          (2) "Active Period of Operation" represents period when waste thickness in cell is significant and/or an effective intermediate cover is placed on the waste.
          (3) "Post-Closure Period" represents period after final cover system has been placed on the entire cell.
          (4) Flow rates are given in liter/hectare/day.
          (5) 65 percent of Cell B3 received final cover at 60 months after start of waste placement.
          (6) NA = not applicable;  ND = not determined.
          (7) Values given represent LCRS and LDS areas, respectively.
          (8) Breakthrough times for steady-state saturated flow through CCL or GCL/CCL component of composite liners are estimated to be 2 to 145 months based on a
             calculation using Darcy's equation and specified hydraulic conductivities, hydraulic gradient of 5 forGCLs and 1 forCCLs, and effective porosity of 0.2.
             For this calculation, it is assumed that flow through the GM component of the composite liner occurs through  small holes and is instantaneous.

-------
            Table 5-5. Average Concentrations of Five Key Chemicals in LCRS and LDS Flows from

                     Landfill Cells with GM/CCL and GM/GCL/CCL Composite Primary Liners.
Ul

ro
ro

Cell
No.
B3








Y2




AD1










Time
Period
(months)
1-4
5-16
17-28
29-40
41-52
53-64
65-76
77-88
89-93
1-10
11-22
23-34
35-46
47-54
1-12
13-20
21-32
33-44
45-51
52-63
64-75
76-87
88-99
100-111
112-121
Chemical'1'
Sulfate (mg/l)
LCRS
282
105
348
104
47
28
<127
<6

108
108
52


6,353
5,830
5,470
4,455
2,223


1,785
4,488
3,633
3,870
LDS
95
1,286

500
14
123
90
301
48
231
299
326




480
498
308
443
338
456
339
296
369
Chloride (mg/l)
LCRS
25
207
352
580
355
899
203
998
1,383
349
590
876


3,930
24,300
10,763
11,590
13,960


13,900
14,550
14,075
14,800
LDS
19
173
241
118

59
58


60
25
89



289
210
185
217
240
131
377
137
114
138
Benzene (ug/l)
LCRS
<11(2)
<1
<1
<5
8
7
<5
<5
<2

10



492
429
33
5
35



26
18
<25
LDS
<25
<1

<5
6
<1
<5
<5
<5






<4
<4
<4
<4
<4
<4
<4
<4
<1
<1
Toluene (ug/l)
LCRS
150
<1
<1
354
233
101
14
<5
<1

720



305
292
133
60
108
<300


186
<30
<25
LDS
<25
<1

<6
24
<1
<6
<4
<1

7




<6
<6
<6
<6
<6
<6
<6
<6
<1
<1
Xylene (ug/l)
LCRS

























LDS


























-------
           Table 5-5. Average Concentrations of Five Key Chemicals in LCRS and LDS Flows

                     from Landfill Cells with GM/CCL and GM/GCL/CCL Composite Top Liners (Continued).
en

*>
CO

Cell
No.
AD7







AK1
AL1


AM1





AM2





A01



Time
Period
(months)
1-12
13-24
25-36
37-48
49-60
61-69
70-81
82-87
1-12
1-29
30-41
42-54
1-9
10-21
22-33
34-45
46-57
58-69
1-9
10-21
22-33
34-45
46-57
58-69
1-5
6-17
18-29
30-37
Chemical
Sulfate (mg/l)
LCRS
2,818
3,620
7,361
8,213
6,867
5,740
6,998
7,480
47
300
225
247
51
<2
<16
<27
<12
<3
96
<2
<13
<7
<2
<3

49
35
69
LDS
340
683
586
954
1,050
1,148
1,168
1,132
16
1,030
900
1,375


1,341
1,200




1,032
1,300
1,730
2,405

88
41
2
Chloride (mg/l)
LCRS
3,214
9,550
10,720
11,535
14,400
15,775
12,875
14,267
104
330
273
400
77
120
159
219
265
240
140
290
353
326
368
262

930
988
570
LDS
109
216
219
469
418
387
387
357
2
89
203
215


2,260
2,600




2,175
2,600
2,700
2,635

58
27
46
Benzene (ug/l)
LCRS
140
612
1168
644
778
687
540
<240
<5
<4
1
2
<21
18
<19
18
14
13
11
17
19
18
15
7

6
10
<5
LDS
<4
<4
<4
<4
<4

<4
<1
<1
<2
<1
<1

<1
<1
<1



<1
<1
<1
<1
<1

<1
<1
<5
Toluene (ug/l)
LCRS
317
892
1,859
2,960
1,660
1,288
906
450
88
133
5
2
219
160
336
290
199
56
22
89
266
286
148
65

230
288
77
LDS
<6
<6
<6
<6
<6
<6
<1
<1
<1
<1(3)
<1
<1

2
<1
<1



<1
<1
<1
<1
<1

<1
44
<5
Xylene (ug/l)
LCRS








30
540
<3
<3
150
90
121
122
90
82
34
57
95
94
115
50

59
45
21
LDS








<3
<1

<1

<1
<1
<3



<2
<1
<3
<2
<2

<3
<3
<10

-------
en
                Table 5-5.  Average Concentrations of Five Key Chemicals in LCRS and LDS Flows
                            from Landfill Cells with GM/CCL and GM/GCL/CCL Composite Top Liners (Continued).

Cell
No.
A02


AR1


AQ1




AQ10




Time
Period
(months)
1-5
6-17
18-31
1-11
12-23
24-36
1-58
59-65
66-77
78-89
90-97
1-15
15-26
27-38
39-50
51-63
Chemical
Sulfate (mg/l)
LCRS

49
35
180
440
520










IDS

89
93
600
170
265










Chloride (mg/l)
LCRS

930
988
1,000
2,200
1,650










IDS

16
24
49
8
41










Benzene (u,g/l)
LCRS

6
10


<100

<5
<8
<12
<5

<10
<10
<6
<5
LDS

<1
<1

<1
<50

<4
<4
<4
<4

<4
<4
<4
<4
Toluene (u,g/l)
LCRS

230
288


<100

<5
<5
<5
<5

<12
<30
<6
5
LDS

<1
<1

<2
<50

<10
<6
<14
<5

<5
<5
<5
<5
Xylene (ug/l)
LCRS

59
45


<100


<5
<8
<5


190
<7
10
LDS

<3
<4

<4
<50










                Notes:
                (1)  Reported concentrations represent average of 1 to 17 individual analysis results (typically on the order of 5) during incremental reporting period.
                (2)  Data preceded by "<" indicates more than half of analysis results for parameter were reported as non-detects; in calculating
                   average values, half of the test detection limit was conservatively used for all results reported as non-detects.
                (3)  For Cell AL1, toluene was not detected in nine often LDS flow samples obtained during the 1-41 months time period.  Toluene was detected
                   at a concentration of 91 |ig/l in month 30. This one detection is attributed to sampling or analysis error and is not included in the average.

-------
      Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and
                GM/GCL/CCL Composite Primary Liners.

Cell
No.



B3







Y2









Monitor.
Period
(months)


93







54
(no key
chemical
data
after 34
months)



Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
46







35








Chemical
Sulfate




not diagnostic due
to fluctuating Co(2)
in both LCRS and
LDS




not diagnostic due
to high C0 in LDS
consolidation
water





Chloride




lower C0 in LDS
than in LCRS and
trend of
decreasing LDS
C0 with time not
indicative of
chloride
breakthrough
in LCRS, C0 =
170to1,160mg/l
with m(2) = 628
mg/l; in LDS, C0 =
8 to 1 40 mg/l with
m = 58 mg/l; no
indication of
chloride
breakthrough
Benzene




not diagnostic
due to very low
C0 in both LCRS
and LDS (i.e., C0
almost always
below DL(3) of 5
H9/I)

no LDS data
available







Toluene




in LCRS, C0 up to
700 |ig/l; in LDS,
C0 typically below
DLof 1 to 10
|ig/l; no indication
of toluene
breakthrough

only one C0
available from
each system (at
11 -22 months):
LCRS C0 = 720
|ig/l and LDS C0
= 7 ug/l
f-^y

Xylene




no data
available






no data
available








Summary of
Observations for Five
Key Constituents


no evidence of
significant leachate
migration into LDS
after almost 8 years of
cell operation, twice
the estimated CCL
breakthrough time

data are insufficient to
draw conclusions;
monitoring period is
about equal to
estimated CCL
breakthrough time;
more chemical data
are needed

en
en

-------
      Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and
                GM/GCL/CCL Composite Primary Liners.
Cell
No.
AD1









AD7













Monitor.
Period
(months)
121









87













Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
70









70













Chemical
Sulfate
in LCRS, C0 =
1,785 to 6,353
mg/l with m =
4,234 mg/l; in
LDS, C0 = 296 to
498 mg/l with m =
392 mg/l; no
indication of
sulfate
breakthrough
in LCRS,
C0=2,818to8,213
mg/l with m=6,137
mg/l; in LDS,
C0=340to 1,168
mg/l with m=882
mg/l; increasing
LDS C0 after 36
months attributed
to decreasing
dilution of consol.
water by
construct, water

Chloride
in LCRS, C0 =
3,930 to 24,300
mg/l, with m =
13,450 mg/l; in
LDS, C0 = 114 to
337 mg/l, with m =
204 mg/l; no
indication of
chloride
breakthrough
in LCRS,
C0=3,214to
15,775 mg/l with
m=1 1,547 mg/l; in
LDS, C0=109to
469 mg/l with
m=320 mg/l;
increasing LDS 0,
after 36 months
attributed to
decreasing dilutior
of consol. water
by construct.
water
Benzene
in LCRS, C0 =
<25to492|ig/l;
in LDS, C0 below
DLof 1 to4|ig/l;
no indication of
benzene
breakthrough



in LCRS, C0 =
<240to 1,168
|ig/l; in LDS, C0
below DL of 1 to
4 |ig/l; no
indication of
benzene
breakthrough






Toluene
in LCRS, C0 =
<25to305|ig/l;
in LDS, C0 below
DLof 1 to6|ig/l;
no indication of
toluene
breakthrough



in LCRS, C0 =
317to2,960|ig/l;
in LDS, C0 below
DLof 1 to6|ig/l;
no indication of
toluene
breakthrough







Xylene
no data
available








no data
available












Summary of
Observations for Five
Key Constituents
no evidence of
significant leachate
migration into LDS
after 10 years of cell
operation and closure,
1 .7 times more than
the estimated CCL
breakthrough time


evidence of possible
breakthrough for
sulfate & chloride at 12
36 months; authors
attribute trend to
decreased dilution of
consolidation water by
construction water; no
evidence of organic
constituent
breakthrough; more
chemical data are
needed

en
en

-------
      Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and
                GM/GCL/CCL Composite Primary Liners.

Cell
No.



AK1









AL1









Monitor.
Period
(months)


12









54








Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
48









70








Chemical
Sulfate




in LCRS, C0 = 7
to 1 1 0 mg/l with m
= 47 mg/l; in
LDS, C0= 10 to
51 mg/l with m =
16 mg/l; no
indication of
sulfate
breakthrough

not diagnostic due
to high C0 in LDS
consolidation
water





Chloride




in LCRS, C0 = 2
to 230 mg/l with m
= 104 mg/l; in
LDS, C0 = 2 to 6
mg/l with m = 4
mg/l; no indication
of chloride
breakthrough


increasing LDS Q,
with time likely
due to decreasing
dilution of
consolidation
water by
construction water


Benzene




not diagnostic
because C0 is
below DL in both
LCRS and LDS






not diagnostic
because C0 is
below DL in both
LCRS and LDS





Toluene




in LCRS, C0 = 5
to 300 |ig/l with m
= 88 |ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
toluene
breakthrough


in LCRS, C0 = up
to 600 |ig/l; in
LDS, toluene
below DL of 1
|ig/l; no indication
of toluene
breakthrough


Xylene




in LCRS, xylene
detected in half
of sampling
events at C0 up
to 79 |ig/l; in
LDS, C0 below
DL of 3 |ig/l; no
indication of
xylene
breakthrough
not diagnostic
because C0 is
below DL in
LDS and, after
29 months, also
in LCRS




Summary of
Observations for Five
Key Constituents


no evidence of
significant leachate
migration into LDS;
however, monitoring
period is only about
1/4th of the estimated
GCL/CCL
breakthrough time;
more chemical data
are needed
no evidence of
significant leachate
migration in to LDS;
monitoring period
somewhat less than
estimated CCL
breakthrough time;
more chemical data
are needed
CJl
ro

-------
      Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and

                GM/GCL/CCL Composite Primary Liners.
Cell
No.
AM1








AM2








Monitor.
Period
(months)
58








58








Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
4








4








Chemical
Sulfate
not diagnostic due
to high C0 in LDS
consolidation
water





not diagnostic due
to high C0 in LDS
consolidation
water





Chloride
not diagnostic due
to high C0 in LDS
consolidation
water





not diagnostic due
to high C0 in LDS
consolidation
water





Benzene
inLCRS, C0= 12
to 20 |ig/l; in
LDS, C0 below
DLof 1 |ig/l; no
indication of
benzene
breakthrough


in LCRS, C0 = 5
to 20 |ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
benzene
breakthrough


Toluene
in LCRS, C0 = 40
to 420 |ig/l with m
= 267 |ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
toluene
breakthrough

inLCRS, C0 = 10
to 400 |ig/l with m
= 146|ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
toluene
breakthrough

Xylene
in LCRS, C0 =
71 to 150|ig/l
with m = 1 22
|ig/l; in LDS, C0
below DL of 1 to
3 |ig/l; no
indication of
xylene
breakthrough

in LCRS, C0 = 2
to 1 30 |ig/l with
m = 71 mg/l; in
LDS, C0 below
DL of 1 to 3
mg/l; no
indication of
xylene
breakthrough

Summary of
Observations for Five
Key Constituents
no evidence of
significant leachate
migration into LDS
after almost 5 years of
cell operation;
monitoring period
more than 12 times
longer than estimated
CCL breakthrough
time
no evidence of
significant leachate
migration into LDS
after almost 5 years of
cell operation;
monitoring period
more than 12 times
longer than estimated
CCL breakthrough
time
CJ1

ro
oo

-------
      Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and
                GM/GCL/CCL Composite Primary Liners.
Cell
No.
A01(4)









A02(4)










Monitor.
Period
(months)
37









31










Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
140









145










Chemical
Sulfate
not diagnostic due
to similar LCRS
and LDS C0
ranges






not diagnostic due
to similar LCRS
and LDS C0
ranges







Chloride
in LCRS, C0 =
320 to 1 300 mg/l
with m = 860 mg/l;
in LDS, C0 = 7 to
1 00 mg/l with m =
40 mg/l; no
indication of
chloride
breakthrough


in LCRS, C0 =
320 to 1 ,300 mg/l
with m = 862 mg/l;
in LDS, C0 = 3 to
34 mg/l with m =
24 mg/l; no
indication of
chloride
breakthrough


Benzene
in LCRS,
benzene
detected in half ol
the sampling
events at C0 = 7
to 12|ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
benzene
breakthrough
in LCRS,
benzene
detected in half o
the sampling
events at C0 = 7
to 12|ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
benzene
breakthrough
Toluene
in LCRS, C0 = 10
to 550 |ig/l with m
= 167|ig/l; in
LDS, C0 below
DL of 1 |ig/l in 2/3
of sampling
events; no
indication of
toluene
breakthrough

in LCRS, C0 = 10
to 550 |ig/l with m
= 167|ig/l; in
LDS, C0 below
DL of 1 |ig/l; no
indication of
toluene
breakthrough



Xylene
in LCRS, C0 =
12to76|ig/l
with m = 34|ig/l;
in LDS, C0
below DL of 3
|ig/l; no
indication of
xylene
breakthrough


in LCRS, C0 =
12to76|ig/l
with m = 34|ig/l;
in LDS, C0
below DL of 3
|ig/l; no
indication of
xylene
breakthrough


Summary of
Observations for Five
Key Constituents
no evidence of
significant leachate
migration into LDS
after 3 years of cell
operation; however,
monitoring period is
only 1/4th of estimated
CCL breakthrough
time; more chemical
data are needed

no evidence of
significant leachate
migration into LDS
after 3 years of cell
operation; however,
monitoring period is
only 1/4th of estimated
CCL breakthrough
time; more chemical
data are needed

en

CO

-------
        Table 5-6.  Evaluation of Chemical Constituent Migration Through Landfill GM/CCL and
                      GM/GCL/CCL Composite Primary Liners.

Cell
No.



AQ1



AQ10



AR1









Monitor.
Period
(months)


97



63



36








Estimated
Advective
Breakthr.
Time for
GCL/CCL
(months)'1'
35



35



2








Chemical
Sulfate




no data available



no data available



not diagnostic due
to similar LCRS
and LDS C0
ranges





Chloride




no data available



no data available



in LCRS, C0 =
600 to 2700 mg/l
with m = 1 625
mg/l; in LDS, C0
= 8 to 74 mg/l with
m = 35 mg/l; no
indication of
chloride
breakthrough
Benzene




not diagnostic
because C0 is
below DL in both
LCRS and LDS
not diagnostic
because C0 is
below DL in both
LCRS and LDS
not diagnostic
because C0 is
below DL in both
LCRS and LDS





Toluene




not diagnostic
because C0 is
below DL in both
LCRS and LDS
not diagnostic
because C0 is
below DL in both
LCRS and LDS
not diagnostic
because C0 is
below DL in both
LCRS and LDS





Xylene




no LDS data
available


no LDS data
available


not diagnostic
because C0 is
below DL in
both LCRS and
LDS





Summary of
Observations for Five
Key Constituents


data are insufficient to
draw conclusions;
more data are needed

data are insufficient to
draw conclusions;
more data are needed

no evidence of
significant leachate
migration into LDS;
monitoring period is
more than 10 times the
estimated GCL/CCL
breakthr. time; data
are not diagnostic;
more data are needed
CO
o
        Notes:
        (1) Advective breakthrough times for steady-state saturated flow through CCL or GCL/CCL component of composite liners were calculated
           using Darcy's equation and specified hydraulic conductivities, hydraulic gradient of 5 for GCLs and 1  for CCLs, and effective porosity of 0.2.
           For this calculation, it is assumed that flow through the GM component of the composite liner occurs  through small holes and is instantaneous.
        (2) C0 = average concentration during incremental reporting period; m = mean concentration for the entire reporting period.
        (3) DL = detection limit.
        (4) Composite liquid quality samples from the LCRSs of Cells AO1 and AO2 were  assumed to represent average conditions at the two cells.

-------
period after the start of landfill cell operation.  Table 5-6 presents the results of the
authors' assessment of the occurrence of key constituent migration through the
composite primary liners.  This assessment is based on a qualitative comparison of the
five key chemical constituents. Table 5-6 also presents an estimate of the advective
breakthrough time for the CCL or GCL/CCL component of each composite primary liner.
The estimated breakthrough times were calculated assuming that the  GM component of
the composite primary liner has one or more holes through which leachate
instantaneously migrates and that leachate migration through the CCL or GCL/CCL
component of the composite liner is governed by Darcy's equation assuming one-
dimensional steady-state saturated flow.   Other assumptions used in the calculations
are given in the table. The effect of chemical retardation was not considered in
calculating the advective breakthrough times. Retardation of chloride  and sulfate should
be negligible. Retardation characteristics for benzene, toluene, and xylene will depend
on the organic carbon content of the CCL or GCL, redox conditions, and other factors.
It is expected, however, that the effective retardation coefficient for these constituents
would have been 2 or more. These organic compounds were chosen  for analysis
notwithstanding their retardation characteristics for a combination of reasons, including
relatively widespread occurrence in leachate, and relatively higher concentrations in
leachate than other organic compounds.   In addition, these three constituents are not
known as laboratory contaminants, in contrast to methylene chloride, a constituent that
is more  mobile but is also a common laboratory contaminant.

The current database is not sufficient to draw definitive  conclusions on the performance
of GM/CCL and GM/GCL/CCL composite primary liners. However, using the data and
comparisons in Tables 5-5 and 5-6, the following observations can be  offered with
respect  to key chemical constituent migration through the composite primary liners of
the 13 considered cells:

   •   There were insufficient data for three  cells (i.e.,  Y2, AQ1, and AQ2) to draw any
       conclusions on primary liner leakage rates based on key chemical constituent
       data.
   •   For the remaining ten cells, key chemical constituent data did not reveal  obvious
       indications of primary liner leakage.
   •   One  of the ten cells (i.e., AD7) exhibited a potential indication of primary  liner
       leakage when sulfate and chloride concentrations in LDS flows increased
       between 12 and 36 months after  construction.  However, the concentrations of
       other chemicals did not increase  over time. The estimated breakthrough time
       for the composite primary liner in this  cell is 70 months, several times greater
       than  the time when sulfate and chloride concentrations increased. The reason
       for the increase in the anion concentrations in the LDS flow from Cell AD7 is
       unclear.
   •   Five  of the ten cells (i.e., B3, AD1, AD7, AM1, and AM2) have  key chemical
       constituent data of sufficient completeness and  duration to conclude that
       leachate migration into the LDS at a rate of any engineering significance has not

                                      5-31

-------
       occurred for a time period exceeding the estimated breakthrough time for the
       CCL component of the composite liner.
   •   Et values were estimated for cells B3, AD1, AD7, AM1, and AM2 using Equation
       5-1, presented in Section 5.2.3, with constituent mass fluxes from the LCRS and
       LDS.  Mass fluxes were calculated using average flow rates and chemical
       concentrations for benzene, toluene, and xylene during the active operation and
       post-closure periods.  With this approach, Et values for these cells were found to
       range from 99.1 to more than 99.9%.
   •   Based on the above data and similar to GM/GCL composite liners, GM/CCL and
       GM/GCL/CCL composite liners of the type evaluated in this study can be
       constructed to achieve Et values of 99.9% or more.  However, Et values in the
       range of 99 to 99.9% will also occur.
   •   Available leakage rate calculation methods for composite liners give leakage
       rates in the same range as the rates estimated from the data for composite
       primary liners presented in Appendix E.  Notwithstanding the uncertainties in
       both the assumptions used in the calculations and the estimated leakage rates,
       this is a useful finding.
   •   In the U.S., landfill cells are typically operated for periods of one to five years,
       occasionally longer, and they are promptly covered with a GM or other low-
       permeability barrier after filling.  This operations sequence defines the timeframe
       for significant leachate generation in a landfill cell that does not contain liquid
       wastes or sludges and that does not undergo leachate recirculation or moisture
       addition.  For the cells in this study, estimated advective breakthrough times
       through CCLs, assuming no chemical retardation, were generally calculated to
       range from about 3 to 12 years (see Table 5-6).  It thus appears that GM/CCL
       and GM/GCL/CCL composite liners are capable  of substantially preventing
       leachate migration over the entire period of significant leachate generation for
       typical modern landfills.
   •   The conclusions given above for GM/CCL composite liners should be
       considered  preliminary. Additional analyses are  recommended using a larger
       database representing a larger time period of operation to confirm or modify
       these preliminary conclusions. The additional analyses should include a more
       thorough analysis of the transport characteristics of a wider array of key
       chemical constituents than considered in this study.

5.2.4.2  Leachate Generation Rates
Average and peak monthly LCRS flow rate data were evaluated for 73 MSW cells at 32
landfills, 56 HW cells at 12 landfills, eight MSW ash or coal ash cells at six landfills, and
three C&DW cells at two landfills.   Most of these landfills are located in the northeast
and southeast; only four of the landfills are located in the west. The LCRS flow rate
data are presented in Table E-3.5 of Appendix E for monitoring periods up to about ten
years.  Almost half of the cells have more than four years of LCRS flow rate data
available. Post-closure data are available for eleven MSW cells at three landfills and 22
HW cells at five landfills. Detailed  results of the data analysis are given in Tables E-5.1
to E-5.5 of Appendix E and are summarized in Tables 5-7 and 5-8 below.  The range of
                                      5-32

-------
average LCRS flow rates for the cells and the mean average LCRS flow rates are
presented in Table 5-7 as a function of waste type, landfill operational stage, and
geographical region of the U.S.  Figure 5-4 illustrates the effects of geographic region
and waste type on LCRS flow rate. Table 5-8 presents the range of average rainfall
fractions (RFs) for the cells and the mean average RFs as a function of the same
variables.  In this report, RF (in percent) is the ratio of average LCRS flow rate to
historical average annual rainfall.
Table 5-7.  Summary of Average LCRS Flow Rates (in Iphd).
Waste
Type
MSW
HW
Ash
C&DW
U.S.
Region
NE
SE
W
NE
SE
W
NE
SE
NE
Initial Period of Operation
Range
1,050-39,900
1,480-43,700
4,980-18,800
480-31,300
42-3,090
2,190-28,600
15,600-19,600
Mean
10,200
10,400
11,000
15,500
480
18,700
17,600
Active Period of
Operation
Range
41-17,700
300-10,900
55-110
1,050-21,300
270-37,100
1-4,280
1,030-35,300
8,940-24,490
3,570-16,200
Mean
3,530
2,930
83(1)
5,380
4,890
990
17,700
17,800
10,600
Post-Closure
Period
Range
55-680
340-1,130
36-1,580
56(1)


Mean
400
780
370
56(1)


Notes:  (1) Values are based on only one or two cells from one landfill.
The major findings from the evaluation of leachate generation rates are given below:

   •   LCRS flow rates during operations (i.e., the initial and active periods of
       operation) can vary significantly between landfills located in the same
       geographic region and accepting similar wastes. Large variations in flow rates
       (e.g., one order of magnitude difference) can even occur between cells at the
       same landfill. Differences in waste placement practices  may be responsible for
       these significant variations.  Limiting the size of the active disposal area and
       using effective measures to minimize rainfall  infiltration into the waste and to
       divert surface-water runoff away from the waste will significantly decrease
       leachate generation rates compared to the rates observed under less controlled
       conditions.
                                       5-33

-------
Table 5-8. Summary of Average Rainfall Fractions (in percent).
Waste
Type
MSW
HW
Ash
C&DW
U.S.
Region
NE
SE
W
NE
SE
W
NE
SE
NE
Initial Period of
Operation
Range
4-160
5-157
21-87
1-73
1-30
8-84
50-63
Mean
39
33
46
33
5
58
56(1)
Active Period of
Operation
Range
0.1-54
1-23
0.5-1
4-81
1-74
0.01-41
4-104
22-60
12-52
Mean
13
8
0.7(1)
21
11
10
55
43
34
Post-Closure
Period
Range
0.2-3
1-4
0.08-3
0.3(1)


Mean
1
3
0.8
0.3(1)


Notes:  (1) Values are based on only one or two cells from one landfill.
       The MSW cells produced, on average, less leachate than the HWand ISW cells.
       Average LCRS flow rates for MSW cells located in the NE and SE ranged from
       1,000 to 44,000 Iphd during the initial period of operation and 40 to 18,000 Iphd
       during the active period of operation.  For this group of cells during the initial
       period of operation, 60%  exhibited average LCRS flow rates less than 10,000
       Iphd and 87% had rates less than 20,000 Iphd.  For the same group of cells
       during the active period of operation, 52% had average LCRS flow rates less
       than 5,000 Iphd and 95% had flow rates less than 10,000 Iphd.  Only two MSW
       cells are located in the W. These two MSW cells had very low average LCRS
       flow rates (i.e., 55 and 110 Iphd).
       RF values calculated for the MSW cells in the NE (means of 39% and 13% for
       the initial and active periods of operation, respectively) were higher than RF
       values for the SE cells (means of 33% and 8% for the  initial  and active periods
       of operation, respectively).  It is possible that the higher water evaporation rates
       and the higher runoff occurring with shorter duration, more intense rainfalls
       associated with the SE offset any potential increases in leachate generation
       rates caused by the higher total amount of rainfall in the SE  as compared to the
       NE. RF values for the two MSW cells that are located at an arid site (average
       annual rainfall of about 430 mm)  in the W were less  than 1%.
       Average LCRS flow rates for HW cells located in the NE and SE ranged from
       500 to 31,000 Iphd during the initial period of operation and  300 to 37,000 Iphd
                                      5-34

-------
en

CO
en
40,000 -,
35,000


— 30,000
T3
.C
LU
^ 25,000
O
^ 20,000
w
or
o
_i
g 15,000
<
or
LU
< 10,000
5,000
n
c

AMSW
• HW "
+ ASH +
u&uw ^ ^
4 NE »
* W fc
+
•
+ + +
A •
• *
+•
X
• A *
S '
m A
A A A A
• HA^A A • E
' A AB "|A * X AA 1
• .it- !;4^ . . - ; |
) 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,0
                                               AVERAGE ANNUAL RAINFALL (mm)
      Figure 5-4.  Average LCRS flow rate versus average annual rainfall during the active period of operation.

-------
during the active period of operation. About 69% of these cells exhibited
average LCRS flow rates greater than 10,000 Iphd during the initial period of
operation and 21% exhibited average LCRS flow rates greater than 5,000 Iphd
during the active period of operation. Average LCRS flow rates from HW cells
during the active period of operation were 50 to 70% higher than flow rates from
MSW cells. The reason for the higher leachate generation rates at the HW cells
in this study is unclear,  but may, in part, be due to differences in waste
characteristics (e.g., initial moisture content, porosity, and permeability) and
operational practices (e.g., waste placement and covering procedures). The ten
HW cells located in the  W had low  average flow rates, ranging from about 1  to
4,000 Iphd during operations.
RF values calculated for the HW cells in the NE (means of 46% and 21 % for the
initial and active periods of operation, respectively) were higher than RF values
for the SE cells  (means of 33% and 11% for the initial and active periods of
operation, respectively). Similar to the  MSW cells, the HW cells  in the SE had
lower RF values than cells in the NE. For most of the HW cells in the W, RF
values were less than 10% during operations.
Average flow rates during operations ranged from 1,000 to 35,000 Iphd for ash
cells (1,000 to 25,000 Iphd for the seven MSW ash cells and 35,000 Iphd for the
coal ash cell) and from 4,000 to 20,000 Iphd for the C&DW cells.  The limited
number of MSW ash, coal ash,  and C&DW cells considered in this study
exhibited average LCRS flow rates during the active period of operation that
were 300 to 600% higher than average LCRS flow rates from MSW cells during
the same period. It is possible that the higher leachate generation rates at the
MSW ash, coal  ash, and C&D waste landfills may, in part, be due to differences
in waste characteristics and operational practices.
Mean RF values were 53% for ash cells and 43% for C&DW cells.
Peak monthly LCRS flow rates were typically two to three times the average
monthly flow rates for all types of waste and regions of the U.S.
Landfill geographic region has a major  impact on LCRS flow rates. For landfill
sites with historical average annual rainfall less than 500 mm, average LCRS
flow rates were  low, typically less than 2,000 Iphd. LCRS flow rates increased
with increasing rainfall up to a point.  In general, for landfills with historical
average annual rainfall greater than 1,100 to 1,200  Iphd, an increase in rainfall
did not appear to cause a corresponding increase in leachate generation rate.
LCRS flow rates were typically two to three times smaller during the active
period of operation than during  the initial period of operation.
Leachate generation  rates for the closed landfills in this study typically
decreased by a factor of four within one year after closure and by one order of
magnitude within two to four years after closure, as shown in Figure 5-5. Six
years after closure, LCRS flow rates were between  5 and 1,200 Iphd (mean of
180 Iphd). Nine years after closure, LCRS flow rates were negligible. These
data show that well designed and constructed cover systems can be very
effective in minimizing infiltration of rainfall into the waste, thus reducing
leachate generation rates to near-zero values.
                               5-36

-------
en
[
[
~ 10000 [
%
< 1000
1
3
u- 100
C/)
a
O
_l
g 10
<
tt
LJJ
^ 1


0



! a
1 a
a g o
MM

n n


0





1—1

01234
OMSW 1
DHW !

:
-
0
0 |
O g
D °

a n :
B B n a i
n
n n :
n :
n
a a D
-
D :
n
1—1 n

567891
                                          YEARS SINCE FINAL CLOSURE
       Figure 5-5. Average LCRS flow rates after closure for eleven MSW cells and 22 HW cells.

-------
5.2.4.3  Leachate Chemistry
Select leachate chemistry data for 59 cells at 50 double-lined landfills were evaluated in
terms of average constituent concentrations and relative detection frequencies.  The
distribution of leachate chemistry data by waste type and start of operation date is
presented in Table 5-9. For the purposes of the discussions on leachate chemistry in
this chapter, MSWash landfill leachate is grouped with leachate from ISW landfills. This
grouping is considered appropriate because MSWash landfill leachate is typically
nonhazardous and has chemical  characteristics more similar to leachate from ISW
landfills than to leachates from MSW or HW landfills. The MSW leachate chemistry
data are from 36 landfills located  in all geographic regions of the U.S.  Based on the
extent of the leachate chemistry data, the data are believed to be representative of
modern MSW landfills in the U.S  operated without leachate recirculation or other special
activities (e.g., special waste disposal, induced aerobic degradation). About 70% of
these landfills began operating in the 1990's. While the data for modern MSW landfills
are extensive, they should not be considered to reflect the full range of leachate
chemistry associated with the anaerobic decomposition process, from the acid stage to
the methane fermentation stage.  Moreover, differences will exist from facility to facility
based on a variety of climate, site, waste,  and operational factors. Additional data are
needed from more facilities over a longer time period to better identify the potential
range of leachate chemistry characteristics throughout the initial, active, and post-
closure  operational periods of a facility.

Table 5-9. Distribution of LCRS Chemistry Database by Waste Type and Start of
           Operation Date.
Waste Type
MSW
HW
MSW Ash
Coal Ash
C&DW
Pre-1990
Start of Operation
11 landfills
13 cells
3 landfills
5 cells
1 landfill
1 cell
1 landfill
1 cell
1 landfill
2 cells
Post-1990
Start of Operation
25 landfills
28 cells
1 landfill
1 cell
6 landfills
6 cells
1 landfill
1 cell
1 landfill
1 cell
Fewer data are available for HW and ISW landfills than for MSW landfills. In addition,
the types of wastes placed in HWand ISW landfills are generally more variable between
landfills than wastes placed in MSW landfills. With the exception of the leachate
chemistry data for MSWash landfills, it is likely that the data presented  in this report do
not characterize the variation in leachate chemistry for HW and ISW landfills.  The
                                       5-38

-------
chemistry data for MSWash landfill leachate may be representative of modern MSW
ash landfills in the U.S. because seven landfills are included in the database and the
chemistry of MSW ash is less variable than that of HW.

The leachate  chemistry data are presented in Table E-3.7 of Appendix E and
summarized in Table 5-10.  Federal MCLs, which are available for two of the heavy
metals and ten of the VOCs considered in this study, are also listed in Table 5-10. The
distributions of select chemistry data for MSW, HW,  and MSWash cells are shown in
Figures E-6.1  to E-6.3 of Appendix E.  For  MSW landfills, the chemical data for older
landfills that started operating before 1990  (pre-1990 cells) and newer landfills that
started operating during 1990 or later (post-1990 cells) are compared (Figure E-6.4 of
Appendix E).  The major findings from  the evaluation of leachate chemistry data are
given below:
   •   For a  given waste type,  many of the leachate constituents exhibited significant
       concentration variations (e.g., several orders of magnitude difference) between
       landfill cells and, sometimes, for a given cell.
   •   For the leachate types for which data are available for more than two landfills,
       the average value of pH (pH units), specific conductance (jimhos), COD (mg/l),
       BOD5  (mg/l), TOC (mg/l), and chloride (mg/l) were, respectively:
             o  MSW leachate: 6.7, 4,470, 2,500,  1,440, 380,  and 560;
             o  HW leachate: 8.2, 22,100, not available,  not available, 1,620, and
                7,760; and
             o  MSW ash leachate: 7.1, 22,100, 1,670, 55, 62, and 10,400.
       The MSW landfill leachates were mineralized, biologically-active liquids with
       relatively low concentrations of heavy metals and VOCs. On average, the
       leachates were slightly acidic (i.e., average pH of 6.7), which is expected
       because carbon dioxide and organic acids are the primary by-products of the
       first stage (i.e., the acid stage) of anaerobic  degradation of organic compounds
       in MSW landfills. The chemistry of these leachates changed with time as the
       organic compounds degraded (see for example, Table E-6.2 of Appendix E).  In
       general, the leachate characteristics for cells receiving waste were more
       indicative of the acid phase of degradation than the second stage (i.e., the
       methane fermentation phase) of anaerobic degradation.  For closed cells, the
       leachate pH typically increased with time and the BOD/COD ratio decreased
       with time, which is expected as the landfill is more fully in the methane
       fermentation phase of degradation.  Of the heavy metals and VOCs considered
       in Table 5-10, chromium, nickel, methylene chloride, and toluene were detected
       at the  highest concentrations in MSW leachates. Average concentrations of
       cadmium, benzene, 1,2-dichloroethane, trichloroethylene, and vinyl chloride in
       MSW landfills leachates exceeded  federal maximum contaminant levels (MCLs)
       (40 CFR 141.11, 141.61, and 141.62) for community drinking water systems.
       None  of the landfills had leachate with average chemical concentrations
       exceeding the MCLs for ethylbenzene, toluene, or xylenes.
                                      5-39

-------
                        Table 5-10. Summary of Landfill Leachate Chemistry Data.
Ol

-k
o
Waste Type
Number of Landfills
Parameter Units
pH pH units
Specific conductance nmhos/cm
IDS mg/l
COD mg/l
BOD5 mg/l
TOC mg/l
Alkalinity mg/l
Chloride mg/l
Sulfate mg/l
Calcium mg/l
Magnesium mg/l
Sodium mg/l
Arsenic ng/l
Cadmium ng/l
Chromium ng/l
Lead ng/l
Nickel ng/l
Benzene ng/l
1,1-Dichloroethane ng/l
1 ,2-Dichloroethane ng/l
cis-1 ,2-Dichloroethylene ng/l
trans-1 ,2-Dichloroethylene ng/l
Ethylbenzene ng/l
Methylene chloride ng/l
1,1,1-Trichloroethane ng/l
Trichloroethylene ng/l
Toluene ng/l
Vinyl chloride ng/l
Xylenes ng/l
MCLs












50
5
100


5

5
70
100
700

200
5
1,000
2
10,000
MSW
10 Pre-1 990
Average
6.62
6,588
5,487
3,878
2,281
1,509
2,295
801
274
444
153
532
19
< 8
68
36
56
< 17
88
< 33
< 64
< 51
40
435
< 68
< 56
491
< 49
117
Minimum
6.30
3,438
2,740
804
< 2
4
1,508
199
< 23
261
84
225
< 4
< 1
5
1
27
< 3
< 5
< 4
< 53
< 32
< 5
< 5
< 5
< 5
< 5
< 7
< 5
Maximum
7.20
8,983
8,640
8,267
4,510
2,852
3,278
2,263
1,943
610
279
1,115
78
< 17
320
90
98
< 36
294
< 100
< 75
< 100
87
1,303
100
114
959
< 100
277
No. of
Landfills
8
8
9
9
10
8
7
10
10
6
6
8
10
8
10
7
9
7
8
6
2
4
7
8
6
7
7
6
6
26 Post-1 990
Average
6.79
3,693
2,758
1,939
976
527
1,536
463
205
398
83
282
23
< 7
38
15
82
< 19
66
< 16
< 57
< 18
35
334
< 55
< 24
228
< 34
83
Minimum
5.90
597
480
< 10
< 2
24
203
5
< 7
66
10
3
< 2
< 1
3
1
10
< 2
< 2
< 1
< 1
< 1
< 1
< 1
< 1
< 1
< 1
< 3
< 5
Maximum
8.09
13,548
8,621
6,800
4,700
2,609
5,800
1,625
1,376
1,994
191
1,219
236
< 20
90
50
220
< 100
260
< 100
436
< 110
118
4,150
270
100
740
< 300
220
No. of
Landfills
22
22
21
22
18
21
22
25
24
22
21
23
21
22
21
22
20
21
22
20
13
16
22
22
20
19
22
20
20
                        Note: (1)" " = not analyzed; < = more than 50% of measurements reported as non-detect.

-------
                        Table 5-10. Summary of Landfill Leachate Chemistry Data (Continued).
Ol
Waste Type
Number of Landfills
Parameter Units
pH pH units
Specific conductance nmhos/cm
IDS mg/l
COD mg/l
BOD5 mg/l
TOC mg/l
Alkalinity mg/l
Chloride mg/l
Sulfate mg/l
Calcium mg/l
Magnesium mg/l
Sodium mg/l
Arsenic ng/l
Cadmium ng/l
Chromium ng/l
Lead ng/l
Nickel ng/l
Benzene ng/l
1,1-Dichloroethane ng/l
1 ,2-Dichloroethane ng/l
cis-1 ,2-Dichloroethylene ng/l
trans-1 ,2-Dichloroethylene ng/l
Ethylbenzene ng/l
Methylene chloride ng/l
1,1,1-Trichloroethane ng/l
Trichloroethylene ng/l
Toluene ng/l
Vinyl chloride ng/l
Xylenes ng/l
MCLs












50
5
100


5

5
70
100
700

200
5
1,000
2
10,000
HW
4
Average
8.17
22,096



1,623

7,758
2,985


5,243
26,710
< 119
124
109
738
< 131
123
< 382

< 79
< 133
161
< 99
< 76
< 173
< 1,475
14
Minimum
7.55
12,302



7

3,783
704


2,514
30
< 5
22
24
285
< 7
< 14
5

< 14
< 5
4
8
33
< 9
< 10
9
Maximum
9.36
39,598



3,239

1 1 ,734
5,267


7,972
79,912
< 233
226
249
1,190
370
< 371
< 1,124

< 143
< 512
< 447
< 347
< 146
616
< 4,405
18
No. of
Landfills
3
3



2

2
2


2
3
2
2
3
2
3
4
3

2
4
4
4
3
4
3
2
MSW ASH
7
Average
7.06
22,083
24,493
1,670
55
62
1,942
10,426
881
900
267
1,181
9
< 12
< 30
23
< 40
< 3
< 12
< 3
< 2
< 3
< 4
< 3
< 7
< 3
< 10
< 5
< 2
Minimum
6.54
10,732
6,067
304
15
39
99
2,940
85
96
113
684
5
< 2
< 1
3
< 24
< 1
< 1
< 1
< 1
< 1
< 2
< 1
< 1
< 1
< 1
< 1
< 1
Maximum
7.44
43,383
46,733
5,607
84
109
5,010
22,400
3,430
1,332
420
1,994
17
49
84
74
48
< 5
< 33
< 5
< 3
< 5
< 7
< 6
< 16
< 5
< 25
< 10
< 3
No. of
Landfills
5
4
6
4
4
3
4
4
5
3
2
5
6
6
6
6
4
3
3
3
2
3
3
3
3
3
3
3
2
                        Note: (1)" " = not analyzed; < = more than 50% of measurements reported as non-detect.

-------
                        Table 5-10. Summary of Landfill Leachate Chemistry Data (Continued).
Ol
Waste Type
Number of Landfills
Parameter Units
pH pH units
Specific conductance nmhos/cm
IDS mg/l
COD mg/l
BOD5 mg/l
TOC mg/l
Alkalinity mg/l
Chloride mg/l
Sulfate mg/l
Calcium mg/l
Magnesium mg/l
Sodium mg/l
Arsenic ng/l
Cadmium ng/l
Chromium ng/l
Lead ng/l
Nickel ng/l
Benzene ng/l
1,1-Dichloroethane ng/l
1 ,2-Dichloroethane ng/l
cis-1 ,2-Dichloroethylene ng/l
trans-1 ,2-Dichloroethylene ng/l
Ethylbenzene ng/l
Methylene chloride ng/l
1,1,1-Trichloroethane ng/l
Trichloroethylene ng/l
Toluene ng/l
Vinyl chloride ng/l
Xylenes ng/l
MCLs












50
5
100


5

5
70
100
700

200
5
1,000
2
10,000
COAL ASH
2
Average
7.70
884
723
11
< 3
6
190
21
383
190
22
46
36
< 7
< 16
< 19
38
< 4
< 4
< 4

< 1
< 3
< 4
< 4
< 4
< 2
< 7
< 4
Minimum
7.66
623
347
11
< 3
6
160
21
178
190
15
46
< 9
< 5
< 9
< 4
38
< 4
< 4
< 4

< 1
< 3
< 4
< 4
< 4
< 2
< 7
< 4
Maximum
7.74
1144
1098
11
< 3
6
220
21
587
190
30
46
62
< 9
22
< 34
38
< 4
< 4
< 4

< 1
< 3
< 4
< 4
< 4
< 2
< 7
< 4
No. of
Landfills
2
2
2
1
1
1
2
1
2
1
2
1
2
2
2
2
1
1
1
1

1
1
1
1
1
1
1
1
C&DW
2
Average
6.43
4815
3553
2414
1126
839
2450
681
255
292
202
304
15
< 3
39
7
< 56
17
92
3


66
417
51
< 11
613
8
210
Minimum
6.43
4815
2880
1139
1126
443
2450
671
48
203
202
284
15
< 1
39
3
< 56
17
92
3


66
417
51
< 11
613
8
210
Maximum
6.43
4815
4225
3688
1126
1235
2450
690
463
382
202
324
15
< 5
39
10
< 56
17
92
3


66
417
51
< 11
613
8
210
No. of
Landfills
1
1
2
2
1
2
1
2
2
2
1
2
1
2
1
2
1
1
1
1


1
1
1
1
1
1
1
                        Note: (1)" " = not analyzed; < = more than 50% of measurements reported as non-detect.

-------
The HW landfill leachates were more mineralized and had a higher organic
content than MSW leachates.  All of the HW leachates were alkaline, with pH
values ranging from 7.5 to 9.4. One possible explanation for the alkaline pH
values is the relatively common practice of solidifying hazardous waste with
pozzolonic additives prior to disposal. These relatively high pHs decrease the
mobility of metals. Even so, the average heavy metals concentrations were
generally several times to several orders of magnitude higher in HW leachates
as compared to MSW leachates. The HW leachates also had higher average
concentrations of all VOCs, except methylene chloride, toluene, and xylenes. Of
the heavy metals and VOCs considered in Table 5-10, arsenic, nickel, 1,2-
dichloroethane, and vinyl chloride were detected at the highest concentrations in
HW leachates. Average concentrations of arsenic, cadmium, chromium,
benzene, 1,2-dichloroethane,  trichloroethylene,  and vinyl chloride in HW landfill
leachates exceeded MCLs. None of the  landfills had leachate with average
chemical concentrations exceeding the MCLs for ethylbenzene, toluene, or
xylenes.
The chemistry of the ISW landfill leachates was highly variable due to the wide
variety of wastes disposed in  ISW landfills. The pH values for these leachates
ranged from 6.4 to 7.7.  The MSW ash leachates, the most mineralized of the
ISW landfill leachates, were even more mineralized than the MSW leachates in
this study, as evidenced by the high specific conductance, TDS, sulfate, and
chloride levels of the MSW ash leachates. Coal ash leachates were the least
mineralized. Both the MSW ash and coal ash leachates had low BOD values
that were several orders of magnitude less than the BOD values for MSW
leachate because  most of the organic materials originally in the MSW and coal
had been combusted. The average BOD value for C&DW leachate, however,
was within range of values reported for MSW leachate.  Heavy metals
concentrations in MSW ash and C&DW leachates were similar to those for MSW
leachates.  Metals concentrations in coal ash leachate were lower, generally at
the lower end of the concentration range for MSW leachates. As expected, the
MSW ash and coal ash leachates did not contain VOCs. However,  published
data show that MSW ash leachates can contain trace amounts of base neutral
extractables (BNAs), polychlorinated dibenzo-p-dioxins (PCDDs), and
polychlorinated dibenzo-furans (PCDFs).  The one C&DW landfill for which
organic chemistry  data are available produced leachate containing VOCs.
Average concentrations of cadmium in MSW ash and coal ash landfill leachates
and benzene, trichloroethylene, and vinyl chloride concentrations in C&DW
landfill leachates exceeded MCLs.
In general, the leachate chemistry data collected for the study fall within the
range of published data.
With the federal solid waste regulations promulgated in the 1980's and early
1990's (e.g., 1980 RCRA Subtitle C regulations for HW in 40 CFR 261 and Land
Disposal Restrictions in  40 CFR 268), it is expected that the quality of MSW and
HW landfill  leachates would have improved over time.  No statistically significant
differences  in concentrations of the considered  trace metals or VOCs in
leachates from older modern MSW landfills constructed prior to 1990 (pre-1990
                               5-43

-------
       landfills) and leachates from newer MSW landfills constructed after 1990 (post-
       1990 landfills) were observed at the 90% confidence level.  However, average
       VOC concentrations were generally lower in leachate from the post-1990
       landfills (Table 5-10). The statistical analysis findings were limited by the data.
       The limited number of landfills contributing to each dataset and the wide range
       of chemical concentrations led to large confidence intervals for each parameter
       in the datasets. To further evaluate the differences in leachate chemistry
       between older and newer MSW landfills, the data for the post-1990 MSW
       landfills were compared to published leachate chemistry data for 61 older MSW
       landfills (i.e., pre-1980 landfills in NUS (1988) and pre-1985 landfills in Gibbons
       et al. (1992)).  The distributions of the leachate chemistry data for the older
       MSW landfills were not known,  so the two data sets could not be compared
       statistically. However, the average concentrations of trace metals and VOCs in
       leachate from the newer landfills were almost always less than the average
       concentrations in leachate from the  older landfills.  Based on the above, it
       appears that the solid waste regulations have resulted in improved MSW landfill
       leachate quality. However, more data are needed to quantify this improvement.
       From the published information summarized in this report, the regulations may
       have also reduced the occurrence of certain chemicals.  For example,
       acetonitrile, cyanide, and naphthalene were detected more frequently in
       leachate from older landfills than in leachate from newer landfills.
   •   Published leachate chemistry data for 33 older HW landfills (i.e., pre-1984
       landfills in Bramlett et al. (1987), pre-1983 landfills in NUS (1988), and pre-1987
       landfills in Gibbons et al. (1992)) were compared to the data presented for  HW
       landfills in this  report (i.e., newer HW landfills).  The dataset for newer HW
       landfills is small; only leachate chemistry data for four landfills are available.
       The concentrations of chemicals in leachate from the newer landfills were found
       to be within the range of published values for the older landfills. The distribution
       of the leachate chemistry data for the older  HW landfills was not known, so the
       two datasets could not be compared statistically.  However, on average, most
       heavy metal concentrations and almost all VOC concentrations were lower in
       leachate from the newer landfills.  This reduction in  leachate strength is likely a
       result of the Subtitle C regulations and the Land Disposal Restrictions.

5.3 Lessons Learned from Waste Containment System Problems at Landfills

5.3.1  Scope of Work
The scope of work for this portion of the project involved an investigation into problems that
have occurred in waste containment systems for 69 modern landfill facilities and five
modern surface impoundment facilities located throughout the U.S. The investigation
focused on landfills, and only landfill-related problems are discussed in this section. The
purpose of the investigation is twofold: (i) to better understand the nature, frequency,  and
significance of identified problems; and (ii) to  develop recommendations to reduce the
future occurrence of problems.
                                      5-44

-------
The scope of work specifically excluded consideration of problems in older waste
containment systems not designed and constructed to current standards and practices.
These problems include, for example, the LCRS and cover system internal drainage
layer failures described by Bass (1986),  Ghassemi et al. (1986), and Kmet et al. (1988).
The scope of work also excluded foundation stability problems at older landfills, such as
the problems described by Oweis (1985), Dvirnoff and Munion (1986), Richardson and
Reynolds (1991), Kenter et al. (1997), Stark and Evans (1997),  and Schmucker and
Hendron (1997). Problems at older facilities are often not relevant to current standards
and practices.

5.3.2  Description of Database
The 80 landfill problems identified during the investigation for this report are categorized
on the basis of two criteria. The first criterion addresses the component or attribute of
the landfill liner  system or cover system  affected by the problem. The specific landfill
components and attributes considered in this study are: (i) liner construction; (ii) liner
degradation; (iii) LCRS or LDS construction; (iv) LCRS or  LDS degradation; (v) LCRS
or LDS malfunction; (vi) LCRS or LDS operation; (vii) liner system stability; (viii) liner
system displacement;  (ix) cover system  as built; (x) cover system degradation;  (xi)
cover system stability; and (xii) cover system displacement.  Specific problems that may
affect these components and attributes are  listed in Tables F-4.1 to F-4.3 in Appendix F.
Other components or attributes not specifically associated with  landfill integrity were not
considered in the investigation.  These include landfill daily and intermediate cover
components (except for cracking of the soil  intermediate cover during the Northridge
earthquake), leachate transmission and  treatment components  beyond the leachate
collection sumps or manholes, and landfill gas extraction and management
components.

The second criterion used to categorize  the problem addresses the principal human
factor contributing to the problem. The principal human factors considered are: (i)
design; (ii) construction; and (iii) operation.  While a principal human factor has been
assigned to each problem, it should be recognized that most problems have complex
causes and several contributing factors.  Hereafter, the problem classifications  are
shown as "component or attribute criterion'Tprincipal human factor criterion" (e.g., liner
system stability/design).

Detailed case histories of the problems are presented in Attachment F-A of Appendix F.
The information  sources for the problems are listed in Table F-2.1 of Appendix F.
Summaries of the identified problems are presented in Table F-2.2 and are repeated in
Table 5-11 below.  The problems are grouped according to the above two criteria  in Table
F-2.3 of Appendix F.
                                      5-45

-------
Table 5-11. Summary of Identified Problems at Landfills.
Problem
Classification*1'
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
operation
liner construction/
operation
liner construction/
design
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner construction/
construction
liner degradation/design
landfill liner
degradation/operation
liner degradation/
design
liner degradation/
construction
liner degradation/
construction
liner degradation/
construction
liner degradation/
design
Facility Designation/
Appendix Section
L-1/F-A.2.1
L-3/F-A.2.2
L-5/F-A.2.3
L-6/F-A.2.4
L-7/F-A.2.5
L-8/F-A.2.6
L-9/F-A.2.7
L-11/F-A.2.8
L-11/F-A.2.9
L-15/F-A.2.10
L-17/F-A.2.11
L-19/F-A.2.12
L-19/F-A.2.13
L-29/F-A.2.14
L-2/F-A.3.1
L-4/F-A.3.2
L-12/F-A.3.3
L-14/F-A.3.4
L-20/F-A.3.5
L-43/F-A.3.6
L-44/F-A.3.7
Problem Summary
leakage through holes in HOPE GM primary liner
leakage through holes in HOPE GM liners
leakage through holes in HOPE GM primary liner
leakage through holes in HOPE GM primary liner
leakage though HOPE GM/CCL composite primary
liner at pipe penetration
landfill gas migrated beyond liner system and into
vadose zone resulting in groundwater contamination
leakage though HOPE GM primary liner at pipe
penetration
construction debris in CCL with initially smooth
surface protruded from CCL after CCL was left
exposed and subsequently eroded
leakage though HOPE GM primary liner at pipe
penetration
sand bag under installed GM liner approved by
CQA consultant
leakage through holes in HOPE GM primary liner
wind uplifted and tore HOPE GM liner during
construction
severe wrinkling of HOPE GM due to thermal
expansion during construction
large folded wrinkles in HOPE GM primary liner at
two exhumed leachate sumps
desiccation cracking of CCL in exposed HOPE
GM/CCL composite liner
HOPE GM/CCL composite liner damaged by waste
fire
leachate extraction well installed in landfill appeared to
puncture GM primary liner
HOPE GM liner damaged by fire believed to be started
by lightning strike
saturation of GCL beneath GM liner when rainwater
ponded on tack-seamed patch over GM hole
water ponded between HOPE GM and CCL
components of composite secondary liner and was
contaminated from a source other than the landfill
landfill gas well punctured GM component of
composite liner and extended into CCL
                                     5-46

-------
Table 5-11. Summary of Identified Problems at Landfills (Continued).
Problem Classification
LCRS or LDS
construction/construction
LCRS or LDS
construction/construction
LCRS or LDS
construction/construction
LCRS or LDS
construction/construction
LCRS or LDS
construction/construction
LCRS or LDS
construction/construction
LCRS or LDS degradation/
design
LCRS or LDS degradation/
design
LCRS or LDS degradation/
construction
LCRS or LDS degradation/
construction
LCRS or LDS degradation/
construction
LCRS or LDS malfunction/
operation
LCRS or LDS malfunction/
design
LCRS or LDS malfunction/
design
LCRS or LDS malfunction/
operation
LCRS or LDS operation/
operation
LCRS or LDS operation/
operation
LCRS or LDS operation/
operation
LCRS or LDS operation/
design
Facility Designation/
Appendix Section
L-10/F-A.4.1
L-15/F-A.4.2
L-16/F-A.4.3
L-28/F-A.4.4
L-32/F-A.4.5
L-33/F-A.4.6
L-9/F-A.5.1
L-11/F-A.5.2
L-13/F-A.5.3
L-18/F-A.5.4
L-30/F-A.5.5
L-12/F-A.6.1
L-22/F-A.6.2
L-36/F-A.6.3
L-37/F-A.6.4
L-5/F-A.7.1
L-23/F-A.7.2
L-34/F-A.7.3
L-35/F-A.7.4
Problem Summary
rainwater entered LDS through anchor trench
sand bags in LCRS drainage layer and debris in
LCRS pipe trench approved by CQA consultant
rainwater entered LDS through anchor trench
excessive needle fragments in manufactured
needlepunched nonwoven GT
HOPE LCRS pipe separated at joints
HOPE LCRS pipe separated at joints
erosion of sand LCRS drainage layer on liner
system side slopes
erosion of sand protection layer on liner system
side slopes
polypropylene continuous filament nonwoven GT
filter degraded due to outdoor exposure
polypropylene staple-fiber needlepunched
nonwoven GT filter degraded due to outdoor
exposure
HOPE LCRS pipe crushed during construction
LCRS pipes were not regularly cleaned and
became clogged, and LCRS drainage layer may
be partially clogged
waste fines clogged needlepunched nonwoven
GT filter wrapped around perforated LCRS pipes
waste fines clogged needlepunched nonwoven
GT filter around LCRS pipe bedding gravel
leachate seeped out landfill side slopes in the
vicinity of chipped tire layers
overestimation of LDS flow quantities due to
problems (e.g., clogging) with automated LDS
flow measuring and removal equipment
valves on LCRS pipes were not opened and
leachate could not drain, and waste and leachate
flowed over a berm into a new unapproved cell
LCRS leachate pump moved air and liquid
causing pump airlock and underestimation of
leachate quantities
LCRS leachate pumps and flowmeters continually
clogged and LDS leachate pumps turned on too
frequently and burned out prematurely
                                    5-47

-------
Table 5-11. Summary of Identified Problems at Landfills (Continued).
Problem Classification
liner system stability/
design
liner system stability/
operation
liner system stability/
design
liner system stability/
design
liner system stability/
design
liner system stability/
design
liner system stability/
design
liner system stability/
design
liner system stability/
design
liner system stability/
operation
liner system stability/
operation
liner system stability/
design
liner system
displacement/design
liner system
displacement/design
liner system
displacement/design
liner system
displacement/design
cover system construction/
construction
cover system construction/
construction
cover system degradation/
design
cover system degradation/
design
Facility Designation/
Appendix Section
L-21/F-A.8.1
L-24/F-A.8.2
L-25/F-A.8.3
L-26/F-A.8.4
L-27/F-A.8.5
L-38/F-A.8.6
L-39/F-A.8.7
L-40/F-A.8.8
L-41/F-A.8.9
L-42/F-A.8.10
L-45/F-A.8.11
L-46/F-A.8.12
L-9/F-A.9.1
L-11/F-A.9.2
L-25/F-A.9.3
L-31/F-A.9.4
C-2/F-A.10.1
C-16/F-A.10.2
C-1/F-A.11.1
C-12/F-A.11.2
Problem Summary
sliding along PVC GM/CCL interface during
construction
sliding along GN/GCL (HOPE GM side) and
GCL(bentonite side)/CCL interfaces during
operation
sliding along HOPE GM/ polyester needlepunched
nonwoven GT and HOPE GM/CCL interfaces
during operation
two tears in HOPE GM liner and cracks in soil
intermediate cover from Northridge earthquake
extensive cracks in soil intermediate cover and
further tearing of GT cushion from Northridge
earthquake
sliding along needlepunched nonwoven GT/HDPE
GM primary liner interface after rainfall
sliding along needlepunched nonwoven GT/HDPE
GM liner interface after rainfall
sliding along gravel/HDPE GM liner interface after
rainfall
sliding along very flexible GM liner/needlepunched
nonwoven GT interface after rainfall
sliding along needlepunched nonwoven GT/PVC
GM liner interface after a thaw
sliding along needlepunched nonwoven GT/HDPE
GM liner interface after erosion of soil anchoring
geosynthetics
sliding along GN/HDPE GM primary liner interface
during construction
uplift of GM by landfill gas after erosion of
overlying sand LCRS drainage layer
uplift of geosynthetics by landfill gas after erosion
of overlying sand protection layer
uplift of composite liner by surface-water
infiltration during construction
uplift of composite liner by surface-water
infiltration during construction
portion of topsoil from off-site source was
contaminated with chemicals
high failure rate of HOPE GM seam samples
during destructive testing
failure of geosynthetic erosion mat-lined
downchute on 3H:1 V side slope
erosion of topsoil layer on 60 m long, 3H:1Vside
slope
                                    5-48

-------
Table 5-11.  Summary of Identified Problems at Landfills (Continued).
Problem Classification
cover system stability/
construction
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
construction
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system stability/
design
cover system
displacement/design
cover system
displacement/construction
Facility Designation/
Appendix Section
C-3/F-A.12.1
C-4/F-A.12.2
C-5/F-A.12.3
C-6/F-A.12.4
C-7/F-A.12.5
C-8/F-A.12.6
C-9/F-A.12.7
C-10/F-A.12.8
C-11/F-A.12.9
C-13/F-A.12.10
C-14/F-A.12.11
C-17/F-A.12.12
C-18/F-A.12.13
C-19/F-A.12.14
C-20/F-A.12.15
C-21/F-A.12.16
C-22/F-A.12.17
C-23/F-A.12.18
C-12/F-A.13.1
C-15/F-A.13.2
Problem Summary
sliding along nonwoven GT/GM interface during
construction
sliding along topsoil/GCL interface after rainfall
sliding along sand/woven GT interface after
rainfall
sliding along sand/GM interface after rainfall
sliding along gap-graded sand/GM interface after
rainfall
sliding along gravel/GT interface during
construction
sliding along sand/calendered nonwoven GT
interface after rainfall
sliding along sand/GM interface after rainfall
sliding along topsoil/nonwoven GT interface
during construction
sliding along PVC GM/CCL interface after a thaw
sliding along geogrid/HDPE GM interface during
construction
sliding along sand/CCL interface during rainfall
sliding along sand/CCL interface immediately after
rainfall
sliding along sand/CCL interface after rainfall
sliding along sand/CCL interface after rainfall
minor cracks in soil intermediate cover from
Northridge Earthquake
215-m long cracks in soil intermediate cover from
Northridge Earthquake
minor cracks in soil intermediate cover from
Northridge Earthquake
cover system settlement caused tearing of HOPE
GM boots around gas well penetrations of GM
barrier
localized cover system settlement during
construction stretched, but did not damage, PVC
GM barrier and opened GCL joints
Note: (1) Each problem classification has two
experienced the problem; and (ii) the principal
parts: (i) the component or attribute of the landfill that
human factor contributing to the problem.
                                          5-49

-------
5.3.3  Study Findings
Based on the results of the investigation into waste containment system problems
presented in Appendix F, the following conclusions are drawn:

   •   This investigation identified 69 modern landfill facilities that had experienced a
       total of 80 waste containment system problems. This number of facilities is
       relatively small in comparison to the over 1,000 modern landfills nationwide.
       The search for problem facilities for this study was not exhaustive, and it is
       certain that there are other facilities with problems similar to those described in
       this report.
   •   About 72% of the landfill problems were liner system related and 28% were
       cover system related.  The ratio of liner system problems to cover system
       problems is probably exaggerated by the fact that a number of the facilities
       surveyed were active and did not have a cover system.
   •   Based on the waste containment system component or attribute criterion, the
       identified landfill  problems were classified as follows, in order of decreasing
       frequency:
            o cover system stability: 23%;
            o  liner construction:  18%;
            o  liner system stability: 15%;
            o  liner degradation:  9%;
            o  LCRS or LDS construction:  8%;
            o  LCRS or LDS degradation:  6%;
            o  LCRS or LDS malfunction: 5%;
            o  LCRS or LDS operation: 5%;
            o  liner system displacement: 5%;
            o cover system construction: 2%;
            o cover system degradation: 2%; and
            o cover system displacement: 2%.
   •   Using this criterion, these problems can also be grouped into the following
       general categories (Figure 5-6):
            o  liner system or cover system slope stability or displacement: 45%
            o  liner, LCRS or LDS, or cover system construction: 28%;
            o  liner, LCRS or LDS, or cover system degradation:  17%; and
            o  LCRS or LDS malfunction or operation:  10%.
   •   Based on the principal human factor contributing to the problem criterion, the
       identified landfill  problems were classified as follows  (Figure 5-7):
            o design: 51%;
            o construction: 35%; and
            o operation: 14%.
                                     5-50

-------
                              Degradation
                                 17%
 Construction
     28%
LCRS or LDS
 Malfunction/
  Operation
    10%
                                               Stability or
                                             Displacement
                                                 45%

Figure 5-6.  General distribution of problems by waste containment system
           component or attribute criterion.
         Construction
            35%
Operation
  14%
                                            Design
                                             51%
Figure 5-7.  Distribution of problems by principal human factor contributing to the
           problem criterion.
                                   5-51

-------
Problems that occurred at two or more landfills and the number of landfills at
which they occurred are as follows, in order of presentation in Chapter F-3 of
Appendix F:
      o  leakage through holes (construction- or operation-related) in an HOPE
         GM primary liner (5 landfills);
      o  leakage through an HOPE GM primary liner or HOPE GM/CCL
         composite primary liner at the LCRS pipe penetration of the liner (3
         landfills);
      o  severe wrinkling of an HOPE GM liner during construction (2 landfills);
      o  liner damage by fire (2 landfills);
      o  liner damage during well installation (2 landfills);
      o  rainwater entered the LDS through the anchor trench (2 landfills);
      o  HOPE LCRS pipe was separated at joints (2 landfills);
      o  erosion of the sand layer on liner system side slopes (2 landfills);
      o  degradation of polypropylene nonwoven GT filters due to outdoor
         exposure (2 landfills);
      o  waste fines clogged the needlepunched nonwoven GT filter in the
         LCRS piping system  (2 landfills);
      o  clogging  and other problems with the leachate pump or flow rate
         measuring system (3 landfills);
      o  liner system slope failure due to static loading (10 landfills);
      o  liner system damage due to earthquakes (2 landfills);
      o  uplift of liner system geosynthetics by landfill gas after erosion of the
         overlying sand layer (2 landfills);
      o  uplift of composite liner by surface-water infiltration during
         construction (2  landfills);
      o  cover system slope failure during construction (4 landfills);
      o  cover system slope failure after rainfall or a thaw (11 landfills); and
      o  soil cover damage due to earthquakes (3 landfills).
For problems that occurred at three or more landfills, the principal human factor
contributing to the problem criterion, detection of the problem, causes of the
problem, and remedy of the problem are described below:
      o  Leakage through holes in an HOPE GM primary liner occurred at five
         landfills.  In each case, the holes were attributed to construction or, at
         one landfill, possibly operation factors. At two of the landfills, leakage
         was first  detected during electrical leak location surveys performed as
         part of CQA and by the relatively high LDS flow rates that occurred
         after rainwater ponded in a landfill. At the remaining three landfills,
         leakage was first detected during operation by the relatively high LDS
         flow rates and the color of and chemical constituents in the LDS liquid.
         The cause of the leakage was attributed to construction-related holes
         in the GM. However, at one landfill, where waste was placed directly
         on liner system geosynthetics (i.e., there is no soil protection layer),
         the GM may have been damaged during waste placement.  The
         leakage problem was resolved at four landfills by repairing the GM
         holes; at the remaining landfill, the problem, clearly identified only
                                5-52

-------
   after the cell had been covered with waste, was partially remedied by
   lowering the "pump on" liquid level in the LCRS sump.
o  Leakage through an HOPE GM primary liner or HOPE GM/CCL
   composite primary liner at the LCRS pipe penetration of the liner
   occurred at three landfills. This leakage was attributed to construction
   factors at two of the landfills and operation factors at the third landfill.
   At two of the landfills, leakage at the pipe penetration was detected
   during construction after rainwater ponded over the penetration and
   LDS flow rates increased. The cause of the leakage was construction
   defects in the pipe penetration; it is difficult to construct a defect-free
   pipe penetration, even when extra measures are taken to enhance the
   integrity of the connection.  At the remaining landfill, leakage was
   detected during operation when the average LDS flow rate increased
   significantly. For this landfill, the pipe penetration was damaged
   during operation when a rubber-tired loader trafficked over it.  The
   pipe penetrations were repaired; however, at one landfill where the
   problem was detected during construction, the repairs did not
   significantly decrease LDS flow rates; thus there must have existed a
   penetration defect that was not located.
o  Clogging and other problems with the leachate pumps or flow rate
   measuring system occurred at three landfills. These problems were
   attributed to design factors at one of the landfills and operation factors
   at the other two landfills. The problems, which were identified during
   routine operations, included: (i) clogging of the air lines and failure of
   the  compressor for the control system;  (ii) drift of the leachate level
   measurement system; (iii) drift of the "pump on" time setting; (iv) burn
   out  of pumps due to control system problems;  (v) clogging of pumps;
   (vi)  clogging of mechanical flowmeters; (vii) damage to electrical
   equipment by electrical storms; (viii) check valve failure; and (ix)
   inaccurate measurement of LCRS or LDS flow rates due to the above
   equipment problems. These problems  appear to have been primarily
   caused by: (i) inadequate overall mechanical system design; (ii) using
   equipment that was less reliable than was needed; (iii) using
   equipment that was not compatible with the landfill leachate; and (iv)
   not  performing equipment maintenance often enough. These
   problems were primarily remedied by equipment maintenance, repair,
   and replacement.
o  Liner system slope failure due to static  loading occurred at ten
   landfills. These problems were attributed to design factors at seven of
   the  landfills and operation factors at the remaining three landfills.
   Slope failure occurred during construction at two of the landfills and
   during operation at the remaining eight landfills.  The problem was
   detected by visual observation of mass movement of the liner system,
   cracking of soil layers near the slope crest, and tearing, tensioning, or
   wrinkling of geosynthetics.  The primary causes of failure were: (i)
   using unconservative presumed values for the critical interface shear
                         5-53

-------
   strength; (ii) not evaluating the critical condition for slope stability
   (e.g., liner system with waste at intermediate grades); (iii) not
   accounting for, or underestimating, seepage pressures; (iv) not
   accounting for moisture at the GM/CCL interface (which weakens the
   interface) due to spraying of the CCL and thermal effects; and (v) not
   maintaining the  drainage layer outlets free of snow and ice, which can
   lead to increased seepage pressures.  The slope failures were
   remedied by reconstructing the damaged liner systems, sometimes
   with different materials, and developing new construction procedures
   to reduce moisture at the  GM/CCL interface.
o  Cover system slope failure during construction occurred at four
   landfills.  These problems were attributed to design factors at two of
   the landfills and construction factors  at the remaining two landfills.
   Slope failure was detected by visual  observation of mass movement
   of the cover system, cracking  of soil  layers near the slope crest,  and
   wrinkling of geosynthetics at the toe  of the cover system slope. The
   primary causes  of failure were: (i) placing soil over the side slope
   geosynthetics from the top of the slope downward, rather from the toe
   of the slope upward; (ii) not considering the effects of variation in the
   tested geosynthetics, accuracy of test methods, and test conditions on
   the interface shear strength to use in design; and (iii) using
   unconservative  presumed values for the critical interface shear
   strength. The problems were remedied by reconstructing the cover
   systems using different cover system materials that result in higher
   interface shear strengths and  placing soil over side slope
   geosynthetics from the toe of the slope upward.
o  Cover system slope failure after rainfall or a thaw occurred at eleven
   landfills.  At all of these landfills, the failures were attributed to design
   factors.  Slope failure occurred during the post-closure period and was
   detected by visual observation of mass movement of the cover
   system, cracking of soil layers near the slope crest, and wrinkling of
   geosynthetics at the toe of the cover system slope. The primary
   causes of failure appeared to  be: (i) not accounting for, or
   underestimating, seepage pressures; (ii) clogging of the drainage
   system, which can lead to increased seepage pressures; and (iii) not
   accounting for moisture at the GM/CCL interface (which weakens the
   interface) due to rain falling on the CCL surface during construction
   and freeze-thaw effects. In general,  the problems were remedied by
   reconstructing the cover systems with new drainage systems or
   different materials.
o  So/7 cover damage due to earthquakes occurred at three landfills.
   These problems all occurred during operation and were attributed to
   design factors.  The damage,  which was detected by visual
   inspection, consisted of surficial cracking of soil intermediate cover
   occurring primarily near locations with contrast in seismic response
   characteristics (e.g., top of waste by  canyon walls). The damage was
                         5-54

-------
         expected and dealt with as an operation issue through post-
         earthquake inspection and repair (i.e., regrading and revegetating the
         cracked soil layers).
Almost all of the problems identified in this investigation were detected shortly
after they occurred by visual observation or evaluation of monitoring data.
Of the problems in this study for which the remedy was identified, six problems
were not completely repaired because their environmental impacts were not
expected to be significant and because:  (i) the source of the problem could not
be identified; (ii) the problem was not worsening; (iii) repair of liner systems or
LCRS pipes after waste placement would be extremely difficult and expensive;
and/or (iv) additional liner system damage could occur in any attempt to
excavate the waste and repair the liner system.
The problems only resulted in an  identified environmental impact to groundwater
or surface-water quality by leachate or landfill gas at one facility,  landfill L-8. At
this MSW landfill, groundwater impact by VOCs was attributed to gas migration
through a relatively permeable soil layer that secured the edge of the GM liner
and extended from the crest of the liner system side slope to beyond the liner
system. The problem was resolved by installing additional gas extraction wells
in the landfill. Without the measures taken to correct the problems at some of
the other facilities, however, adverse environmental impacts could have
eventually occurred  at these facilities.
The main impacts of the problems identified in this investigation are interruption
of waste containment system construction and operation,  increased
maintenance, and increased costs.
The identified problems that most often disrupted construction and were
required to be repaired before construction proceeded were related to:
      o holes in GM liners and  at pipe penetrations of liners;
      o large wrinkles in HOPE GM liners;
      o degradation of exposed geosynthetics;
      o uplift of constructed liners by groundwater or infiltrating surface water;
         and
      o erosion of unprotected  soil layers (CCLs, sand drainage layers, soil
         protection layers).
Problems that disrupt operation are generally more severe in terms of required
repairs than those that interfere with construction and may require waste
relocation.  Consequently, problems that disrupt operation generally require
more time to remedy than problems that are identified and repaired during
construction.  Problems that involve major breaches of liner systems or cover
systems (e.g., failure of landfill liner system slopes) may require months to
repair. The identified problems that most often disrupted operation and were
required to be repaired before operation proceeded were related to:
      o holes in GM liners and  at pipe penetrations of liners;
      o failure of one or more components of a liner system or cover system
         on landfill slopes; and
      o clogging of GTs in LCRSs.
                               5-55

-------
   •   Problems that require maintenance may be more severe in terms of required
       repairs than those that interfere with construction, but are generally less severe
       than those that interfere with operation.  In addition, problems that require
       maintenance are more likely to be reoccurring.  The identified problems that
       most often required maintenance were related to:
             o  erosion of soil layers (sand drainage layers, soil protection layers);
             o  repair of LCRS or LDS flow rate measuring and  removal systems; and
             o  cracking of soil intermediate cover after earthquakes.
   •   The costs of remedying the problems can be significant. For the identified
       problems, the costs at the times the remedies were implemented ranged from
       less than $10,000 for repairs of GM holes identified by leak location surveys
       during construction to more than several million dollars for repair of a liner
       system slope failure  that occurred during cell operation.  In general, problems
       that impacted operation were more expensive than those that impacted
       construction or maintenance.  However,  certain problems that impact
       maintenance, such as erosion of soil layers, may ultimately be more costly than
       other problems if these problems reoccur.
   •   Even though there was only evidence of environmental impact at one of the
       waste containment systems in this study, the landfill industry should do more to
       avoid future problems in order to: (i) reduce the potential risk of future
       environmental impact; (ii) reduce the potential health and safety risk to facility
       workers, visitors, and neighbors; (iii) increase public confidence  in the
       performance of waste containment systems; (iv) decrease potential impacts to
       construction, operation, and maintenance; and (v) reduce costs  associated with
       the investigation and repair of problems.
   •   Importantly, all of the design,  construction, and operation problems identified in
       this investigation can be prevented using available design approaches,
       construction materials and procedures, and operation practices. It is the
       responsibility of all professionals involved in the design,  construction, operation,
       and closure of waste containment systems to improve the  practice of waste
       containment system  engineering.  Owners must be prepared to  adequately fund
       the levels of design and CQA activity necessary to properly design and construct
       waste containment systems.  Design engineers must improve their practice to
       avoid the types of problems identified herein. Earthwork contractors,
       geosynthetics installers, and landfill operators all must be properly trained,
       supervised, and committed to the "quality goals" necessary to eliminate
       problems.

5.3.4 Recommendations
Based on an evaluation of the identified waste containment system problems, the
following general and specific design, construction, and  operation recommendations are
made to reduce the incidence of these problems. These measures are  not new; they
have been used extensively for other engineered structures, such as dams. The
measures include widely available design approaches, construction procedures,  and
operation practices.  Many recommendations for landfill  liner systems also apply to
                                      5-56

-------
cover systems, and vice versa. Because of this, the recommendations are grouped to
apply to the following broad categories:

   •   general;
   •   liners and barriers;
   •   drainage systems;
   •   surface layers and protection layers;
   •   liner system and cover system stability; and
   •   liner system and cover system displacements.

Recommendations for each of these categories are presented below.

General recommendations intended to reduce the occurrence of problems include:
   •   information dissemination (e.g., this report);
   •   training of design engineers to better understand waste containment system
       design fundamentals and to avoid the types of design problems described in this
       report;
   •   training of design engineers to be better prepared to develop waste containment
       system specifications and CQA plans that are complete and precise, that include
       the construction-related assumptions made during design, and that require
       construction and CQA procedures to identify and prevent the kinds of
       construction problems identified in this report;
   •   training of CQA personnel in standard CQA procedures to avoid the types of
       construction problems identified in this report; for engineering technicians, this
       training can be demonstrated through the National Institute for Training in
       Engineering Technologies (NICET)  certification program;
   •   training of contractors  to avoid the types of construction problems identified in
       this report;
   •   development of better construction materials, techniques, and quality
       control/quality assurance procedures to prevent the kinds of construction
       problems identified in this report;
   •   development of better operations manuals to describe and provide controls for
       procedures to be followed by landfill operations personnel;
   •   training of facility operators to better avoid the types of operation problems
       identified in this report;
   •   training of facility operators to better detect and quickly report problems
       occurring during operation; and
   •   performing  periodic independent audits to verify that the specified operation
       procedures are being practiced.

Specific recommendations are presented below in Table 5-12.
                                      5-57

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems.
                      Recommendations
Category
Design
Construction
Operation
Liners and
Barriers
  Resin used to manufacture
  HOPE GM should be resistant
  to stress cracking.  This is
  currently evaluated using the
  notched constant tensile load
  test (ASTM D 5397). This test
  should be required in project
  specifications.
  Project specifications should
  require that both the inner and
  outer tracks of GM fusion seam
  samples taken for destructive
  testing meet the project seam
  requirements.  Failure of one
  track is generally indicative of
  overall seaming problems and
  can result in increased stress
  concentrations in the adjacent
  track.  In addition, testing both
  tracks may allow seaming
  problems to be identified and
  corrected quicker.
  The potential for GM damage
  during placement of a soil layer
  over a GM can be reduced by
  protecting the GM. Measures
  for GM protection should be
  incorporated into the design
  and specifications. Measures
  include placing a protection
  layer (e.g., thick GT cushion or
  GC drainage layer) over the
  GM, using a greater initial lift
  thickness of soil above the GM,
  and using construction
  equipment with low ground
  pressure to place soils over the
  GM. The protection measures
  should be selected based  on
  the characteristics of the soil to
  be placed (e.g., angularity,
  maximum particle  size), the
  thickness of the soil layer, the
  type of equipment placing  the
  soil, and whether CQA will be
  performed during soil
  placement. If the soil layer is
  placed during operation without
  CQA, extra GM protection is
  necessary.
  GMs located in areas
  subjected to high static and
  dynamic stresses from
  construction equipment, such
  as beneath temporary access
  roads, require an even higher
  level of protection  than GMs
  Construction equipment should
  be inspected for fuel and oil
  leaks, and those leaks should
  be repaired prior to using the
  equipment in liner construction
  to avoid liner and IDS
  contamination.
  Liners and barriers should be
  constructed in manageable
  increments that ensure
  protection of the liner and
  barrier materials under
  seasonal weather changes.
  CCLs should not be
  constructed with materials
  containing construction debris
  or large particles, even if prior
  to GM installation the CCL has
  a smooth surface and meets
  the hydraulic conductivity
  criterion. The debris may
  adversely impact the hydraulic
  conductivity of the CCL and/or
  damage an overlying GM.
  CCLs should not be left
  unprotected for an extended
  period of time.  They can
  desiccate and crack due to
  evaporation of water in the
  CCL, degrade when exposed
  to freezing and thawing
  actions, and be eroded by wind
  and water.
  Prior to deploying a GM, all
  extraneous objects (e.g., tools,
  sand bags) should be removed
  from the surface on which the
  GM is to be placed to avoid
  GM damage and, for
  composite liners, promote
  good contact between the GM
  and underlying GCL or GCL.
  HOPE GMs should be installed
  so that they are essentially
  stress-free at their lowest
  expected temperatures to
  avoid GM straining and,
  potentially, rupture.
  GMs should be covered with
  thermal insulation layers at
  very low temperatures since
  GM strain at break decreases
  with decreasing temperature.
  The  leading edge of an
  uncovered GM should be
  secured to prevent wind from
  flowing beneath the GM and
  uplifting it. This is typically
• Landfill operations manuals
  should include limitations on
  the types of equipment that
  may traffic over the liner
  system before the first lift of
  waste is  placed to prevent liner
  damage.
• Landfill operations personnel
  should be aware  of sensitive
  areas of  a liner system, such
  as at pipe penetrations or
  sumps, and should  protect
  these areas from damage.
  Sensitive areas can be
  identified with cones, flags, or
  other markers. They can also
  be isolated from traffic by
  berms, bollards, or  other
  means.
• Landfills  should be  operated to
  minimize the potential for
  waste fires. Measures to be
  taken could include not
  depositing loads  of hot waste
  in a landfill  and covering waste
  with a soil cover to decrease
  waste access to oxygen.
• Care should be taken to not
  damage  the liner system
  components when drilling into
  landfilled waste.  Settlement of
  the waste surface must be
  taken into account when
  selecting the depth  of drilling,
  and boreholes should not
  extend close (e.g., within 1 m)
  to the primary liner. Also, the
  limits of waste containment
  systems  should be  identified
  with markers or other means to
  reduce the  potential for liner
  system or cover system
  damage  by drilling or other
  invasive  activities.
                                                       5-58

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems (Continued).
                      Recommendations
Category
Design
Construction
Operation
Liners and
Barriers
(continued)
  not subjected to high stresses.
  These protection measures
  should be incorporated into the
  design and specifications.
  GM should be protected during
  waste placement over the GM.
  Protection measures should be
  incorporated into the design
  and specifications. Measures
  include installing a protection
  layer  (e.g., thick GT cushion,
  GC drainage layer, or soil
  layer) over the GM, using
  spotters to direct equipment
  operators during waste
  placement over the GM, and
  placing only select waste over
  the GM.  Protection measures
  should be selected with
  consideration of waste
  characteristics and the
  equipment placing the waste.
  Sensitive areas  of a liner
  system (e.g., at  pipe
  penetrations) should be
  designed to be untraffickable
  by berms, bollards, or other
  means to decrease the
  potential for damage to these
  areas.
  It is difficult to construct pipe
  penetrations of liners to be
  defect free. Until new methods
  for constructing  better
  connections between GMs and
  ancillary structures have been
  developed and tested, designs
  without pipe penetrations (i.e.,
  designs with internal sumps)
  should be preferred.
  Internal sumps typically have
  sustained leachate heads at
  greater depths than other
  locations within the landfill and
  have  seamed corners, which
  may contain holes. To
  decrease the rate of leakage
  through GM holes at sumps,
  the sump design should
  include additional liner
  components, such as a GCL,
  beneath the GM liner in the
  sump area, even if the GM is
  already underlain by a CCL. A
  design with a prefabricated GM
  sump may also be considered.
  The potential for landfill gas to
  migrate over the geosynthetics
 accomplished by seaming
 adjacent panels of GM shortly
 after deployment and placing a
 row of sandbags along the
 edge of the GM.
 If sand bags are used to
 secure GM panels, the installer
 should ensure that the sand
 bags, and all other extraneous
 objects, are not trapped
 beneath the GM after seaming
 to avoid GM damage and, for
 composite liners, promote
 good contact between the GM
 and underlying CCL or GCL.
 For HOPE GMs, fusion seams
 are preferred over extrusion
 seams because fusion seams
 have higher seam integrity and
 lower stress concentrations at
 seams.  Extrusion seams
 should be minimized in the
 field by using prefabricated
 pipe boots, careful GM
 installation, etc.
 HOPE GMs must be cleaned
 along the seam path before the
 seam is constructed since dirt
 in the seam adversely impacts
 seam integrity. To minimize
 the potential for dirt to collect in
 the seam path, GM should be
 seamed shortly after
 deployment. A temporary
 protective plastic film may also
 be placed on the GM edges at
 the factory and removed from
 the GM just prior to seaming.
 In general, holes in HOPE GM
 seams should not be repaired
 by reseaming.  This reheating
 of seams can embrittle the
 HOPE at the repair and make it
 more susceptible to stress
 cracking.
 To the extent practicable, holes
 in GMs liners installed over
 GCLs should be repaired as
 soon as possible to avoid
 swelling of the GCL in case of
 hydration. GCL swelling
 results in a decrease in GCL
 shear strength and may impact
 landfill slope stability. Holes
 located in  areas where
 rainwater may pond should be
 patched first. The  patches
 should be sealed with a
                                                     5-59

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems (Continued).
                     Recommendations
Category
Design
Construction
Operation
Liners and
Barriers
(continued)
  at the edge of the liner system
  must be considered in design.
  The potential for gas migration
  into the subsurface can be
  reduced by collecting gas
  generated in the landfill, using
  low-permeability soils over the
  edge of the liner system, and
  modifying the edge of the liner
  system so that the liner
  extends back up to the ground
  surface (like a reverse anchor
  trench).
 permanent seam and not only
 tack welded.
 When a GM is placed over a
 GCL, the GM should be
 covered with soils as soon as
 possible to minimize swelling
 of the GCL in case of
 hydration. GCL swelling
 results in a decrease in GCL
 shear strength and may impact
 landfill slope stability.
 Connections between GMs
 and ancillary structures should
 be carefully constructed and
 inspected to decrease the
 potential for construction-
 related GM defects.
 To decrease the potential for
 construction-related  GM
 defects in sumps, the GM
 panel layout should be
 configured to minimize seams
 in sumps or prefabricated
 sumps should be used.
 With respect to the potential for
 leakage, pipe penetrations are
 generally the most critical
 locations in landfills without
 internal sumps. If pipe
 penetrations are used, they
 should be carefully constructed
 and inspected.
 Sumps and pipe penetrations
 of liners should be leak tested
 by ponding tests, leak location
 surveys, gas tracer tests, or
 pressure  tests of double pipe
 boots as part of liner system
 CQA.  Leak testing of the liner
 on the landfill base (where
 leachate heads are the
 highest) may also be
 considered. Identified holes
 should be repaired.
 The entire installed GM should
 be inspected for damage and
 any damage should be
 repaired prior to placement of
 overlying materials.
 GM should be covered with a
 soil layer as soon as
 practicable after installation,
 but not during the hottest time
 of the day if the GM is
 significantly wrinkled, to reduce
 GM wrinkles, prevent GM
 uplift by wind, and protect the
 GM from  damage.	
                                                    5-60

-------
Table 5-12.  Specific Recommendations to Reduce Landfill  Problems (Continued).
                      Recommendations
Category
Design
Construction
Operation
Liners and
Barriers
(continued)
                                 Prior to placing soil over a GM,
                                 the GM should be inspected for
                                 wrinkles.  Excessive GM
                                 wrinkles and wrinkles that may
                                 fold over should be removed
                                 by waiting to backfill until the
                                 GM cools and contracts during
                                 the cooler nighttime and early
                                 morning hours, pulling the
                                 wrinkles out, or cutting the
                                 wrinkles out. The latter
                                 method is less desirable than
                                 the former methods because it
                                 requires intact GM to be cut,
                                 and it results in more GM
                                 seaming and testing.
                                 On long side slopes, it may be
                                 preferable to use textured GM
                                 rather than smooth GM to
                                 decrease the size of GM
                                 wrinkles that develop,
                                 especially near the slope toe.
                                 Composite liners and barriers
                                 constructed with a CCL should
                                 be covered with an insulation
                                 layer as soon as practicable to
                                 prevent CCL desiccation
                                 related to heating  or freeze-
                                 thaw action.
Drainage
Systems
  Adjacent materials conveying
  water should be designed to
  decrease the clogging potential
  of the downgradient material
  using filter criteria calculations
  and/or laboratory testing.
  If gap-graded soils are used as
  drainage materials, the effect
  of particle migration should  be
  evaluated during design using
  filter criteria calculations and/or
  laboratory testing. In fact, the
  effect of particle migration from
  all granular drainage materials
  should be evaluated since
  these materials have fines.
  Perforated pipes bedded in
  gravel should not be wrapped
  with a GT because the GT is
  useless, and, in some cases,
  even detrimental because the
  GT in this location is prone to
  clogging. Instead, the design
  should include a GT between
  the gravel and the surrounding
  soil or, possibly, no GT.
  Geosynthetic anchor trenches
  should be backfilled with low-
  permeability soil and the soil
  The drainage system should
  be kept free of debris that may
  impede the flow of liquid. In
  general, all sandbags should
  be removed from the drainage
  system. However, if the sand
  in the bags meets the project
  specifications for the overlying
  drainage layer material, the
  bags can be cut and removed
  and the sand left in place.
  GTs and GCs should be
  covered as soon as possible
  after installation to protect
  them from the environment
  (e.g., ultraviolet  light, water,
  high temperature, animals).
  The CQA consultant should
  verify that  all connections
  required for adjacent drainage
  system pipes have been made.
  When pipe is connected by
  butt fusion seaming, the seam
  should be  inspected for
  defects.
  Care should be taken to not
  damage drainage system pipes
  during  construction. The
  contractor should maintain
• Leachate may seep from
  landfill side slopes if the
  leachate can perch on layers of
  less permeable materials (e.g.,
  daily and intermediate cover
  materials) within the waste or
  drain from layers of more
  permeable materials (e.g.,
  tires) in the waste that are
  located relatively close to the
  side slope. The potential for
  seepage can be decreased by:
  (i) not placing layers of the
  more permeable materials near
  the side slopes; (ii) sloping
  layers of the less and more
  permeable materials away from
  the side slopes; (iii) distributing
  the more permeable materials
  throughout the waste; (iv)
  constructing leachate chimney
  drains to the LCRS around
  these layers; (v) removing
  perched leachate from wells
  installed to these layers; and
  (vi) using alternate daily covers
  (e.g., foams, tarps) that do not
  result in layers of less
                                                      5-61

-------
Table 5-12.  Specific  Recommendations to Reduce Landfill Problems (Continued).
                      Recommendations
Category
Design
Construction
Operation
Drainage
Systems
(continued)
  should be well compacted to
  reduce the potential for water
  to infiltrate into the trenches
  and flow into LCRSs or LDSs.
  If this is not practicable, the
  anchor trenches should be
  designed to drain freely and/or
  covered with a barrier, such as
  a GM. In addition, the ground
  surface should be graded away
  from the  trenches to  reduce
  runon from infiltrating into the
  trenches.
  Project specifications for
  needlepunched nonwoven GTs
  should require that the  GTs be
  needle-free and should require
  a certification from the
  manufacturer attesting to  this.
  Needles, if present, may
  damage  a nearby GM.
  The CQA Plan should require
  that deployed GTs near GMs
  be  inspected for needles
  before the GTs are covered
  with overlying materials.  If
  needles are found, the  GT
  should be rejected.
  If a GT is to be exposed to the
  environment for an extended
  time period after installation,
  the potential for degradation of
  the GT should be evaluated
  under all the anticipated
  environmental conditions.  EPA
  recommends that the effect of
  ultraviolet light on GT
  properties be evaluated using
  ASTM D 4355 (Daniel and
  Koerner, 1993).  This test is
  typically run for 500 hours;
  however, it can be run for
  longer time periods to meet
  project-specific conditions.  In
  any case, prior to covering the
  exposed GT, the  condition of
  the GT should be evaluated by
  laboratory testing to verify that
  the GT is still satisfactory.
  If test results indicate that the
  GT will not have the  required
  properties after exposure
  (typically a specified  strength
  retention), the GT should  be
  protected with a sacrificial
  opaque waterproof plastic tarp,
  soil layer, or other means.
  When the waste in a
 sufficient soil cover between
 construction equipment and
 the pipes during construction.
 Equipment operators should be
 aware of pipe locations, since
 pipes can be crushed by
 trafficking equipment.  Also,
 soil around pipes should be
 compacted using hand
 operated or walk-behind
 compaction equipment.
 After construction of a cell with
 an external sump, the pipe
 from the cell to the sump
 should be inspected to verify
 that the pipe is functioning as
 designed. The inspection may
 be performed by surveying the
 pipe with a video camera,
 pulling a mandrel through the
 pipe, flushing the pipe with
 water, or other means.
  permeable materials in the
  waste.
  Drainage system pipes should
  be maintained by cleaning the
  pipes at least annually and
  more frequently, if warranted.
  Landfills with external sumps
  may also include riser pipes at
  the low point of LCRSs as a
  precautionary measure to allow
  for leachate removal from the
  landfill, if necessary.
  Leachate flow measurement
  systems should be calibrated
  and adjusted as needed at
  least annually to ensure that
  the quantities measured are
  accurate.
  Due to the potential for
  problems in automated
  leachate metering and
  pumping equipment, landfill
  operations plans should
  include a verification program
  and contingency method for
  estimating the quantities of
  liquid removed from the LCRS
  and LDS.
  Leachate sump pumps should
  be self priming so the pumps
  will not become airlocked and
  shut down if air is pulled into
  the pumps.
  Leachate sump pumps should
  be selected to be compatible
  with sump geometries and
  anticipated leachate recharge
  rates so pump cycles are
  appropriate (e.g., not so  short
  that the  pumps turn on and off
  too frequently and burn out
  prematurely).
  The "pump on" levels in
  internal  sumps should be kept
  as low as practicable to reduce
  leakage if there are holes in
  the GM  liner in the sump,
  especially if the GM is not
  underlain by a GCL. It is
  recognized, however, that
  "pump on" liquid levels in
  internal  sumps may need to be
  larger than 0.3 m to achieve
  efficient sump pump operation.
  The potential for clogging of
  water-level indicators, pumps,
  and flowmeters must be
  considered when selecting the
                                                      5-62

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems (Continued).
                     Recommendations
Category
Design
Construction
Operation
Drainage
Systems
(continued)
  containment system contains
  some fine particles that may
  migrate to the LCRS, the
  potential for LCRS clogging
  may be reduced by allowing
  those fine particles to pass
  though the drainage system to
  the LCRS pipes, which can
  subsequently be cleaned. The
  fine particles will pass more
  easily through the LCRS if no
  GTs are used in the LCRS or if
  the LCRS contains relatively
  thin open nonwoven GTs
  rather than thicker nonwoven
  GTs with a smaller apparent
  opening size.  Note that the
  above does not apply to an
  LCRS with only a GN drainage
  layer. Though a GN drainage
  layer has a high transmissivity,
  it is thin and is, therefore,
  generally more susceptible to
  clogging by sedimentation than
  a granular drainage layer.
                                types of equipment to use at a
                                facility.
                                Outlets of cover system
                                drainage layers should be kept
                                free of snow and ice so that
                                these layers can drain freely.
Surface
Layers
and
Protection
Layers
  Erosion of soil protection layers
  on liner system side slopes
  should be anticipated and dealt
  with in design.  The potential
  for erosion can be reduced by
  grading the liner system to
  avoid concentrated runoff and
  using a relatively permeable
  soil in the protection layer. In
  areas where the potential for
  erosion is relatively high,
  erosion control structures (e.g.,
  silt fence) can be used to
  reduce the need for intensive
  maintenance of soil protection
  layers. Protection layers can
  also be covered with a tarp or
  temporary erosion control mat.
  When a landfill is constructed
  on top of an existing landfill
  (vertical expansion), an
  exposed  GM liner can be
  uplifted by gases from the
  underlying landfill.  Therefore,
  in the case of a vertical
  expansion, unless gases from
  the underlying landfill are well
  controlled, GMs must be
  covered by a soil layer to
  prevent GM uplift and
  precautions must be taken to
  prevent erosion of this soil.
 Though it may be less costly
 for the owner to construct
 several landfill cells at once,
 this can leave new cells
 exposed to the environment for
 a significant time period.
 These cells will experience
 more erosion than cells filled
 sooner and will have more
 opportunity for liner damage.
 Additionally, every time an
 eroded soil layer is pushed
 back up the side slopes there
 is an opportunity for the
 underlying liner system
 materials to be damaged by
 construction equipment.
                                                     5-63

-------
Table 5-12.  Specific Recommendations to Reduce Landfill  Problems (Continued).
                     Recommendations
Category
Design
Construction
Operation
Surface
Layers
and
Protection
Layers
(continued)
  Better methods for protecting
  exposed soil layers on liner
  system side slopes from
  erosion or alternatives to soil
  layers (e.g., sand filled mats,
  Styrofoam sheets) are needed.
  Post-construction plans should
  be developed for portions of
  landfills that may sit idle for an
  extended period of time.  The
  plans should include
  procedures describing how the
  liner systems should be
  maintained prior to operation.
  For liner systems where soil
  protection layers are placed
  incrementally during landfill
  operation, a geosynthetic
  cushion (supercushion) better
  than the usual thick nonwoven
  GT needs to be developed to
  protect the liner system during
  soil placement.
  Erosion of surface layers on
  cover system side slopes
  should be anticipated and dealt
  with in design. In areas where
  the potential for erosion is
  relatively high, erosion control
  measures (e.g., silt fence, turf
  reinforcement and revegetation
  mat) can  be specified to
  reduce the need for intensive
  maintenance of soil layers.
  However, the erosion control
  measures themselves require
  maintenance.
  The length of cover system
  slopes between ditches or
  swales where runoff is
  collected should be selected to
  limit erosion to acceptable
  amounts (e.g., 5 tonnes/ha/yr).
  At a minimum, the potential for
  erosion should be evaluated
  using the universal soil loss
  equation. Cover system
  slopes may need to be 41-1:1 V
  or less and intercepted by
  swales at 6-m vertical intervals
  to meet acceptable erosion
  levels (EPA, 1994).
  Design flow velocities in
  drainage channels should be
  calculated so the appropriate
  channel lining can be selected.
                                                    5-64

-------
Table 5-12.  Specific Recommendations to Reduce  Landfill  Problems (Continued).
                      Recommendations
Category
Design
Construction
Operation
Liner System
and
Cover System
Stability
  The stability of liner system
  and cover system slopes
  should always be evaluated
  using rigorous slope stability
  analysis methods that consider
  actual shear strengths of
  materials, anticipated seepage
  pressures, and anticipated
  loadings.
  The majority of the slides
  described herein occurred
  along geosynthetic/
  geosynthetic interfaces. For a
  number of these cases, the
  interface shear strengths were
  estimated on the basis of
  published test data.  This
  approach should be avoided
  because there may be
  significant differences in
  interface shear strengths
  between similar materials from
  different manufacturers and
  even identical materials in
  different production lots from
  the same manufacturer.
  Because of this, geosynthetic
  interface shear strengths
  should be measured and not
  estimated.
  Interface shear strength test
  conditions (moisture, stresses,
  displacement rate, and
  displacement magnitude)
  should be representative of
  field conditions.
  The effects of variation in the
  tested geosynthetics, accuracy
  of test methods, and test
  conditions must be considered
  when selecting the design
  interface shear strength.
  Freeze-thaw of CCLs can have
  a significantly detrimental
  impact on GM/CCL interface
  shear strength and should be
  considered when selecting the
  interface shear strength to use
  in slope stability analyses.
  However, freeze-thaw effects
  on interface strength should
  not actually be a design
  consideration, since CCLs
  should be protected from
  freezing in the first place.
  The effect of construction on
  moisture conditions at the
  GM/CCL interface should be
  Soils should be placed over
  geosynthetics from the toe of
  slope upward to avoid
  tensioning the geosynthetics.
  Methods of soil placement that
  are not toe to top should be
  pre-approved by the engineer
  who analyzed the stability.
  Geosynthetic reinforcement
  should be  anchored prior to
  placing the soil layer to be
  reinforced.
  Outlets of drainage layers
  should be kept free of snow
  and ice so these layers can
  drain freely and prevent the
  buildup of seepage pressures.
  Soils or waste should be
  placed over geosynthetics from
  the toe of slope upward to
  avoid tensioning the
  geosynthetics.  Methods of
  waste placement that are not
  toe to top should be pre-
  approved by the engineer who
  analyzed the stability.
  Surficial cracking of soil cover
  layers during seismic loading,
  especially near locations with
  contrast in seismic response
  characteristics  (e.g.,  top of
  waste by rock canyon walls),
  should be anticipated and  dealt
  with as an operation  issue
  through post-earthquake
  inspection and  repair.
  Proposed changes to the
  landfill  filling sequence should
  be reviewed by the design
  engineer to ensure that these
  changes  will not adversely
  impact slope stability.
  Soil layers anchoring
  geosynthetics should be
  maintained during  landfill
  construction and operation.
                                                      5-65

-------
Table 5-12.  Specific Recommendations to Reduce Landfill  Problems (Continued).
                     Recommendations
Category
Design
Construction
Operation
Liner System
and
Cover System
Stability
(continued)
  considered when developing
  the specification for CCL
  construction and selecting the
  strength of liner system
  interfaces for slope stability
  analyses. The CCL
  construction specification
  should generally include
  limitations on maximum
  compacted moisture content,
  restrictions on applying
  supplemental moisture, and
  requirements for covering the
  CCL and overlying GM as soon
  as practical to minimize
  moisture migration to the
  GM/CCL interface.  If a CCL on
  a slope becomes desiccated, it
  should be reworked and not
  just moistened.
  Cover systems incorporating a
  low-permeability barrier layer
  should include a drainage layer
  above the barrier when the
  cover system side slopes are
  steeper than 5H:1V (EPA,
  1994). The purpose of this
  drainage layer is to prevent the
  buildup of seepage pressures
  in the cover system soil
  layer(s) overlying the barrier
  layer.
  When liner systems or cover
  systems are constructed over
  wastes, the potential for the
  wastes to generate gases that
  uplift the liners or barriers must
  be considered.  The  gas
  pressures decrease the shear
  strength along the bottom
  interface of the uplifted layer
  and may lead to slope
  instability.  Gas collection
  systems, therefore, may be
  required to prevent the buildup
  of gas pressures.
  Cover system drainage layers
  should be designed to handle
  the total anticipated flow to the
  drainage layer calculated  using
  a water balance or other
  appropriate analysis (e.g.,
  Giroud and Houlihan, 1995).
  Soong and Koerner (1997)
  recommend using a short-
  duration intensive storm in the
  water balance and do not
                                                    5-66

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems (Continued).
                     Recommendations
Category
Design
Construction
Operation
Liner System
and
Cover System
Stability
(continued)
  recommend the EPA HELP
  computer model for this
  purpose.  The drainage layer
  flow rates output from the
  HELP model are an average
  for a 24-hour period and may
  be much less than the peak
  flow rates calculated using
  other methods if the
  precipitation data used in the
  HELP model are not carefully
  selected.
  Water collected in the drainage
  layer must be allowed to outlet
  to prevent the buildup of
  seepage pressures.
  Containment systems should
  be designed to limit seismic
  displacements to tolerable
  amounts.  To do this, designs
  may incorporate predetermined
  slip surfaces to confine
  movements to locations where
  they will cause the least
  damage (i.e., above the GM
  liner) and inverted liner system
  keyways to provide more
  resistance to movement.  For
  example,  a GM with a smooth
  top surface and a textured
  bottom surface could be  used
  in certain  liner systems to
  create a predetermined slip
  surface above the GM.
  Liner system anchor trenches
  should be designed to secure
  geosynthetics during
  construction, but release the
  geosynthetics before they are
  damaged by displacements
  during earthquakes.  An
  alternative is to unanchor the
  liner system after construction
  and secure it on a bench with
  an overlying soil layer.
  Stress concentrations at or
  near the liner system side
  slope crest should be avoided.
  Areas with stress
  concentrations are more
  problematic when subjected  to
  seismic displacements.  In
  particular, GM seams should
  generally  not be sampled near
  the slope  crest.	
                                                   5-67

-------
Table 5-12.  Specific Recommendations to Reduce Landfill Problems (Continued).
                  Recommendations
Category
Design
Construction
Operation
Liner System
and
Cover System
Displacements
 When liner systems or cover
 systems are constructed over
 existing wastes, the potential
 for the wastes to generate
 gases must be considered.
 The gases may uplift GMs,
 causing excessive stresses in
 the GMs and may impact slope
 stability. Some landfills may
 be generating little or no gas at
 the time of construction and
 may not need a gas collection
 system. Other landfills may be
 generating significant amounts
 of gas and may require a gas
 collection system beneath the
 entire liner system.
 Surface-water runoff should be
 managed to reduce foundation
 uplift problems during and after
 construction. Temporary and
 permanent surface-water
 diversion structures located
 near a cell may need to be
 lined to reduce infiltration,
 especially if the structures are
 located on relatively permeable
 soils and convey relatively
 large amounts of water.
 Runoff should not be allowed
 to pond near the cell, where it
 can infiltrate into the cell.
 Liner systems constructed over
 compressible, low shear
 strength waste materials
 should  be designed to
 accommodate the anticipated
 settlements. When GCL is
 used, seam overlaps should be
 wider than normal.
 Gas extraction well boots
 should  be designed to
 accommodate the anticipated
 landfill settlements.
 Cover systems with soil layers
 placed over compressible, low
 shear strength waste should
 use lightweight construction
 equipment and have good
 control of the thickness of soil
 placed over the waste so as
 not to cause bearing capacity
 failure of the waste and
 excessive displacement of the
 cover system.
5.4  Assessment of EPA HELP Model Using Leachate Generation Data

5.4.1  Introduction
The HELP model was developed by the U.S. Army Engineer Waterways Experiment
Station for the EPA Risk Reduction Engineering Laboratory, Cincinnati,  OH, to help
landfill designers compare the hydraulic performance of alternative waste containment
system designs.  However, the HELP model has increasingly been used to design
LCRS drainage layers and to estimate leachate generation  rates in order to size
leachate storage tanks.  There is little published information on the adequacy of the
HELP model for these purposes.  In this section of the  report, the HELP model is
assessed by comparing LCRS flow rate data from six landfill cells to leachate
                                            5-68

-------
generation rates predicted for these cells from HELP model simulations with typical
input parameters. The cells were selected to represent different waste types and
geographical regions of the U.S.  General data for the six landfill cells are included in
the landfill performance database described in Section 5.2.2.

The HELP model theoretical development and data requirements are described in
Section 5.4.2. A review of published literature on leachate generation rate predictions
by the HELP model is presented  in Section 5.4.3. The evaluation of the HELP model as
a design tool is described in Section 5.4.4, and results of this evaluation are presented
in Section 5.4.5.

5.4.2  Description of HELP Model
The HELP model simulates hydrologic processes for a landfill by performing daily,
sequential water budget analyses using a quasi-two-dimensional, deterministic
approach (Schroeder et al., 1994a, 1994b).  The hydrologic processes considered in the
HELP model include precipitation, surface-water storage (i.e., storage as snow),
interception of precipitation by foliage, surface-water evaporation, runoff, snow melt,
infiltration, plant transpiration, soil water evaporation, soil water storage, vertical drainage
(saturated and unsaturated) through non-barrier soil layers, vertical percolation (saturated)
through soil barriers, vertical percolation through GM and GM/soil composite barriers, and
lateral drainage (saturated).

Five main routines are used in  the HELP model to estimate runoff, evapotranspiration,
vertical drainage, vertical percolation, and lateral drainage.  Several other routines interact
with the main routines to generate daily precipitation, temperature, and solar radiation
values and simulate snow accumulation and melt, vegetative growth, interception, and
leakage through GM and GM/soil composite barriers. Runoff is computed using the runoff
curve-number method of the U.S.  Department of Agriculture Soil Conservation Service
(USDA-SCS) (USDA-SCS, 1985). Evapotranspiration is computed using a two-stage
modified Penman energy balance method developed by Ritchie (1972). Vertical drainage
is computed using Darcy's equation, modified to allow drainage under unsaturated
conditions using an unsaturated soil hydraulic conductivity calculated using an equation by
Campbell (1974). Percolation through a soil liner or barrier is also evaluated using Darcy's
equation, but under saturated conditions. Lateral drainage is modeled by an analytical
approximation to the steady-state solution of the Boussinesq equation.  Leakage through
GMs and GM/soil composite liners or barriers is evaluated based on the work of Giroud
and Bonaparte (1989a, 1989b) and Giroud et al. (1992).

Version 1 of the HELP model  was issued in 1984, and the model has been updated
several  times since then.  At the time this  report was prepared, Version 3 was the most
current. Data requirements for Version 3 of the HELP model are summarized in Table
5-13.  HELP requires daily and general climatic data, material properties data for the
landfill components being modeled, and landfill design data.  Required daily weather

                                       5-69

-------
data are precipitation, mean temperature, and total global solar radiation.  Daily
precipitation may be input manually, selected from a historical database (e.g., 1974-1977
data in HELP database, NOAATape, or Climatedata™), or generated stochastically
using a weather generation model developed by the U.S. Department of Agriculture-
Agricultural Research Service (USDA-ARS) (Richardson and Wright, 1984) with simulation
parameters available for 139 U.S. cities. Other daily climatologic data are generated
stochastically using the USDA-ARS routine. Required general weather data include
average annual wind speed  and latitude.  Default general weather data for 183 U.S.
cities are used by the model. The material properties of each layer being modeled are
either selected from the HELP model database of  default material properties or are
specified by the model user. Landfill design data,  including landfill general information
and layer configuration, are user specified.
Table 5-13.  Data Requirements for the EPA HELP Model, Version 3.	
WEATHER DATA REQUIREMENTS

     Evapotranspiration Data
          Default evapotranspiration option
               Location
               Evaporation zone depth
               Maximum leaf area index
          Manual evapotranspiration option
               Location
               Evaporation zone depth
               Maximum leaf area index
               Dates starting and ending the growing season
               Normal average annual wind speed
               Normal average quarterly relative humidity

     Precipitation Data
          Default precipitation option
               Location
          Synthetic precipitation option
               Location
               Number of years of data to be generated
               Normal mean monthly precipitation
          Create/Edit precipitation option
               Location
               One or more years of daily precipitation data
          NOAA Tape  precipitation option
               Location
               NOAA ASCII print file of Summary of Day daily precipitation data in as-on-tape format
          Climatedata™ precipitation option
              Location
              Climatedata™ prepared file containing daily precipitation data
          ASCII precipitation option
               Location
              Files containing ASCII data
              Years
                                         5-70

-------
Table 5-13.  Data Requirements for the EPA HELP Model, Version 3 (Continued).
          Precipitation Data (continued)
          HELP Version 2 data option
                Location
                File containing HELP Version 2 data

    Temperature Data
          Synthetic temperature option
                Location
                Number of years of data to be generated
                Years of daily temperature values
                Normal mean monthly temperature
          Create/Edit temperature option
                Location
                One or more years of daily temperature data
          NOAA Tape temperature option
                Location
                NOAA ASCII print file of Summary of Day data file containing years of daily maximum
                  temperature values or daily mean temperature values in as-on-tape format
                NOAA ASCII print file of Summary of Day data file containing years of daily minimum
                  temperature values or daily mean temperature values in as-on-tape format
          Climatedata M temperature option
                Location
                Climatedata™ prepared file containing daily maximum temperature data
                Climatedata™ prepared file containing daily minimum temperature data
          ASCII temperature option
                Location
                Files containing ASCII data
                Years
          HELP Version 2 data option
                Location
                File containing HELP Version 2 data

     Solar Radiation Data
          Synthetic solar radiation option
                Location
                Number of years of data to be generated
                Years of daily solar radiation values
                Latitude
          Create/Edit solar radiation option
                Location
          One or more years of daily solar radiation data
          NOAA Tape solar radiation option
                Location
                NOAA ASCII print file of Surface Airways Hourly solar radiation data in as-on-tape
                  format
          Climatedata™ solar radiation option
                Location
                Climatedata™ Surface Airways prepared file containing years of daily solar radiation
                  data
          ASCII Solar Radiation Option
               Location
                Files containing ASCII data
               Years
                                            5-71

-------
 Table 5-13.  Data Requirements for the EPA HELP Model, Version 3 (Continued).
      Solar Radiation Data (continued)
                HELP Version 2 Data Option
                     Location
                File containing HELP Version 2 data

 MATERIAL PROPERTY AND DESIGN DATA REQUIREMENTS

      Landfill General Information
           Project title
           Landfill area
           Percentage of landfill area where runoff is possible
           Method of initialization of moisture storage
           Initial snow water storage

      Layer Data
           Layer type
           Layer thickness
           Soil texture (Default, User-built, or manual options)
                Porosity
                Field capacity
                Wilting point
                Saturated hydraulic conductivity
           Initial volumetric soil water content
           Rate of subsurface inflow to layer

      Lateral Drainage Layer Design Data
           Maximum drainage length
           Drain slope
           Percentage of leachate collected from drainage layer that is recirculated
           Layer to receive recirculated leachate from drainage layer

      GM Liner Data
           Pinhole density in GM liner
           GM liner installation defects
           GM liner installation quality
           GM liner saturated hydraulic conductivity (vapor diffusivity)
           GT transmissivity

      Runoff Curve Number Information
           User-specified runoff curve number  used without modification
           User-specified runoff curve number  modified for surface slope and slope length
	Curve number calculated by HELP	
 5.4.3  Literature Review
 A number of researchers have performed field studies and analytical assessments to
 evaluate the HELP model (EPRI, 1984; Thompson and Tyler, 1984; Peters et al., 1986;
 Peyton and Schroeder, 1988; Barnes and Rodgers, 1988; Udoh, 1991; Lane et al., 1992;
 Benson et al., 1993;  Peyton and Schroeder, 1993; Field and Nangunoori, 1994;  Khire et al.,
 1994; Lange et al., 1997).  In particular, the studies evaluated the reliability of the HELP
 model as a tool to predict trends and magnitudes of the different landfill water balance


                                           5-72

-------
components (i.e., infiltration, runoff, etc.). The conclusions of these studies are not always
in agreement. For example, some of these studies found that HELP over-predicted
infiltration in humid climates and under-predicted infiltration in arid climates, but other
studies concluded just the opposite. In some cases, the HELP model was not able to
predict short-term trends. However, for a number of cases the HELP model analysis was
shown to give reasonable predictions of cumulative longer-term water balances. Despite
the wide use of the HELP model to predict landfill leachate generation rates, to the
author's knowledge, there are very few published studies comparing leachate generation
rates for modern landfills predicted with HELP model to actual leachate generation rates
measured at these landfills.  Two such studies were performed by Field and Nangunoori
(1994) and Lange et al. (1997). Results from these two studies are summarized below.

Field and Nangunoori (1994) used the HELP model, Version 2 to predict leachate
generation rates at an active, modern double-lined HW landfill in New York. They
compared the predicted rates to measured leachate generation rates at the landfill during
1992. Site-specific and default data were used in the model simulations. Actual site
rainfall data during 1992 were input manually.  The HELP model default values for a city
near the site were used for all other required weather data (i.e., evapotranspiration,
temperature, and solar radiation data).  The waste and liner system geometries used in
the model were selected to represent the actual landfill conditions in 1992. Data for the
intermediate and daily covers, waste, LCRS, and liner system layers (i.e., porosity, field
capacity, wilting point, and saturate hydraulic conductivity) were selected from the HELP
model database of default material characteristics. The permeabilities of the liner
system drainage layers (i.e., a 0.3-m thick sand layer and a GN) were modified to match
laboratory results or manufacturer's data.  The  authors did not provide modeling data
related to the runoff curve number or the percentage of the landfill area where runoff is
possible. The average annual leachate generation rate predicted by HELP was 36% of
the measured average annual LCRS flow rate.  Field and Nangunoori (1994) reported
that increasing the default hydraulic conductivity of the waste from 2x10"6 to 2x10"5 m/s
caused the predicted leachate generation rate to be within 17% of the measured rate.  It
is noted that the HELP model does not contain default properties for HWs, and Field and
Nangunoori used the default  properties for MSW in their simulations.

Lange et al. (1997) presented a case study of the use of the HELP model to predict
leachate generation rates at a modern MSW landfill in northeastern Ohio.  The liner
system for the landfill consisted of a GM/GCL/CCL composite liner and a GN drainage
layer. At the landfill permitting stage, the landfill designer used Version 2.05 of the HELP
model to estimate leachate generation rates. The designer used the model's default
weather data for a nearby city and default material property data.  The designer
assumed intermediate covers would be bare (i.e., unvegetated) and no surface water
runoff would leave the modeled landfill area.  The average annual leachate generation
rate estimated using the HELP model was 12,200 Iphd. The landfill began waste
placement operations in December 1992 on a 5-ha portion of the landfill.  Additional

                                      5-73

-------
areas were utilized with time, up to 25 ha by October 1995.  Actual leachate generation
rates were measured between December 1992 and February 1997.  Rates were highest,
up to 15,000 Iphd, during the first few months of operations and decreased with time. By
July 1996, LCRS flow rates had reached relatively steady levels of 800 to 900 Iphd.  In
this case, the leachate generation rate estimated during the design phase was in the
range of measured rates during the first few months of the landfill operation, but was not
representative of rates measured in the  following years.

Lange et al. (1997) evaluated the ability of the HELP model, Version 3.05 to predict
leachate generation rates that occurred at different times in the life of the landfill by using
input data that model the conditions of the landfill at these times.  In particular, actual
precipitation and temperature data recorded at a nearby city between 1992 and 1996 were
used in the HELP model. For comparison, one-year and five-year simulations were
performed.  Furthermore, the landfill was  modeled in terms of five areas with different cover,
slope, and vegetation conditions.  The extent of these areas varied during the operation of
the landfill. The leachate generation rate for the landfill at a certain time was estimated as
the area-weighted average of the average leachate generation rates for the five areas
during the considered simulation period (i.e.,  one year or five years). The landfill geometry
was modeled  based on aerial topographic maps obtained at different times in the life of the
landfill. The fraction of runoff that could exit the landfill areas was assumed to be zero for
the active area (i.e., working face) and areas with no waste, 75% for areas covered with
daily cover, and 100% for all other areas.  Good vegetation was assumed for intermediate
cover placed over areas that reached final grades. The material property data were, for the
most part, selected from the HELP default values.  The predicted leachate generation  rates
at the five different times in the life of the  landfill were between 90 and 230%  of the
measured rates.  It is noted that the length of the simulation period had limited effect on the
predicted average leachate generation rates.

5.4.4  Evaluation of HELP Model
The performance of the HELP model was evaluated as a "design  tool" to estimate
landfill leachate generation rates using a specific simulation methodology (i.e.,  specific
procedures for selecting simulation period and model input parameter values).  The
evaluation was conducted by comparing leachate generation rates estimated by HELP
Version 3.04a when used with the specific simulation methodology for six landfill cells to
measured LCRS flow rates at these cells. Table 5-14 presents operation information
and LCRS and liner system details for the six cells.  Four of the cells contain MSW, one
contains HW, and one contains MSW ash.  All of the MSW cells and the ash cell are
located in the NE; the HWcell is located in the W.  The cells varied in area between 2.2
and 6.4 ha and varied in maximum waste height between 21 and  46  m.  Average annual
rainfall at the  landfill sites was lowest (i.e., 280 mm) for Cell AC2 located in the Wand
highest (i.e., 1,190  mm) for Cells Y1  and Y2, located in the NE. As shown in Table 5-
14, all of the cells, except for Cell AC2, have a sand LCRS drainage  layer.  Cell
                                      5-74

-------
en
•Ij
en
                       Table 5-14. Operation Information and Liner System Details for Six Landfill Cells
                                        Modeled  Using EPA HELP Model.
Cell
No.



B1
B3
12
Y1
Y2
AC2
U.S.
Region11'



NE
NE
NE
NE
NE
W
Avg.
Annual
Rainfall

(mm)
1,070
1,070
990
1,190
1,190
280
Waste
Type«



MSW
MSW
MSW
ASH
MSW
HW
Cell
Area


(ha)
3.3
6.4
2.4
2.2
3.0
4.2
Max.
Waste
Height

(m)
21
25
46
10
15
30
Avg.
Liner
Base
Slope
(%)
2.0
2.0
2.5
5.5
5.5
2.0
LCRS Material
Type14'



S
S
S
S
S
G/GN
Thick.


(mm)
450
450
600
600
600
300/5
Hydraulic
Cond.(5)

(m/s)
1X10"4
1X10"4
1X10"4
5X1 0"5
5X1 0"5
5X10"3/0.1
LCRS Collector Pipe
Size (mm)
& Material'61


152PVC
152PVC
ND(3)
152PVC
152PVC
ND
Spacing


(m)
38
30
30
30
30
30
Primary Liner
Liner
Type'"


GM
GM/CCL
GM
GM/CCL
GM/CCL
GM/CCL
GM
Material16'


CSPE
CSPE
HOPE
HOPE
HOPE
HOPE
Thick.

(mm)
0.9
0.9
1.5
2.0
2.0
1.5
CCL
Thick.

(mm)
NA(3)
600
NA
450
450
450
Hydraulic
Cond.(5)
(m/s)
NA
1X10"9
NA
1X10"9
1X10"9
1X10"9
Notes:
(1) Regionsof the U.S. are:  NE = northeast, SE = southeast, W=west
(2) Waste types are: MSW = municipal solid waste, HW= hazardous waste, ASH = MSW ash
(3) ND = not determined; NA = not applicable
(4) LCRS material types are: S = sand, G = gravel, GN = geonet
(5) Hydraulic conductivity values shown represent minimum values for the LCRS drainage layer and maximum values for the CCL, as required in the project
   material specifications.
(6) Collector pipe and primary liner GM materials are:  PVC = polyvinyl chloride, HOPE = high-density polyethylene,  CSPE = chlorosulfonated polyethylene
(7) Liner types are: GM = geomembrane,  CCL = compacted clay liner

-------
AC2 has a gravel drainage layer underlain by a GN. LCRS collector pipes were used at
all of the cells and were spaced 30 to 38 m apart.

The simulation methodology used herein involved modeling four landfill scenarios
representing conditions that typically occur at different times and in different areas
within a cell during landfill operations. The first scenario assumes that essentially no
waste has been placed on the liner system and no measures have been implemented to
prevent direct infiltration of rainwater into the LCRS. This scenario may occur during
the first few months of cell operation.  The second scenario assumes waste has been
placed in the cell, and either no daily cover has been placed on the waste or no
measures have been implemented to divert clean rainwater from the cell. The third
scenario models an area of the  cell that has received waste and has been covered with
daily cover. It was assumed that the daily cover can shed away 50% of the storm-water
runoff in the cell. The fourth scenario assumes an intermediate cover has been placed
on waste that has almost reached final grades. The intermediate cover is assumed to
be vegetated with good grass and capable of diverting 100% of storm-water runoff.

The simulation methodology used herein utilized average annual (not peak) HELP
model results calculated over a  100-year simulation period.  Other methodologies could
also be used, although no other methodologies were used in  the preparation of this
report. Weather data for the simulation (i.e., daily precipitation, temperature,  and solar
radiation values) were  generated stochastically by the  HELP  program for the closest city
to the landfill site for which HELP has built-in weather parameters.  One hundred years
of data were generated and used in the HELP simulation to represent a wide range of
weather conditions that the landfill may experience.  The normal mean  monthly
precipitation values were modified to match the historical average annual precipitation
at the site. Table 5-15 summarizes HELP soil and design parameter values used for
the four different scenarios modeled at the six landfill cells. Daily and intermediate
covers and waste were modeled as vertical percolation layers. The LCRS drainage
material was modeled as a lateral drainage layer, and the CCL component of the
primary liner was modeled as a barrier soil liner. The layer material properties were
selected from the HELP model database of default material properties to represent
material properties described in the landfill  design plans and specifications.  LCRS
material hydraulic conductivity was modified in some cases to reflect values required by
the project specifications.  The cell geometry parameters (i.e., drainage length and
slope, waste height,  and surface slope and length) were selected based on the cell
design plans and on anticipated waste placement sequence and practices.

5.4.5 Study Findings
The results of the HELP model simulations are summarized in Table 5-16.  Reported for
each cell are actual average LCRS flow rates measured at the cell during the initial and
active periods of operation, as well as average flow rates obtained using HELP with the
four cell modeling scenarios.  Average annual LCRS flow rates over the 100-year

                                      5-76

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells.
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

MSWCellBI located in the NE
1
0 for no waste and uncovered areas, 50 for areas with daily
cover, and 100 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)
1
30
10
0.398
0.244
0.136
1.2x10'4
0
Waste

1
varies (1)
18
0.671
0.292
0.077
1x1 0'3
0
Drainage
Material
2
45
1
0.417
0.045
0.018
1x10'2
0
GM CCL

4 NA
0.09 NA
40 NA
NA NA
NA NA
NA NA
3x1 0'12 NA
0 NA

20
2
0

NA


1
5
Good
NA
No
Waste (1)
2
20
1
Bare
75
No
Cover (1)
5
20
18
Bare
82
Daily
Cover (1)
5
20
5
Bare
85
Intermed.
Cover (1)
25
45
10
Good
82







Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore, no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                                5-77

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells (Continued).
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

MSW Cell B3 located in the NE
1
0 for no waste and uncovered areas, 50 for areas with daily
cover, and 100 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)
1
30
10
0.398
0.244
0.136
1.2x10'4
0
Waste

1
varies (1)
18
0.671
0.292
0.077
1x1 0'3
0
Drainage
Material
2
45
1
0.417
0.045
0.018
1x10'2
0
GM CCL

4 3
0.09 60
40 16
NA 0.427
NA 0.418
NA 0.367
3x1 0'12 1x1 0'7
0 0

17
2
0

NA


1
5
Good
NA
No
Waste (1)
2
17
1
Bare
76
No
Cover (1)
5
17
18
Bare
82
Daily
Cover (1)
5
17
5
Bare
85
Intermed.
Cover (1)
25
60
10
Good
82







Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore, no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                               5-78

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells (Continued).
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

MSW Cell 12 located in the NE
1
0 for no waste and uncovered areas, 50 for areas with daily
cover, and 100 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)
1
30
10
0.398
0.244
0.136
1.2x10'4
0
Waste

1
varies (1)
18
0.671
0.292
0.077
1x1 0'3
0
Drainage
Material
2
60
1
0.417
0.045
0.018
1x1 0'2
0
GM CCL

4 NA
0.15 NA
35 NA
NA NA
NA NA
NA NA
2x1 0"13 NA
0 NA

76
2.5
0

NA


1
3
Good
NA
No
Waste (1)
2.5
76
1
Bare
74
No
Cover (1)
5
20
18
Bare
82
Daily
Cover (1)
5
25
5
Bare
85
Intermed.
Cover (1)
20
35
10
Good
83







Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore, no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                               5-79

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells (Continued).
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

MSWAsh Cell Y1 located in the NE
1
0 for no waste and uncovered areas and for areas with
daily cover; 1 00 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)

Not
Applicable





Waste

1
varies (1)
32
0.450
0.116
0.049
1x1 0'2
0
Drainage
Material
2
60
1
0.417
0.045
0.018
5x1 0'3
0
GM CCL

4 3
0.2 45
35 16
NA 0.427
NA 0.418
NA 0.367
2x1 0'13 1x1 0'7
0 0

17
5.5
0

NA


1
3
Good
NA
No
Waste (1)
5.5
17
1
Bare
76
No
Cover (1)
5
17
32
Bare
97
Daily
Cover (1)
5
17
5
Bare
85
Intermed.
Cover (1)

Not
Applicable









Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore, no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                               5-80

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells (Continued).
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

MSW Cell Y2 located in the NE
1
0 for no waste and uncovered areas, 50 for areas with daily
cover, and 100 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)
1
30
10
0.398
0.244
0.136
1.2x10'4
0
Waste

1
varies (1)
18
0.671
0.292
0.077
1x1 0'3
0
Drainage
Material
2
60
1
0.417
0.045
0.018
5x1 0'3
0
GM CCL

4 3
0.2 45
35 16
NA 0.427
NA 0.418
NA 0.367
2x1 0'13 1x1 0'7
0 0

17
5.5
0

NA


1
3
Good
NA
No
Waste (1)
5.5
17
1
Bare
76
No
Cover (1)
5
17
18
Bare
82
Daily
Cover (1)
5
17
5
Bare
85
Intermed.
Cover (1)
30
30
10
Good
83







Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore,  no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                               5-81

-------
Table 5-15.  HELP Model Soil and Design Input Parameters for Select Cells (Continued).
Landfill General Information
- Project title
- Area of modeled portion of landfill (ha)
- Percentage of landfill area
where runoff is possible (percent) (1)
- Moisture storage initialization method
- Initial snow water storage (cm)
Layer Data

- Layer type
- Layer thickness (cm)
- Material texture default number
Porosity (vol. /vol.)
Field capacity (vol. /vol.)
Wilting point (vol./vol.)
Saturated hydraulic conductivity (cm/s)
- Rate of subsurface inflow to layer
Lateral Drainage Layer Design Data
- Maximum drainage length (m)
- Drain slope (%)
- Percentage of leachate collected from
drainage layer that is recirculated (%)
- Layer to receive recirculated leachate
from drainage layer
Geomembrane Liner Data
- Pinhole density per hectare
- Installation defects per hectare
- Geomembrane liner installation quality
- Geotextile transmissivity (m2/sec)
Runoff Curve Number Calculated by HELP

- Surface slope (%)
- Slope length (m)
- Default soil texture
- Quantity of vegetative cover
- Runoff curve number calculated by HELP

HW Cell AC2 located in the W
1
0 for no waste and uncovered areas, 50 for areas with daily
cover, and 100 for areas with an intermediate cover
Program initialized to near steady state
0
Daily
Cover (1)
1
15
5
0.457
0.131
0.058
1x1 0'3
0
Intermed.
Cover (1)
1
30
10
0.398
0.244
0.136
1.2x10'4
0
Waste

1
varies (1)
18
0.671
0.292
0.077
1x1 0'3
0
Protect.
Material
1
30
5
0.457
0.131
0.058
1x10'3
0
Drainage GM CCL
Material
1/2 4 3
30/0.5 0.15 45
1/20 35 16
0.417/0.85 NA 0.427
0.045/0.01 NA 0.418
0.018/0.005 NA 0.367
0.5/10 2x10"13 1x10"7
0 00

40
2.0
0

NA


1
3
Good
NA
No
Waste (1)
2
17
5
Bare
76
No
Cover (1)
5
17
18
Bare
82
Daily
Cover (1)
5
17
5
Bare
85
Interm.
Cover (1)
25
30
10
Good
83







Notes:
(1) Four scenarios are analyzed which model different conditions at different areas of the cell: (i) the no waste
   scenario models conditions at an area which has not received waste and where measures have not been
   implemented to prevent rainwater from entering the LCRS; (ii) the no cover scenario models an active waste
   disposal area which has not received daily or intermediate covers, and therefore, no runoff is allowed (waste
   thickness assumed at 3 m); (iii) the daily cover scenario models an area which received daily cover and allows
   rainwater runoff from 50% of the area (waste thickness assumed at 6 m); and (iv) the intermediate cover
   scenario models an area which received an intermediate cover and allows rainwater runoff from the entire area
   (waste thickness assumed at 12 m).
                                               5-82

-------
simulation were selected for comparison with measured flow rates. As shown in Table
5-16, estimated flow rates are highest for the "no waste" scenario and decrease
significantly with placement of waste and daily cover.  The lowest flow rates were
estimated for the "intermediate cover" scenario.  Therefore, when the HELP model is
used in the manner described above, it seems capable of modeling the trend  of
decreasing leachate generation rate with time observed at landfills. This trend is more
fully investigated in Appendix E.

Table 5-16.  Summary of Measured and Estimated LCRS Flow Rates for Six
            Landfill Cells.
Cell
No.


B1
B3
12
Y1
Y2
AC2
Initial Period of Operation
Time
Period
(months)
ND(1)
1-4
1-7
ND
1-10
1-6
Avg. LCRS
Flow Rate
(Iphd)
ND
15,304
6,627
ND
23,368
272
Active Period of Operation
Time
Period
(months)
20-54
5-93
8-76
13-78
11-54
7-88
Avg. LCRS
Flow Rate
(Iphd)
3,816
3,748
728
19,319
7,918
18
LCRS Flow Rates (Iphd) Estimated Using
HELP for Four Different Scenarios
No
Waste
14,335
16,075
11,698
17,685
17,685
1,885
No
Cover
9,622
10,820
7,290
17,066
12,597
1,250
Daily
Cover
9,124
10,142
7,462
15,504
12,225
1,025
Intermediate
Cover
4,392
5,035
3,533
NA(1)
6,708
130
Notes:  (1) ND = not determined, NA = not applicable.
       (2) Reported measurements are average values over the indicated time period. HELP
          simulation estimates represent average values for 100-year simulation period.

As shown in Table 5-16, the MSWash Cell Y1 exhibited much higher measured LCRS
flow rates than the MSW and HW cells.  During the active period of operation, the
average LCRS flow rate for Cell Y1 was about 19,300 Iphd and the average flow rates
for the other cells ranged from about 20 to 7,900 Iphd.  The higher flow rate in this range
is for MSW Cell Y2, located at the same site and having the same liner system details
and cell geometry as Cell Y1.  The higher LCRS flow rate for Cell Y1 than for Cell Y2
may be attributed to the slow placement rate of ash, the high hydraulic conductivity of
the ash, and lack of storm-water diversion from the ash cell.  The small thickness of
waste, high ash hydraulic conductivity, and lack of storm-water diversion allow for
relatively unimpeded infiltration of rainwater through the waste into the LCRS, and,
therefore, result in high leachate flow rates.  The lowest measured flow rate during the
active period of about 20 Iphd occurred for HW cell AC2,  located in the W at a site with
an average annual rainfall of only 280 mm.

LCRS flow rates  estimated using the HELP simulation methodology exhibited similar
trends as the measured flow rates.  The MSW ash cell Y1 had the highest estimated
flow rates; the HW cell in the W had the lowest estimated flow rates. Leachate
generation rates  of approximately 17,100 Iphd and 12,600 Iphd were estimated for the
MSW ash Cell Y1 and MSW Cell Y2, respectively,  for the "no cover" scenario,
demonstrating that, at least for the  considered cells, the HELP simulation methodology
is capable of predicting relative differences in LCRS flow rates for different waste types.
                                      5-83

-------
Figure 5-8 presents a comparison of average measured LCRS flow rates for the six
cells and the LCRS flow rates estimated for these cells using the HELP simulation
methodology. In this figure, the ranges of estimated LCRS flow rates given in Table
5-16 are plotted against the average measured LCRS flow rates given in the same
table. In particular, estimated LCRS flow rates for the "no waste" and "no cover"
scenarios are directly compared to average measured flow rates during the initial period
of operation and estimated LCRS flow rates for the "daily cover" and "intermediate
cover" scenarios are directly compared to average measured flow rates during the
active period of operation. Each complete data set is represented by a box when the
data points are connected as done in the figure. As shown in Figure 5-8, except for HW
Cell AC2 located in the W, the estimated leachate generation rates using the HELP
simulation methodology are generally of the same order of magnitude. The estimated
LCRS flow rates were somewhat higher than measured flow rates for MSW cells and
somewhat lower than the measured flow rate for the MSW ash cell.  For MSW Cells B1
and B3, the estimated flow rates were somewhat higher than measured flow rates. For
example, for Cell B3, estimated LCRS flow rates were in the range of 5,000 to 16,100
Iphd, while average measured flow rates were 15,300 Iphd during the initial period of
operation and 3,700 Iphd during the active period of operation.  For MSW ash Cell Y1,
the HELP methodology somewhat underpredicted leachate generation rates. Average
measured flow rates were 19,300 Iphd during the active period of operation;  the
estimated flow rates using the HELP simulation methodology were in the range of
15,500 to 17,700 Iphd. For MSW Cell 12 and especially for HW Cell AC2, the estimated
rates were significantly higher than measured rates. Average measured LCRS flow
rates for Cell AC2 were about 270 Iphd during the initial period of operation and 20 Iphd
during the active period of operation; the estimated flow rates using the HELP
simulation methodology were in the range of 130 to 1,000 Iphd.

For the evaluations performed in this section, the HELP model responded as expected
to changes in waste type and site climate. When used with default parameters, the
HELP model may generally overpredict LCRS flow rates for MSW landfills and landfills
in arid climates;  however, too few sites were evaluated to draw definitive conclusions.
Part  of the conservatism of the HELP model in predicting LCRS flow rates, especially as
waste is placed, may lie in the default moisture content value for the waste.  In the
HELP model, the waste is assumed to be at field capacity. However, MSW is typically
placed at moisture contents less than field capacity.  The moisture storage capacity of
waste is particularly important in arid climates since this storage capacity, if utilized,
may  hold essentially all the rainwater infiltrating the waste, and little leachate will be
generated. Also, some landfill cells,  and especially HW cells, have special methods of
handling rainwater that may not be taken into account in the  HELP model. For
example, in some HW cells, part of the waste is covered with a GM during cell
operation.  Rainwater collected on the GM is removed from the cell separate from
leachate, and, if clean, is discharged.

                                     5-84

-------
          25,000 -r
                                            Y1    XY2    oAC2
       LLI
                        5,000     10,000     15,000     20,000    25,000
                        MEASURED LCRS FLOW RATES (Iphd)
Figure 5-8. LCRS flow rates estimated using HELP versus measured LCRS flow
           rates for six landfill cells.
The authors believe that the HELP model can appropriately be employed as a tool to
estimate long-term average leachate generation rates to use with an appropriate level of
conservatism in the design of LCRS drainage layers and the sizing of leachate
management system components. The authors recommend that users develop a
consistent simulation methodology (analogous to the methodology used herein, with the
same or different underlying assumptions) for the HELP model and that they evaluate
the simulations, similar to the evaluations in Table 5-16 and Figure 5-8, using data from
existing local landfills. These simulations can be enhanced by performing parametric
analyses for key input parameters, such as initial waste moisture content.  With this
consistent, locally calibrated approach, the usefulness of the HELP model as a design
tool can be improved.
                                     5-85

-------
5.5 References
Barnes, F. J. and Rodgers, J. E. (1988), "Evaluation of Hydrologic Models in the Design
   of Stable Landfill Final Covers", U.S. Environmental Protection Agency, Office of
   Research and Development, Cincinnati, OH, EPA Project Summary, Report No.
   EPA/600/S-88/048.
Bass, J. M. (1986), "Avoiding Failure ofLeachate Collection and Cap Drainage Layers",
   EPA/600/2-86/058, U.S. Environmental Protection Agency,  Hazardous Waste
   Engineering Research Laboratory, Cincinnati, OH, 142 p.
Benson, C. H., Khire, M. V., and Bosscher, P. J. (1993),  "Final Cover Hydrologic
   Evaluation, Final Cover-Phase II", Environmental Geotechnics Report No. 93-4,
   University of Wisconsin, Madison, Wl, 151 p.
Bonaparte, R. and Gross B. A. (1990), "Field Behavior of Double-Liner Systems", Waste
   Containment Systems: Construction, Regulation, and Performance, ASCE
   Geotechnical Special Publication No. 26, R. Bonaparte ed., New York, pp. 52-83.
Bonaparte, R., Othman, M. A., Rad, N. S., Swan, R. H., and Vander Linde, D.  L. (1996),
   "Evaluation of Various Aspects of GCL Performance", Appendix F in Report of 1995
   Workshop on Geosynthetic Clay Liners, D.E. Daniel and H.E. Scranton, authors,
   EPA/600/R-96/149, EPA National Risk Management Research Laboratory,
   Cincinnati, OH, pp. F1-F34.
Bramlett,  J., Furman, C., Johnson, A., Ellis, W. D., Nelson, H.  and Vick, W. H. (1987),
   "Composition ofLeachates from Actual Hazardous Waste Sites",  U.S. Environmental
   Protection Agency, Office of Research and Development, Cincinnati, OH, EPA/600/2-
   87/043, 113 p.
Campbell, G. S. (1974), "A Simple Method for Determining Unsaturated Hydraulic
   Conductivity from Moisture Retention Data", Soil Science, 117(6), pp. 311-314.
Daniel, D. E. and Koerner, R. M. (1993), "Technical Guidance Document: Quality
   Assurance and Quality Control for Waste Containment Facilities", EPA/600/R-93/182,
   EPA Risk Reduction Research Laboratory, Cincinnati, OH, 305 p.
Dvirnoff, A. H. and Munion,  D. W. (1986), "Stability Failure of a Sanitary Landfill",
   International Symposium on Environmental Geotechnology, Lehigh, Pennsylvania,
   pp. 25-35.
EPA (1989), "Technical Guidance Document: Final Covers on Hazardous Waste Landfills
   and Surface Impoundments", EPA/530/SW-89/047, U.S. Environmental Protection
   Agency, Office of Solid Waste and Emergency Response, Washington, D.C., 39 p.
EPA (1994), "Seminar Publication Design, Operation, and Closure of Municipal Solid
   Waste Landfills",  EPA/625/R-94/008, U.S. Environmental Protection Agency, Center
   for Environmental Research Information, Cincinnati, Ohio, 86 p.
EPRI (1984), "Comparison of Two Groundwater Flow Models-UNSATID and HELP',
   Electric Power Research Institute, Topical Report, EPRI CS-3695, Project 1406-1.
Field, C. R. and Nangunoori, R. K. (1994), "Case Study-Efficacy of the HELP Model: A
   Myth or Reality?", Proceedings of the Waste Tech '94 Conference, National Solid
   Wastes Management Association, Charleston, SC., 9 p.
Ghassemi, M., Crawford, K., and Haro, M. (1986), "Leachate Collection and Gas
   Migration and Emission Problems at Landfills and Surface Impoundments",

                                     5-86

-------
  EPA/600/2-86/017, U.S. Environmental Protection Agency, Hazardous Waste
  Engineering Research Laboratory, Cincinnati, Ohio, 206 p.
Gibbons, R. D., Dolan, D., Keough, H., O'Leary, K., and O'Hara, R. (1992), "A
  Comparison of Chemical Constituents in Leachate From Industrial Hazardous Waste
  & Municipal Solid Waste Landfills", Proceeding of Fifteenth Annual Madison Waste
  Conference, University of Wisconsin - Madison, Sep, pp. 251-276.
Giroud, J. P. and Bonaparte, R. (1989a), "Leakage Through Liners Constructed with
  Geomembranes - Part I. Geomembrane Liners", Geotextile and Geomembranes, Vol.
  8, No. 1,  pp. 27-67.
Giroud, J. P. and Bonaparte, R. (1989b), "Leakage Through Liners Constructed with
  Geomembranes - Part II. Composite Liners", Geotextile and Geomembranes, Vol. 8,
  No. 2, pp. 77-111.
Giroud, J. P., Badu-Tweneboah, K., and Bonaparte, R. (1992),  "Rate of Leakage
  Through a Composite Liner Due to Geomembrane Defects",  Geotextile and
  Geomembranes, Vol.  11, No. 1, pp. 1-28.
Gross, B. A., Bonaparte, R., and Giroud, J.P. (1990), "Evaluation of Flow From Landfill
  Leakage Detection Layers", Proceedings of Fourth International Conference on
  Geotextiles, Vol. 2, The Hague, pp. 481-486.
Kenter, R. J., Schmucker, B. 0., and Miller, K. R. (1997), "The Day the Earth Didn't
  Stand Still: The Rumpke Landfill", Waste Age, Mar, pp. 66-81.
Khire, M. V., Benson,  C. H., and Bosscher, P. J. (1994), "Final Cover Hydrologic
  Evaluation-Phase III", Environmental Geotechnics Report No. 94-4, University of
  Wisconsin-Madison, Wl, 142 p.
Kmet, P., Mitchell, G., and Gordon, M. (1988), "Leachate Collection System Design and
  Performance - Wisconsin's Experience", Proceedings of ASTSWMO National Solid
  Waste Forum on Integrated Municipal Waste Management, Lake Buena Vista,
  Florida.
Lane, D. T., Benson,  C. H., and Bosscher, P. J. (1992), "Hydrologic Observations and
  Modeling Assessments of Landfill Covers", Final report No. 92-10, University of
  Wisconsin-Madison, 406 p.
Lange, D. A., Cellier, B. F., and Dunchak, T. (1997), "A Case Study of the HELP Model:
  Actual Versus Predicted Leachate Production Rates at a MSW Landfill in North-
  eastern Ohio", Proceedings of the SWAN A Conference, Sacramento, CA, 18 p.
NUS (1988), "Draft Background Document Summary of Data on Municipal Solid Waste
  Landfill Leachate Characteristics "Criteria for Municipal Solid Waste Landfills" (40
  CFR Part 258) Subtitle D of Resource Conservation and Recovery Act (RCRA)", U.S.
  Environmental Protection Agency, Office of Solid Waste, Washington, D.C.,
  EPA/530-SW-88-038, Jul.
Oweis, I. S.  (1985), "Stability of Sanitary Landfills", Geotechnical Aspects of Waste
  Management, seminar sponsored byASCE Metropolitan Section, New York.
Peters, N., Warner, R. S., Coates, A. L,  Logsdon, D. S., and Grube, W. E.  (1986),
  "Applicability of the HELP Model in Multilayer Cover Design:  A Field Verification and
  Modeling Assessment", Land Disposal of Hazardous Waste-Proceedings of the 1986
  Research Symposium, U.S. Environmental Protection Agency, Cincinnati, OH.

                                    5-87

-------
Peyton, R. L. and Schroeder, P. R. (1988), "Field Verification of HELP Model for
   Landfills", Journal of Environmental Engineering, ASCE, Vol. 114, No. 2, pp. 247-
   269.
Peyton, R. L. and Schroeder, P. R. (1993), "Water Balance for Landfills", Geotechnical
   Practice for Waste Disposal, D.E. Daniel, Ed., Chapman & Hall, London, pp. 214-
   243.
Richardson, G. and Reynolds, D. (1991), "Geosynthetic Considerations in a Landfill on
   Compressible Clays", Proceedings of Geosynthetics '91, Atlanta, Georgia, 1991, Vol.
   2, pp. 507-516.
Richardson, C. W. and Wright, D. A. (1984), "WGEN: A Model for Generating Daily
   Weather Variables", ARS-8, U.S. Department of Agriculture, Agricultural Research
   Service, 83 p.
Ritchie, J. T. (1972), "A Model for Predicting Evaporation from  a Row Crop with
   Incomplete Cover", Water Resources Research, 8(5), pp. 1204-1213.
Schmucker, B. 0. and Hendron, D. M. (1997), "Forensic Analysis of 9 March 1996
   Landslide at the Rumpke Sanitary Landfill, Hamilton County, Ohio", Slope Stability in
   Waste Systems, seminar sponsored by ASCE Cincinnati and Toledo Sections, Ohio.
Schroeder, P.  R., Lloyd, C. M., and Zappi, P. A. (1994a), "The Hydrologic Evaluation of
   Landfill Performance (HELP) Model, User's Guide for Version 3", U.S. Environmental
   Protection Agency, Office of Research and Development, Washington, D.C., Report No.
   EPA/600/R-94/168a.
Schroeder, P.  R., Dozier, T. S., Zappi, P. A., McEnroe, B. M., Sjostrom, J. W., and Peyton,
   R. L. (1994b), "The Hydrologic Evaluation of Landfill Performance (HELP) Model
   Engineering Documentation for Version 3", U.S. Environmental Protection Agency,
   Office of Research and Development,  Washington,  D.C., Report No. EPA/600/R-
   94/168b, 116 p.
Soong, T. Y. and Koerner, R. M. (1997), "The Design of Drainage Systems Over
   Geosynthetically Lined Slopes", GRI  Report #19,  Geosynthetic Research Institute,
   Philadelphia, Pennsylvania, 88 p.
Stark, T.  D. and Evans, W. D. (1997), "Balancing Act", Civil Engineering, Aug,  pp. 8A-
   11 A.
Thompson, F. and Tyler S. (1984),  "Comparison of Two Groundwater Flow Models
   (UNSAT1D and HELP) and their Application to Covered Fly Ash Disposal Sites",
   EPRI Document Series, Electric Power Research  Institute, Palo Alto, California, Aug.
Udoh, F.  D. (1991), "Minimization of Infiltration  Into Mining Stockpiles Using Low
   Permeability Covers", Dissertation Proposal,  Dept. of Materials Science and
   Engineering,  Mining Engineering Program, University of Wisconsin-Madison.
UDSA-SCS (1985), "Hydrology", Section 4 in National Engineering Handbook, U.S.
   Government  Printing Office, Washington, D.C.
                                     5-88

-------
                                 Chapter 6
                    Summary and Recommendations

6.1  Rationale and Scope of Chapter
The study discussed in this research report addressed three important areas of waste
containment system design and performance, namely:

   •   geosynthetic materials (puncture protection of GMs using GTs, wave behavior in
       HOPE GMs, plastic pipe behavior under high overburden stresses, and service
       life prediction of GTs  and GMs);
   •   natural soil materials  (slope stability of final cover systems with GCLs, kfieid of
       natural soil CCLs and soil-bentonite admixed CCLs, and hydraulic performance
       of CCLs in final cover systems); and
   •   field performance (LCRS and LDS flow quantities and chemical quality at
       landfills, assessment of EPA HELP computer code as a design tool using LCRS
       flow rate data, and lessons learned from waste containment problems at
       landfills).

All three areas were addressed through multiple tasks, each important in its' own right,
but also complementary to the other tasks because of the interrelationships between
waste containment system components. The ultimate goals of this study were to
assess the field performance of waste containment systems and to develop
recommendations for further improving the performance of these systems in
comparison to the current state-of-practice.

This chapter presents a summary of the tasks conducted for this study and provides
recommendations on practices to further improve the performance of waste
containment systems. These recommendations were developed, in part, using the
results of the various tasks.  Some, however, go beyond the scope of this study and are
offered by the authors with the understanding that the current level of "good" field
performance can be further improved within current material, design, testing, and
installation technology and practices.

6.1.1  Geosynthetics
As discussed  in Chapter 1, geosynthetics, including GMs, GTs, GNs,  GCs, plastic pipe,
and GCLs, are used in waste containment systems for a variety of functions.  Most
modern waste containment systems contain one or more geosynthetic components.
Notwithstanding their broad use, issues related to geosynthetic materials persist.
Indeed, the relative newness of these materials compared  to natural soil construction
materials requires that they continue to be studied and evaluated.  Chapter 2 of this
report described the results of the geosynthetic-related tasks of this research project.
These tasks addressed:
                                   6-1

-------
   •   protection of GMs from puncture using needlepunched nonwoven GTs;
   •   behavior of waves in HOPE GMs when subjected to overburden stress;
   •   plastic pipe stress-deformation behavior under high overburden stress;
   •   service life prediction of GTs; and
   •   service life prediction of GMs.

Key findings of the geosynthetic-related tasks are given below:

   •   Needlepunched nonwoven GTs can provide adequate protection of GMs against
       puncture by adjacent granular soils. A design methodology for GM  puncture
       protection was developed from the results of laboratory tests and was
       presented.
   •   Temperature-induced waves (wrinkles) in GMs do not disappear when the GM is
       subjected to overburden stress (i.e., when the GM is covered with soil), rather
       the wave height decreases somewhat, the width of the wave decreases  even
       more (i.e., the height-to-width ratio (H/W) of the wave increases), and the void
       space beneath the wave becomes smaller.  Residual stresses in HOPE  GMs
       installed in the field may be on the order of about 1 % to 22% of the  GM's short-
       term yield strength in the vicinity of GM waves, with higher residual  stresses
       associated with higher H/W values.  Significant residual stresses can reduce the
       GM service life.  The relationship between GM type, residual stress magnitude,
       and service life requires further investigation.
   •   If GM waves after backfilling are to be avoided,  light-colored (e.g., white) GMs
       can be used, GMs can be deployed and seamed without intentional slack,  GMs
       can be covered with an  overlying light colored temporary GT until backfilling
       occurs, and backfilling can be performed only in the coolest part of the day or
       even at night.
   •   Based on finite element modeling  results, use of the Iowa State formula  for
       predicting plastic pipe deflection under high overburden stress is reasonable.  In
       comparison to the FEM  predictions, the Iowa State formula overestimated  pipe
       vertical deflection under short-term conditions (which is conservative) and
       slightly underestimated pipe vertical deflection under long-term conditions
       (which is slightly unconservative, but typically accommodated by the
       incorporation of a factor of safety).
   •   PP GTs are slightly more susceptible to UV degradation than PET GTs,  and
       lighter weight GTs degrade faster than heavier GTs.
   •   GTs that are partially degraded by UV light do not continue to degrade when
       covered with soil, i.e., the degradation process is not auto-catalytic.
       Nonetheless, good practice dictates that GTs be covered with overlying
       protective materials in a timely manner to minimize exposure. Also, GTs should
       be protected  from exposure prior to installation (i.e., by keeping the  GT rolls in
       opaque bags).
                                   6-2

-------
   •   Buried HOPE GMs have an estimated service life that is measured in terms of at
       least hundreds of years. The three stages of degradation and approximate
       associated times for each as obtained from the laboratory testing program
       described in this report are: (i) antioxidant depletion (« 200 years), (ii) induction
       (« 20 years), and (iii) half-life (50% degradation) of an engineering property («
       750 years).  It is noted that these durations were obtained from the extrapolation
       of a number of laboratory tests performed under a limited range of conditions. It
       is recommended that additional testing be performed under a broader range of
       conditions to develop additional insight into the ultimate service life of HOPE
       GMs, and other types of GMs as well.

6.1.2 Natural Soils
CCLs, including those constructed from natural clay soils and those constructed from
soil-bentonite mixtures, have long been used in waste containment systems as
hydraulic barriers to inhibit liquid migration from the waste management unit. Either
used alone, or with a GM component in the form of a GM/CCL composite liner, CCLs
form an essential part of many liner systems and final cover systems. Other natural soil
materials used in liner and final cover systems include sands and gravels used for gas
conveyance systems or liquid drainage and collection systems, and soil layers used for
filtration, separation, or protection. Notwithstanding the widespread use of natural soil
materials in liner systems and final cover systems, questions and issues persist relative
to their use.  Several of these questions and issues were investigated, and the results
were reported  in Chapters 3 and 4 of this report. The subject areas that were
addressed are:

   •   slope stability of GCLs in final cover systems, as assessed from field test plots;
   •   kfieid of low-permeability natural soil CCLs;
   •   kfieid of admixed (soil-bentonite) CCLs; and
   •   CCL hydraulic performance in final cover systems;

These topics were selected on the basis of past research indicating areas where
additional insight was required,  or on the basis of concerns developed from relatively
recent field experience. Key findings of the natural soils related tasks are given below:

   •   Slope stability monitoring of final cover system test plots incorporating GCLs
       demonstrated acceptable performance for test plots constructed on 3H:1V
       slopes, but several of the test plots constructed on 2H:1V slopes failed.
       Importantly, for internally-reinforced GCLs, these failures were not due to
       inadequate internal strength, but inadequate interface strength.  Clearly, proper
       characterization of GCL interface shear strength is an important design step.
   •   The key to achieving low kfieid for natural soil CCLs is to ensure that 70 to 80%,
       or more, of the field-measured compaction (w vs. yd) points lie on or above the
       line of optimums for the particular CCL being placed.
                                    6-3

-------
   •   Practically no correlation was found between kfieid and frequently measured soil
       characterization parameters, such as plasticity index and percentage of clay,
       indicating that natural soil CCLs can be constructed with a relatively broad range
       of soil materials.
   •   Compaction density appears to be more significant than water content for
       achieving low kfieid in soil-bentonite liners.
   •   The long-term hydraulic performance of low-permeability (i.e., kfieid ^ 10~7 cm/s)
       CCLs in final cover systems may not be good in light of the effects of
       desiccation, freeze-thaw, root penetration, animal intrusion, and subsidence.

6.1.3  Field Performance
The premise of this portion of the study was that "modern"  waste containment systems
have been installed for up to a decade or more allowing for an assessment of their field
performance. Information on actual field performance can  be used to evaluate how
waste containment systems are performing now, and for extrapolation of their long-term
performance. Chapter 5 presented a discussion of the following specific topics related
to the performance of waste containment systems for modern landfills:

   •   evaluation of published information on field performance;
   •   collection and analysis of liquids management data;
   •   identification and assessment of problems;  and
   •   assessment of the EPA HELP model as a tool for LCRS design.

These topics were selected to develop an improved understanding of the actual field
performance of modern landfill liner systems, and, to the extent possible, to develop
answers to the questions identified in Section 5.1.1 of this report.  Key findings of the
field performance tasks are:

   •   LDSs from double-lined landfills will almost always  exhibit flow.  Much of this
       flow may be from sources other than primary liner leakage, particularly  in the
       time frame just after construction when construction water can be a significant
       source, and for GM/CCL composite liners, following waste placement when
       consolidation water from the CCL can be a major source.
   •   Average monthly active-period LDS flow rates for cells with HOPE GM primary
       liners  constructed with CQA (but without ponding tests or electrical leak location
       surveys) will often be less than 50 Iphd, but occasionally in excess of 200 Iphd.
       These flows are attributable primarily to liner leakage and, for cells with sand
       LDSs, possibly construction water.  Average monthly active-period LDS flow
       rates attributable to leakage through GM/GCL primary liners constructed with
       CQA will often be less than 2 Iphd, but occasionally in excess of 10 Iphd.
       Available data suggest that average monthly active-period LDS flow rates
       attributable to leakage through GM/CCL and GM/GCL/CCL primary liners
       constructed with CQA are probably similar to those for GM/GCL primary liners
       constructed with CQA.
                                   6-4

-------
•   Single liner systems with GM liners (installed on top of a relatively permeable
    subgrade) should not be used in applications where a true hydraulic efficiency
    above 90% must be reliably achieved, even if a thorough CQA program is
    employed. In these cases, single-composite liner systems or double-liner
    systems should be used. An exception to this may be made for certain facilities
    where electrical leak location surveys or ponding tests are used to identify GM
    defects and the defects are repaired. Higher true hydraulic efficiencies of 99%
    to more than 99.9% can be achieved by GM/GCL, GM/CCL, and GM/GCL/CCL
    composite liners constructed with good CQA.
•   Based on the existing data, GM/CCL and GM/GCL/CCL composite liners are
    capable of substantially preventing leachate migration over the entire period of
    significant leachate generation for typical landfill operation scenarios (i.e., for a
    landfill cell filled over a number of years, that does not undergo leachate
    recirculation or disposal of liquid wastes or sludges,  and that is capped with a
    final cover system designed to minimize percolation into the landfill; based on
    our existing understanding of their performance capabilities, these types of
    composite liners are capable of substantially preventing leachate  migration for a
    much longer  period, although field performance data of the type presented in
    this report do not yet exist for this longer period.
•   LCRS flow rates during operations (i.e., the initial  and active periods of
    operation) can vary significantly between landfills  located in the same
    geographic region and accepting similar wastes.  Large variations in flow rates
    (e.g., one order of magnitude difference) can even occur between cells at the
    same landfill.
•   LCRS flow rates were highest at the beginning of cell operations and decreased
    as waste thickness  increased and  daily and intermediate covers were  applied to
    the waste. Leachate generation rates decreased, on average, by a factor of four
    within one year after closure and by one order of magnitude two to four years
    after closure. Within nine years of closure, LCRS flow rates were negligible for
    the landfill cells evaluated in this study.
•   MSW cells produced, on average,  less leachate than HWand ISW cells.
•   For cells of a given waste type,  rainfall fraction (RF) values were highest in the
    northeast U.S. and  lowest in the west.
•   In general, HW landfills produced the strongest leachates and coal ash landfills
    produced the weakest  leachates.  MSW ash leachate was more mineralized
    than MSW leachate and the other  ISW leachates.
•   The solid waste regulations of the  1980s and 1990s have resulted in the
    improved quality of  MSW and HW landfill leachates.
•   The EPA HELP computer model, when applied  using an appropriate simulation
    methodology and an appropriate level of conservatism, provides a reasonable
    basis for designing LCRSs and  sizing leachate management system
    components. Use of the HELP  model for these purposes can  be enhanced
    through calibration to leachate generation rates at other landfills in the region
    and through parametric analyses that consider the potential range of values for
    key input parameters (e.g., initial moisture contents of waste).  Due to the
                                6-5

-------
       complexity and variability of landfill systems, however, the model will generally
       not be adequate for use in a predictive or simulation mode, unless calibration  is
       performed using site-specific measured (not default) material properties and
       actual leachate generation data.
   •   The frequency of occurrence of design, construction, and operational problems
       at landfills is significant.  The most common types of problems encountered
       involved liner system and final cover system slope stability.  Almost all of the
       problems were detected  shortly after they occurred, and  environmental impact
       due to the problems was only identified at one facility, which has since been
       remediated.  The main impacts of the  problems were interruption of waste
       containment system construction and  operation, increased maintenance, and
       increased costs.  Importantly, all of the problems identified in this investigation
       could have been prevented  using available design approaches, construction
       materials and procedures, and operation practices.

In light of the significant findings in each of the three areas of investigation,  it is obvious
that landfills are complex structures that require careful and  thorough design, testing,
construction and operation/maintenance. Procedures exist to avoid the types of issues
and problems identified in this report. Unfortunately, as most clearly demonstrated by
Appendix F of this report, landfill industry personnel  do not always utilize adequate
design, testing, construction, and operation/maintenance practices. The authors feel
strongly that current practices can and should be improved.  In the next four sections  of
this report, the authors highlight a number of areas related to landfill design,
construction, and operation where they believe practice improvement can be achieved
using readily available technology.

6.2  Liner Systems
Liner systems for the containment of solid waste consist of at least a low-permeability
barrier (liner) and an overlying LCRS. Depending on the nature  of the waste (and
obviously the pertinent regulations)  a single-liner system or a double-liner system with a
LDS between the two liners may be required.  In all cases, geosynthetics and/or natural
soils are typically utilized for the  liners, drainage layers, or both.  The design of these
multi-component and multilayered systems (see Figures 1-1 and 1-3) requires the
application of sophisticated engineering analysis methods. These systems also require
careful construction methods and CQA if they are to function as  intended.  This section
of the report is intended to  highlight several of the more important challenges faced by
engineers and contractors  in designing and constructing these systems. It is noted that
some of these challenges go beyond the tasks directly evaluated in this project;
however, these challenges are identified because they are important to waste
containment system performance.
                                    6-6

-------
6.2.1   Construction Quality Assurance
CQA has been shown to be of direct benefit in minimizing the potential leakage through
liner systems. This finding was originally put forth by Bonaparte and Gross (1990) on
the basis of sparse data and has been reinforced with the considerable additional data
generated since that time, including data presented in this study. Considerable
guidance exists for the development and implementation of liner system and cover
system CQA plans. Among the many requirements for such plans, the authors make
note of the following:

   •   soil and geosynthetic material conformance with the project specifications;
   •   proper pre-conditioning and placement of CCL lifts;
   •   proper compaction moisture content and density of CCLs;
   •   protection of CCLs from desiccation and freezing;
   •   placement of GMs without excessive waves and covering or backfilling the GMs
       in a manner that minimizes the trapping of waves; the goal of these measures is
       intimate contact between the GM and the underlying CCL or GCL;
   •   prevention of premature GCL hydration;
   •   inspection of GM seams, including nondestructive and destructive testing; and
   •   protection of GMs from puncture by adjacent materials or equipment.

6.2.2  Liner System Stability
This category of stability involves the liner system prior to waste placement.  The main
concern regarding liner system stability is for natural soils (particularly sand and gravel
drainage soils) or geosynthetics (particularly GTs and  GNs) to slide on underlying
geosynthetic surfaces. Sliding of drainage soils or sliding of drainage soils and GT
cushions on underlying GMs is unfortunately too common. The  instability is induced by
low shear  strength interfaces, steep and/or long slopes, equipment loads, seepage
forces, and/or seismic forces.  An area requiring particular attention is at access ramps
into below-grade landfills. These ramps are needed for operations, but are sometimes
overlooked in the assessment of landfill cell slope stability.  In some cases, ramps have
been installed by landfill operations personnel, without an evaluation of their effect on
liner system stability.  Another type of liner system  stability problem that requires careful
attention is sliding of GM layers on underlying CCLs or GCLs prior to waste placement.

Design of  liner systems for adequate slope stability is well within the design state-of-
practice. The available technical literature contains more than adequate  information to
design liner systems to be stable (see for example, Giroud and Beech, 1989; Koerner
and Hwu,  1991; Giroud et al., 1995; and Koerner and Soong, 1998). However, in the
authors' experience, the available methods are often not adequately utilized in design.
For example, it is not uncommon for seepage forces to be inadequately addressed
during the design process. Another significant design issue involves the  inadequate
characterization of interface shear strengths, apparently due to insufficient effort
                                   6-7

-------
expended in the laboratory evaluation of these strengths. Testing must be performed
under both project-specific and material-specific conditions, with considerable attention
given to the many variables that can influence the interface shear test results (e.g.,
boundary conditions, normal stresses,  hydration times, moisture conditions, and
displacement rates).  A number of papers, including those by Dove et al. (1997), Eid
and Stark (1997), Gilbert et al. (1997),  Sharma et al.  (1997), Sabatini et al. (1998),
Breitenbach and Swan (1999), and Sabatini et al. (2001), discuss variables that can
influence test results.

It is somewhat fortunate that many liner system slope stability problems can be repaired
at relatively small cost and with no environmental impact. This is particularly true of
slides that occur above the GM component of the liner system.  However, these facts
certainly do not justify a  less rigorous or careful design approach,  and overall
improvement in the rigor with which some owners and engineers address this design
issue is warranted.

6.2.3  Waste Stability
Of potentially greater significance than instability of the liner system before waste is
placed is a failure that occurs after waste has been deposited on top  of the liner system.
The Kettleman Hills landfill failure (Byrne et al.,  1992) is perhaps the  best known of this
type of occurrence. Design to resist this type of instability requires that the design
engineer specify acceptable waste configuration (e.g., intermediate slope angles) and
waste placement procedures, in addition to appropriately using slope stability analysis
methods and selecting liner system interface shear strengths.  For many facilities, waste
placement operations will need to be carefully sequenced.  Canyon-type landfills and
landfills built on soft foundation soils represent two classes  of facilities for which waste
mass stability deserves particular attention.

As with liner system stability, the technical analysis for waste mass stability is within the
state-of-practice, relying principally on  limit equilibrium slope stability  methods
developed in geotechnical engineering. The validity of the analysis is dependent  on the
choice of analysis methods, waste geometry and properties, interface shear strengths,
and moisture conditions in the landfill.  Particularly important with respect to waste
placement operations are the slope of the exposed surface of the waste, distance of this
exposed surface from the liner system  sideslopes, height of the waste, waste density,
and waste shear strength.  Discussion  of solid waste shear strengths to use in design
can be found in Kavazanjian et al. (1995). To help assure waste stability, the authors
recommend that the operations plans developed for landfills provide detailed criteria for
waste placement so that the landfill operator does not unknowingly fill the facility in a
potentially unstable manner.
                                    6-8

-------
Of particular importance in choosing waste and interface shear strengths is deformation
compatibility. It must be recognized that the amounts of deformation needed to
generate peak shear strengths in waste and along geosynthetic interfaces  are very
different. As discussed by Byrne (1994), Stark and Poeppel (1994), Gilbert et al.
(1997), and Sabatini et al. (2001), careful consideration must be given to the shear
strength deformation conditions used in design (i.e., peak, large displacement, or
residual).

It is interesting to note that several of the larger waste failures reported in the literature
occurred after periods of high rainfall, which had the effect of temporarily increasing the
density of the waste (Reynolds, 1991).  High rainfall can also impose seepage forces,
which will decrease stability accordingly.

Also important in some cases is seismic stability of the waste mass. While the
performance of several lined earthquakes in the 1994 California Northridge earthquake
was very good (Matasovic et al., 1995; Matasovic and Kavazanjian, 1996)  more needs
to be learned about this subject, particularly with respect to the seismic response of the
landfill and the determination of the acceptable magnitude of seismically-induced liner
system deformation. With respect to this latter criterion, it is the authors' experience
that design engineers often select a seismic deformation criterion of 150 to 300 mm
based on Seed and Bonaparte (1992).  However, these values may not be appropriate
in all applications. Careful consideration should be given to selection of an acceptable
level of deformation for design.  For example, all other factors being equal, a lower
allowable deformation should be used if the critical  interface is below the GM
component of the liner system (because excessive deformation would cause the GM to
rupture) than above  it.  Guidance on the seismic design of landfills can be found in
Richardson et al. (1995), Anderson and Kavazanjian (1995), and Kavazanjian (1998).

6.2.4  Performance of Composite Liner
For over a decade it has been known through theoretical analyses, laboratory tests, and
limited field data that composite liners are superior to either GMs alone or CCLs alone
for the containment of leachate or other liquids (Brown et al.,  1987; EPA, 1987; Giroud
and Bonaparte, 1989a,b; Bonaparte and Gross, 1990; Bonaparte and Othman, 1995).
This report has presented significant new field data that confirms the very good
performance characteristics of GM/GCL, GM/CCL,  and  GM/GCL/CCL composite liners
versus current types of single liner materials.

As discussed in Section 1.4.1.4, the basic premise of using a composite liner is that
leakage through a hole or defect in the GM upper component is impeded by the
presence of  a CCL or GCL lower component.  The GM  improves the performance of the
composite liner relative to that for a CCL or GCL alone by greatly limiting the portion of
the CCL or GCL exposed to leachate, and, for CCLs, lowering the potential for
                                   6-9

-------
desiccation cracking. Another benefit derived by using a composite liner is reduced
potential for diffusive transport through the liner. Diffusion is not an important transport
mechanism for inorganic ions through GMs.  However, as shown by Rowe (1998) and
others, diffusion rates of certain organic contaminants through GMs can be significant
when the concentrations of these contaminants are relatively high (i.e., diffusion through
GMs is generally not a concern at MSW landfills, but may be a concern at landfills or
impoundments where the liquid above the liner has relatively high concentrations of
volatile organic compounds). A CCL, and to a lesser extent GCL, component of a
composite liner will help retard organics that diffuse through the GM. Analysis methods
to design composite liners to account for diffusive transport are given by Rowe (1998).
If diffusion is the primary concern for a specific project, a CCL is preferred to a GCL as
the soil component of the composite liner.  To maximize retardation potential, adequate
thickness of CCLs is more important than low permeability.  One  approach to achieve a
composite liner with both low advective transport and low diffusive transport potential is
to specify a GM/GCL/CCL composite liner.  Giroud et al. (1997) present equations to
evaluate advective leakage rates through composite liners containing GCLs. While a
GM/GCL/CCL composite liner has advantages with respect to minimizing contaminant
transport potential, it may also create challenges with respect to slope stability factors of
safety. Shear strengths for both the GM/GCL interface and GCL/CCL interface require
careful evaluation.

6.2.5 Single vs. Double Liner System
As discussed in Section 1.2, federal regulations under Subtitle C  of RCRA require
permitted HWfacilities to be underlain by double-liner systems with leak detection
capability.  Also as discussed in Section 1.2, federal minimum design criteria for MSW
landfills include a single composite liner system. Several states have gone beyond
these minimum criteria for MSW landfills by requiring double-liner systems. A 1998
survey of 43 states has shown that for MSW landfills:

   •   31  (72%) states require single liner systems;
   •   6 (14%) states require double liner systems; and
   •   6 (14%) states provide options for the use of either a single liner or double liner
       system.

This survey highlights the differences in perspective (due to regional political
differences, population attitudes, hydrogeology, climate, drinking  water resources, and
other factors) between states as to the minimum requirements for liner systems at
RCRA Subtitle D landfills. These differences are even greater when it is realized that
the federal Subtitle D regulations contain both federal minimum design criteria and
performance-based criteria.  The performance-based critera require technical
demonstrations that are often made using the EPA HELP and MULTIMED computer
models, which do not address the potential for the migration of any landfill-generated
                                   6-10

-------
gas through the liner system. With respect to selection of the type of liner system for a
specific project, the authors offer the following thoughts:

   •   Caution should be exercised in using the EPA HELP  model to make a technical
       demonstration that the Subtitle D performance standard can be achieved with a
       liner system less (e.g., without a GM) than the federal minimum design criteria.
       Input parameters to the model can be selected to demonstrate a lesser potential
       for leachate generation than actually exists.  For example, the discussion in
       Chapter 5 of this report indicated that modeled leachate generation rates are
       sensitive to the assumed initial moisture content of the waste.  Because of the
       sensitivity of the HELP model results to the input parameters, when the model is
       used to make a technical performance demonstration, the model should be
       calibrated against  data (i.e., LCRS flow rates) from lined landfills in the same
       geographic area.  In addition,  the potential for  landfill  gas impacts to
       groundwater should also be considered as part of the technical demonstration.
   •   Based on the landfill operation data presented in this report, Subtitle D single-
       composite liner systems meeting federal  minimum design criteria can achieve a
       very high hydraulic efficiency and are capable  of preventing adverse impacts to
       groundwater. This conclusion is consistent with the previous conclusion
       reached by EPA regarding the performance capabilities of liner systems meeting
       federal minimum design criteria.
   •   Caution should be exercised in substituting a GCL alone for the CCL as the low-
       permeability soil component of a Subtitle D single-composite liner on the base of
       a landfill. While the hydraulic efficiency of a GM/GCL composite liner is as
       good, or better, than a GM/CCL composite liner, the GM/GCL  composite liner is
       more susceptible to diffusive transport  (Rowe,  1998)  and puncture than the
       GM/CCL composite liner. These concerns are less important for sideslope
       areas of the landfill where leachate heads are  lower;  thus, a GM/GCL composite
       liner is more likely  be appropriate for sideslopes than for base  areas from a
       hydraulic perspective. Also, a GM/GCL/CCL composite liner may be an
       effective low-permeability soil component for a single-composite liner.  In this
       case, it may be acceptable to specify a maximum hydraulic conductivity on the
       order of 1 x 1f>5 cm/s for the CCL of a  three-component composite liner used at
       MSW landfills.
   •   There may exist situations for MSW landfills where a  double-liner system would
       be preferred to a liner system meeting  the federal minimum design criteria.  In
       addition to the obvious situation where a  state regulation requires use  of a
       double-liner system, the project conditions favoring selection of a double-liner
       system include: (i) sites with especially vulnerable hydrogeology; (ii) sites where
       groundwater cannot be reliably monitored due to the  presence of complex
       hydrogeology, karst,  or other factors; and (iii) sites where, for whatever reason,
       a higher degree of reliability/redundancy  is required of the liner system than can
       be achieved by the Subtitle D federal design criteria.  In some  cases, it may be
       desirable to use a  double-liner system  beneath the base of the landfill, and,  for
       cost-effectiveness, a single-composite liner system beneath the sideslopes.
                                   6-11

-------
   •   The authors endorse a design strategy of providing additional protection at
       critical locations in the landfill. For a landfill underlain by a single-composite
       liner system, this strategy might take the form of installing a GM/GCL/CCL liner
       beneath each landfill sump. This design feature adds little cost to the overall
       project, but significant benefit in terms of design reliability and redundancy at the
       project's most critical location.
   •   Where double-liner systems are used, the authors prefer that the secondary
       liner be a GM/CCL or GM/GCL/CCL composite to a GM alone or a GM/GCL
       alone. Furthermore, the authors prefer a GM/GCL composite liner to either a
       GM alone or GM/CCL composite liner as the primary liner component of the
       double-liner system.  The preferred type of composite primary liner is much
       easier to construct on top of the LDS and secondary liner than is a GM/CCL
       composite primary liner and it minimizes LDS flow rates compared to the rates
       associated with the other types of primary liners considered in this report.
   •   Notwithstanding the specificity of the recommendations given above, the authors
       do endorse the use of creative thinking and good engineering to develop better-
       performing, more cost-effective liner systems.  The key to this approach,
       however, should be good engineering, and not, for example, manipulation of the
       HELP model or any other data or design tool to achieve a pre-conceived desired
       outcome.

6.2.6  Fate of Liner Systems
Of critical importance to the long-term performance of a liner system is the service life of
the GM component of the system.  Both CCLs and GCLs consist of geologic materials.
These materials can be expected to have service lives in excess of the design lives  that
have been defined for MSW, HW,  and LLRW disposal facilities constructed in the
United States.  This conclusion is only valid, however, if these materials stay hydrated
and stable, are adequately protected, and, for GCLs, are not subjected to unacceptable
chemical interactions (i.e., an increase in hydraulic conductivity of bentonite may result
if sodium in the bentonite is replaced by other cations present in the permeant). For the
CCL or GCL component of a composite liner beneath a waste mass these conditions
will often be met.

Perhaps the most important factor governing the service life of GMs is the polymer type,
and resin formulation. Of the variety of choices, HOPE is the most widely used polymer
and the resin formulation includes carbon black and an antioxidant package. As
described in this report and in Hsuan and Koerner (1998), lifetime predictions are
measured in terms of hundreds of years (but not "forever"). However, this report has
also pointed out that additional research in this area is warranted.

Perhaps the most significant issue  related to the use of HOPE GMs is the potential for
premature stress cracking before the end of its design life.  Currently, the notched
constant tensile load (NCTL) test is used with a minimum onset of brittle behavior of 100
                                   6-12

-------
hours.  (This is equivalent to a single point value, per ASTM D5397, of 200 hours). At
the designer's discretion, these values can be increased, and, depending on site-
specific conditions, this is encouraged.  Regarding HOPE formulations, the antioxidant
package included in the formulation is critically important, and specifications should
include a minimum OIT along with a minimum OIT retained value after oven aging and
laboratory simulated UV exposure.

6.3 Liquids Management
The liquids management strategy for a landfill generally refers to all liquids including:

   •   leachate collection and removal at the bottom of the waste mass, above the
       primary liner system;
   •   leakage collection and removal at the bottom of the waste mass between the
       primary and secondary liners;
   •   rainwater collection and removal via the final cover system drainage layer above
       the barrier material;
   •   gas condensate collection and removal via the gas collection piping system;  and
   •   groundwater collection and control via the pore pressure relief system in areas
       of high groundwater.

For the first three systems, drainage layers transmit liquid by gravity to a low point
where the liquid empties into a sump or gravity drain or is discharged from the waste
containment system, in the case of a final cover system drainage layer. In the case of a
sump, the liquid is withdrawn using submersible pumps or bailers.  For a gravity drain,
the liquid flows by gravity through a pipe that penetrates the liner system and
discharges to a storage or treatment system outside the limits of the landfill. From final
cover system  drains, the liquid flows by gravity either as sheet flow to the surrounding
land, or, more typically, into a perimeter stormwater collection and conveyance
structure.  For gas condensate collection and removal systems, liquids collected  in gas
collection piping systems typically drain to a low point in the piping system. From this
location, condensate is usually introduced back into the waste; however, sometimes
condensate is removed from the waste containment system and treated. With respect
to pore pressure relief system, these systems may consist of a series of wells or
perimeter trenches that are pumped to lower the groundwater table or may include a
drainage layer and sump installed beneath the liner system.

These liquid collection and removal systems were discussed in Section 1.4.2 of this
report.  That discussion is not repeated in this chapter.
                                   6-13

-------
6.3.1   Construction Quality Assurance
The authors provide the following commentary on CQA of liquids management systems:

   •   Placement of natural drainage soils on previously placed geosynthetics should
       be done with great care.  Minimum thicknesses of soil should be maintained
       between construction equipment and the underlying geosynthetics.  Only low
       ground pressure equipment should be allowed to spread the soil. Particular
       attention should be focused on preventing accumulation of excess slack and the
       propagation of waves in the underlying GM.
   •   When shallow trenching is conducted within a natural drainage soil in order to
       place plastic pipes above the previously installed GM, the trenching should be
       conducted by hand, with special lightweight backhoes equipped with rubber
       bucket tips, or by an alternate method that will not damage the GM.
   •   Pipe fittings and joints need careful observation during installation to confirm that
       the fittings are mated with one another and are fabricated by the same
       manufacturer.
   •   Sumps, sideslope or vertical riser pipes, and pipe penetrations through GMs
       should be carefully constructed since they are located in areas with the highest
       sustained hydraulic head on the liner. To this end, GM  seams in sumps should
       be limited to the extent possible.  The placement of an enhanced liner in the
       sump (e.g., GM/GCL/CCL instead of a GM/CCL) should also be considered. In
       addition, extra care should be taken to protect the liner around these features  by
       ensuring that required cushion layers are properly placed and that construction
       equipment maintains adequate separation from the GM.
   •   GN placement should be observed to ensure that no gaps exist between roll
       ends and edges and that sufficient plastic ties are used per the specifications.
   •   GNs must be carefully inspected for excess soil particles, fugitive bentonite from
       GCLs, vegetative growth, or other foreign matter, and these materials should be
       removed.  Flushing of water through the installed GN could be considered .
   •   Full coverage of required GT filters over drainage layers should be provided
       whether the drain is natural soil or a GN.
   •   If select waste is to be placed directly over the drainage stone (with no
       intervening filter layer), its placement should have full-time inspection by a CQA
       monitor to assure that the underlying materials are not disturbed.

6.3.2  Potential lor Clogging and Reduction of Flow Capacity
An important question regarding drainage systems is whether or not the system will clog
excessively. The phrase "excessively" is used because all drainage systems will,  over
time, undergo a decrease in their flow capability from the original installation or
manufacture; the issue is to what degree. The authors provide the following
commentary and recommendations on this topic:

   •   For LCRSs, the authors believe that the often cited regulatory value for drainage
       layer hydraulic conductivity of 1 x 10~2 cm/s for natural soil is too low in many
                                   6-14

-------
       applications. Furthermore, a value of 1 x 10~3 cm/s, which  is sometimes
       specified, will almost always be too low.  Hydraulic conductivities at these values
       result in drainage layers with substantial liquid storage (capillary) capacity and
       slow drainage rates. These conditions result in increased hydraulic head on the
       liner and, consequently, increased potential for clogging and leakage.  Design of
       LCRSs should be performed on a site  specific basis, using an adequate factor of
       safety. The soil should be free draining, with few fines, and little or no capillarity.
   •   For design of LCRSs, the HELP model can be an appropriate design tool for
       estimating leachate generation rates (see Chapter 5).  As previously indicated,
       however, HELP model results are sensitive to the input parameters provided.
       The authors believe design engineers  can do much more to calibrate their HELP
       model runs using data from already active landfills in the region.  In this regard,
       design engineers and landfill operators are encouraged to collect and
       disseminate this information.
   •   Landfill LCRS design should include not only an evaluation of leachate quantity,
       but also leachate quality.  This report presents considerable new data on landfill
       leachate characteristics.  From a design perspective, it is important to identify
       conditions (e.g., sludge co-disposal, special waste disposal) that would create a
       leachate with more than usual potential to clog a drainage layer.  For example,
       Koerner et al. (1994) identified leachate with high TSS and/or BOD5  values
       (e.g., above 10,000 to 15,000 mg/l) as a condition requiring special design
       consideration.  Interestingly, in the study of liquids management data described
       in Chapter 5, none of the landfill cells for which leachate chemistry data are
       available had average BOD5 values greater than 5,000 mg/l.
   •   For the internal drainage layer in a final cover system, water is the medium
       being transmitted and clogging of the drainage layer by water is generally not
       considered.  The primary issue for this layer is inadequate drainage capacity
       and the  buildup of seepage forces in the final cover system, leading to slope
       instability. A significant number of seepage-induced final cover system failures
       were identified in Chapter 5. The HELP  model must be used with caution to
       calculate liquid heads in the final cover system drainage layer, as experience
       has shown that these heads may be underpredicted if the peak daily rainfall
       used in the model is too low. Guidance on using the HELP model for this
       purpose is given in the upcoming EPA technical  guidance document titled,
       "Technical Guidance for RCRA/CERCLA Final Covers" (Bonaparte et al., 2002).
       Also, the manual procedure in Koerner and Daniel (1997) can be used to
       estimate liquid heads in the final cover system drainage  layer.

6.3.3 Perched Leachate
Perched leachate (which does not have full hydraulic connection to the underlying
LCRS) can occur as a result of a number of conditions in a landfill.  Excessively clogged
filters above the drainage layer, low-permeability buffer (or protection) soils placed
above the LCRS, low-permeability daily cover, and high moisture content sludges
(industrial or sewage) within the waste mass all can lead to the trapping of moisture in
                                   6-15

-------
pockets within the waste.  The perched leachate can increase the unit weight of the
waste and impact waste stability. Saturated conditions within the zone of perched
leachate will inhibit the generation of landfill gas and reduce the effectiveness of gas
extraction wells in the area. In addition, the "breakout" of perched leachate as seeps
has contaminated nearby surface waters, created odor problems, and killed vegetation.

6.3.4 Fate of Liquids Management Systems
Data in Chapter 5 and Appendix E indicate that in modern landfills, within a few  years to
a decade after a final cover system with a GM  is placed, LCRS flow rates become very
low to negligible. At a minimum, an LCRS should be designed for the anticipated peak
leachate flows during the facility's active life assuming  that some amount of clogging will
occur (the amount being specific to the anticipated leachate and the details of the
design). The authors recommend, however, that the system also be designed to retain
a significant flow capacity even after closure. The rationale for this recommendation is
that it provides the ability to continue to collect  and remove leachate at some future
date, should that need ever become necessary. The future need could arrive out of an
unforeseen, even if improbable, future event such as undetected damage to the final
cover system that allows significant new infiltration into the landfill.  The need could also
arrive out of a planned future event, such as the future use of a closed landfill cell as a
bioreactor or the placement of additional waste as in "piggybacking" operations. Finally,
the authors encourage the increasing tendency of design engineers to design LCRSs to
performance levels better than those required by regulatory  minimums.  Examples
include specifying higher permeability natural drainage materials (e.g., materials having
hydraulic conductivities of at least five to ten times larger than the regulatory minimum
of 1x10-2 cm/s specified by some states for MSW landfills), and designing to a  lower
maximum leachate head than the maximum allowed by regulation (e.g., designing to a
maximum leachate head of 0.03 or 0.1  m, rather than the maximum value  of 0.3 m
allowed by Subtitle C and Subtitle D regulations).

The internal drainage layer above the hydraulic barrier in a final cover system must
function for as long as the final cover system is required. Thus, the design must result
in a stable hydraulic condition within the final cover system over its' design life.   This will
typically require careful selection of the protection soil above a filter or drainage  layer so
that hydraulic equilibrium can be established (i.e., so that particles of the protection soil
are retained by the filter or drainage layer without clogging the layer).

6.4  Final Cover Systems
Conventional final cover systems placed over solid waste typically consist  of a barrier
material, internal drainage layer, cover soil, and surface material. Regulatory
requirements for these conventional systems were discussed in Section 1.2 of this
report. Increasingly, alternative design concepts are being applied to final cover
systems. These alternative concepts include evapotranspiration (ET) or capillary
                                   6-16

-------
barriers, rather than low-permeability hydraulic barriers such as GMs, CCLs, and GCLs.
ET and capillary barrier cover systems are finding increasing use at arid and semi-arid
sites.  These alternative cover systems are discussed in detail in the upcoming EPA
technical guidance document titled, "Technical Guidance for RCRA/CERCLA Final
Covers" (Bonaparte et al., 2002).

Both geosynthetics and natural soils are commonly used in final cover systems. Great
care is required during both design and construction in order to achieve adequate
performance.  While many of the authors' comments in Section 6.2 of this report on liner
systems also apply to the final cover systems, there are several differences between the
two.  For cover systems in comparison to liner systems:

   •   the barrier is meant to keep liquid  out of the waste mass, rather than containing
       liquid within;
   •   the liquid to be managed is infiltrating rainwater (and snow melt) which
       percolates through the cover soil rather than leachate;
   •   upward rising gases from the waste may need to be captured beneath the
       barrier and effectively transmitted  for proper management;
   •   the upward rising gases usually contain volatile constituents from the leachate,
       albeit at low concentrations for landfills (though potentially at higher
       concentrations at remediation sites), thus chemical mass transport and chemical
       compatibility of systems in contact with the gas should be considered;
   •   final cover systems slopes may be relatively steep and long, resulting in
       significant  slope stability design issues;
   •   final cover systems are subjected  to different environmental stresses than liner
       systems; these stresses include freeze-thaw and desiccation-wetting cycles; and
   •   the impact of waste settlements, both total and differential, on final cover system
       integrity should be considered for  proper design of all system components.

Several of the more important issues with respect to design, construction, and
maintenance of landfill final cover systems are discussed below.

6.4.1  Construction Quality Assurance
It seems  intuitive that if proper CQA produces improved performance for liner systems,
the same will be true for final cover systems.  The authors believe that in addition to the
CQA items for liner systems mentioned in Section 6.2.1 of this report, the following
items  require special attention when performing CQA of final cover systems:

   •   evaluation of the subgrade upon which the final cover system is to be placed to
       assure adequate bearing capacity and that buried waste will not damage
       overlying final cover system components;
   •   careful construction according to the design details for connections of GMs and
       GCLs to pipe vents;
                                   6-17

-------
   •   attention to construction details of the gas drainage layer (beneath the barrier)
       connection to the vent system to prevent GM blowouts;
   •   careful construction for connections of cover system internal drainage layers to
       their outlets;
   •   attention to proper location of haul roads and access roads according to lines
       and grades of the plans; and
   •   inspection of the erosion control features to verify that measures have been
       taken to obtain a healthy vegetative cover before the conclusion of construction
       activities.

6.4.2 Compacted Clay Barriers
There are serious concerns with respect to the use of CCLs in final cover systems,
particularly when used alone. These concerns are as follows:

   •   based on the  case histories presented in Section 4.3 of this report, desiccation
       of CCLs is a distinct long-term possibility in almost every final cover system
       application where the CCL  is not covered by a GM; it has been shown that upon
       rewetting, a desiccated CCL does not return to its original low hydraulic
       conductivity;
   •   the freeze-thaw sensitivity of CCLs has been demonstrated in laboratory studies
       whereby the CCL hydraulic conductivity is increased significantly and self-
       healing of the thawed CCL  is not likely (Othman et al., 1994);
   •   as discussed  in Section 4.3 of this report, there are documented cases of
       moisture migration through some CCLs used in final cover systems due to CCL
       degradation;
   •   depending upon the thickness and properties of the final cover system materials
       above the CCL, root penetration of CCLs may occur; these roots can cause the
       development of channels for water migration into the underlying waste;
   •   again depending upon the thickness and properties of the final cover system
       materials above the CCL, burrowing animal intrusion into CCLs is a possibility;
       animal intrusion could lead to relatively large pathways for water migration into
       the underlying waste mass; and
   •   distortion of CCLs due to total and (more importantly) differential settlement of
       the underlying waste may lead to CCL tensile strains that exceed the ultimate
       tensile strain by orders of magnitude; based on studies by Leonards and Narain
       (1963), Ajaz and Parry (1975a,b, 1976), and others, most CCLs tested under
       unconfined or low confinement conditions exhibit failure  at extensional strains of
       0.5% or less.

For these reasons, it  is felt that CCL barriers should typically not be used alone in the
final cover systems of landfills (particularly MSW landfills, which contain wastes that
undergo significant settlement) and that GMs or GCLs, by themselves, or as part  of a
composite cap, will typically be preferable.
                                   6-18

-------
6.4.3 Final Cover System Stability
Notwithstanding the availability of proven slope stability design methods (e.g., Koerner
and Hwu, 1991; Giroud et al., 1995), the sliding of cover soils on underlying
soil/geosynthetic and geosynthetic/geosynthetic interfaces has been a relatively
common problem for landfill final cover systems.  In evaluating final cover system
stability, consideration must be given to a variety of potential destabilizing forces (i.e.,
the gravitational mass of the cover soil,  equipment loadings, seepage forces, and
seismic forces).  As for liner systems, attention to detail by a qualified design engineer
has sometimes been lacking. This attention to detail should apply to the selection of the
input parameters to the slope stability analysis, to the evaluation of seepage, seismic,
and/or equipment forces to be applied to the cover system, the factor of safety used in
the analysis, and the analysis itself.  As for the evaluation of liner system stability, it is
recommended that the shear strengths of cover system materials and interfaces be
evaluated using the results of project-specific laboratory shear tests conducted in a
manner to simulate the anticipated field conditions.

In the experience of the authors, factors that contribute to the observed high frequency
of final cover system slope failures include:

   •   relatively steep slopes with long uninterrupted surfaces; these conditions can be
       mitigated by using flatter slopes, benches, intermediate berms, and/or tapered
       cover soil thicknesses;
   •   equipment loadings, which can be minimized by limiting the ground pressure of
       equipment and orienting the equipment in predetermined (and properly
       designed) paths; the  effect of even  low ground pressure equipment on cover
       system stability should be checked  by the design engineer;
   •   build-up of seepage forces within the drainage layer and/or cover soils due to
       inadequate drainage capacity, which is often the result of not performing a water
       balance for the internal drainage layer  and evaluating the potential for seepage
       forces; if the HELP model is used to estimate seepage forces, considerable care
       is needed in selecting a design storm event and other input parameters that do
       not lead to an underestimate of liquid head buildup in the drainage layer; as
       previously noted, the manual calculation method of Koerner and Daniel (1997)
       can also be used to estimate liquid  heads;
   •   inadequate design of drain transitions and outlets, such that water backs up in
       the drain and causes a buildup of pore pressure within the cover soil mass; and
   •   instability caused  by seismic forces, which is clearly a site-specific situation and
       one requiring careful  design and interpretation; paradoxically,  current regulations
       require seismic design of many MSW landfills but do not do so for HW landfills
       or abandoned landfills.

6.4.4 Cover Soil Erosion
The evaluation of cover soil erosion is also an  important step in the design of a landfill
cover system.  A possible design strategy to avoid seepage forces within a cover soil is
                                    6-19

-------
to use low permeability materials thereby preventing infiltration. Steep slopes, long
uninterrupted slope lengths, and/or poor vegetative cover all tend to increase runoff and
the potential for erosion of cover soils. The authors would like to highlight the following
design considerations with respect to cover soil erosion:

   •   temporary erosion control materials should be used more widely than current
       practice so as to minimize erosion until a healthy stand of vegetation is obtained;
   •   cover system construction projects often conclude in the late fall, with little time
       to establish vegetation before the end of the growing season; this condition
       should be avoided, if possible; it is critical to have a good stand of surface
       vegetation  prior to the end of the growing season if the potential for severe
       winter erosion is to be avoided;
   •   a careful choice of vegetation is critical in providing year-round protection of
       topsoil; a diverse  mixture of native vegetation that closely emulate a selected
       local "climax" community is preferred;
   •   channelization of  runoff is critical, and the design of soft or hard armor surface
       drainage swales and channels is necessary;  let-down chutes represent a
       particular design challenge due to the high water velocities that occur on steep
       slopes; and
   •   erosion control in  arid and semi-arid sites takes a completely different strategy
       than just described; the use of hard armor surfaces, particularly rock riprap, is
       common, with the selection of rock size being an important design output.

6.4.5 Fate of Final Cover Systems
Final cover systems play  a critical long-term role at landfills.  A properly functioning final
cover system will largely eliminate the long-term post-closure leachate generation
potential at solid waste landfills where there is no other input source for liquid (which is
usually the case).  If liquids are prevented from entering the waste mass, there is, in the
long term after the waste has biodegraded and/or stabilized, no significant potential for
continuing leachate and/or gas generation. Thus, the barrier component of the final
cover system should be as durable as the liner component of the liner system. Current
regulations call for a 30-year post-closure care period, and many design engineers
assume that this time frame represents the required design life for the final cover
system barrier.  The authors of this report recommend that the barrier in the final cover
system generally be designed for a longer design life, for example, a 100-year design
life.  The authors also offer the following  observations:

   •   The choice of GM resin in a final  cover system is influenced  by a number of
       factors.  For MSW landfills where settlement potential is significant, high out-of-
       plane deformation capability is a  desirable characteristic. This design criterion
       favors VFPE, fPP and PVC GMs. Long-term durability considerations favor
       HOPE (recall Sections 2.5 and 6.2.6), which does not perform well above the
       pressure rate of 7 kPa/min given in the out-of-place deformation test ASTM
                                    6-20

-------
D5617.  At lower pressure rates, where stress relaxation can occur, the situation
is different but the test is rarely conducted in a slow strain rate or creep mode.
In the current state-of-practice, chemical compatibility is rarely considered for
final cover system GMs since the upper surface of the GM is only exposed to
water infiltrating the cover soil. However, the lower surface of the GM may be
exposed to landfill gas, which invariably contains low concentrations of volatile
components present in the leachate. Thus, chemical  resistance is an issue that
should be considered based on site-specific conditions.
Both durability and chemical compatibility are issues with respect to the
reinforcing fibers or yarns of reinforced GCLs placed on sideslopes.  While the
GCL test plots described in Chapter 3 go far to show the validity of such GCL
reinforcement, GCLs have not been installed for a long enough time to
demonstrate the adequacy of this reinforcement over  a 30 or 100-year time
frame.
The design of internal drainage layers in final cover systems is too often
inadequate, i.e., the flow capacity is too low and outlets and transitions do not
have adequate flow capacity.  The potential for fines migration through the
drainage layer filter is not always considered.  The potential for freezing or other
blockage of the  drainage layer outlets is sometimes not assessed.
The design of final cover systems in seismic impact zones requires careful
consideration. The potential for amplification of free-field ground motions by the
waste mass combined with low shear strength geosynthetic interfaces makes
seismic performance an important consideration. EPA guidance (Richardson et
al. 1995) and Anderson and Kavazanjian (1995) provide procedures for
evaluating the potential for seismically-induced final cover systems
deformations. Considerations applicable to seismically-induced deformations of
liner systems (discussed in Section 6.2.3) are also applicable to final cover
systems.  An  additional consideration for final cover systems is that in high
seismic zones (e.g., near major active faults in  California), it may not be feasible
to design sloping final cover systems containing geosyntnetics to sustain non-
damaging deformations during major earthquakes.  As discussed by
Kavazanjian (1998), in these circumstances, it may be appropriate to design the
final cover system to an acceptable damage criterion. Acceptable damage
levels would be  based on preventing adverse environmental impact, cost of
repair, ease of repair, and any other impacts associated with the damage (e.g.,
loss of serviceability).  This approach would necessitate development of a
detailed post-earthquake response action plan  coupled with financial
assurances to provide the required funds to make the repairs at the time when
they are needed.
The fact that the waste mass is subsiding over time means that sideslope
angles are progressively decreasing. The amount is waste-dependent,  but the
mechanism is one that tends to progressively increase final cover system slope
stability factors of safety.
                            6-21

-------
6.5 Gas Management
The degradation of any putrescible organic fraction of the solid waste in a landfill
produces a number of gases. The condition is mostly applicable to MSW, but other
types of waste may also produce some type of gas by biological or chemical means.
The anaerobic decomposition of MSW produces two principal gases, methane and
carbon dioxide, in roughly equal quantities (i.e., 40 to 50% each of total gas volume)
and much smaller quantities of other gases. The gases produced in a MSW landfill is
generated over a relatively long period of time, especially for landfills at arid sites.
Using the EPA LandGEM computer program with Clean Air Act (CAA) default
parameters for gas generation at a temperate  site and considering an increment of
waste placed in one year, 10% of the gas from this waste is produced within two years
of waste placement, 40%  is produced  within 10 years of waste placement, and 80% is
produced within 30 years.  For an arid site, 20% of the gas from the waste is produced
within 10 years of waste placement, 45% is produced within 30 years, and 80% is
produced within 80 years.  The  gases  move within and from the landfill primarily by
convection, but also by diffusion. Gas emissions from MSW landfills are currently
governed by the RCRA Subtitle D regulations, which address the personal and
fire/explosion aspects of landfill gas, and the CAA regulations, which regulate emissions
of non-methane organic compounds as a surrogate to total landfill gas emissions.
Under the CAA, MSW landfills greater than a certain size must collect and combust their
landfill gas.

Some design engineers collect and vent or extract MSW landfill gases with vertical,
perforated collection wells (typically 5 wells per 2 hectares) without a continuous gas
transmission layer beneath the barrier system. This approach  can be justified if the
waste itself is sufficiently permeable to gas, if the gas wells are relatively closely
spaced, or, perhaps, at arid sites, where gas is generated relatively slowly. With gas
wells, the gas moves within the waste  to the perforations in the pipe and then flows or is
drawn out of the system. Another approach to venting or extracting gas from a  landfill
involves installing a continuous  gas transmission layer beneath the final cover system
barrier layer.  Shallow gas venting or extraction pipes will tie into the gas transmission
layer. Gas collection trenches with periodic vent or extraction pipes represents  a third
approach to gas collection beneath the final cover system.  Also, a combination of
these three gas venting/extraction systems can be used.

In any case (deep wells penetrating the waste, a continuous gas transmission layer
beneath the final cover  barrier layer, and/or collection trenches), the system outlets are
typically plastic pipes extending up through the final cover system.  Gas flow through
the pipes can be either  passive (vented to the atmosphere or flared) or active (collected
through a header using a blower system to create a small vacuum).  Without a gas
management system, gas pressure will build up in the landfill.  Note that with a GM in
the final cover system and relatively small cover soil thicknesses, gas pressures can
                                   6-22

-------
cause GM uplift.  Even if the GM is not physically lifted, positive gas pressure beneath
the GM can lower the effective stress at the interface between the GM and underlying
material (e.g., GCL), thereby reducing interface shear strength and potentially
contributing to a slope failure.

6.5.1   Construction Quality Assurance
As with all aspects of a waste containment system, CQA plays an important role in
achieving acceptable performance of a gas management system. For deep wells, the
number, location and extent of the pipe perforations are important.  Also, the wells must
be kept safely above the liner system beneath the waste.  Several examples exist
where gas well borings have extended into the liner system because of inadequate
survey control and not accounting for landfill settlement. For continuous gas
transmission layers beneath the barrier, continuity is important for either soil or
geosynthetic gas transmission layers.  If the latter, the material is often a GN with GTs
bonded to both sides. The overlapping of the GN along its edges and ends is important
as well as its joining with plastic ties per the specifications.  Both upper and lower GTs
need to be continuous with generous overlaps (often 300 mm) or sewn together to
prevent soil from entering and clogging the GN.

Lastly, the penetration of gas wells or vents through a GM barrier should have tightly
fitting prefabricated boots.  Unlike boots for liner penetrations at the bottom of the
landfill, boots for the final cover system GM must be designed to function while
accommodating the anticipated landfill settlement.  GCL tie-ins have similar
considerations.

6.5.2  Gas Uplift
As indicated above,  when using a GM in an MSW landfill final cover system, gas uplift
pressures will be exerted on the GM unless the gas is efficiently conveyed to the wells,
vents, or collection trenches.  If gas is not adequately managed,  uplift pressure will
either cause GM bubbles (or "wales") to occur displacing the cover soil and appearing at
the surface, or it will decrease the normal stress between the GM and  the underlying
material.  At several facilities, this latter effect has led to slippage of the GM and
overlying cover materials creating high tensile stresses as evidenced by compression
ridges in the cover soil and folding of the GM at the slope toe and tension cracks in the
cover soil near the slope crest.  Three situations need careful design consideration:

   •   if gas removal is by deep wells, the uppermost pipe perforations should be
       effective in capturing gas in the upper layers of waste;
   •   if gas removal is by a gas transmission layer beneath the GM and vents, the gas
       transmission layer should be designed with adequate long-term transmissivity;
       and
                                   6-23

-------
   •   if gas is removed by horizontal collection trenches, some of the trenches should
       be placed in close proximity to the bottom of the final cover system to prevent
       gas accumulation and uplift pressure on the cover system GM.

6.5.3   Landfill Settlement
The design of final cover system drainage systems and gas collection systems
(including the gas wells, vents, and/or collection trenches, and the network of piping for
gas and condensate transmission systems) is complicated by the magnitude of waste
settlement that typically occurs at solid waste landfills. Post-closure total settlement
may equal 10 to 20% of the landfill height for MSW landfills and up to 20 to 30% of the
waste height for some abandoned dumps. The design, construction, and  maintenance
of both the final cover system and gas management system must take these
settlements into account. Figure 6-1 illustrates the magnitude of post-closure
settlements that can occur at MSW landfills.  The settlement magnitudes given in this
figure should be considered to represent the upper range of values potentially
applicable to modern landfills because the database used to develop the figures
includes data for not only MSW landfills, but also abandoned dumps. Of equal concern
(but largely unquantified) is the differential settlement that may occur in isolated areas  of
the landfill.
      CO
      0)
      .
      o

      0)
      E
      0)
      "CD
                          10
       100

Time in Days (log)
1000
10000
Figure 6-1. Total post closure settlement data for MSW landfills and abandoned
           dumps [after Edgers et al. (1990); Konig et al. (1996); Spikula (1996)].
                                  6-24

-------
The time frames over which both total and differential settlement may occur are quite
long and depend on the many factors including the liquids management strategy
practiced at the site. Table 6-1 presents a framework for evaluating likely post-closure
total and differential settlements at MSW landfills and abandoned dumps.
Table 6-1. Impact of Liquids Management Practice on Final Cover System
           Settlement at MSW Landfills and Abandoned Dumps(1-2) (Koerner and
           Daniel, 1997).
Leachate
Management Practice
Standard leachate
withdrawal
Leachate recirculation
Total Settlement
Amount
10-20%
10-20%
Time
< 30 yrs.
< 1 5 yrs.
Differential Settlement(3>4)
Amount
Little to moderate
Moderate to major
Time
< 20 yrs.
< 10 yrs.
None, e.g., at abandoned   Up to 30%   > 30 yrs.   Unknown              > 20 yrs.
  landfills or dumps
1HW landfills, ISW landfills, and MSW ash monofills usually have much less settlement than the amounts
 listed in this table.
2The estimates in this table regarding the impact of the liquids management practice on settlement of
 landfill final cover systems are based on sparse data. They are meant to be a guide only, and site-
 specific estimates are required to develop more appropriate figures for any particular final cover system
 project.
3The estimates in this table regarding differential settlement amount and time are also based on
 very sparse data.  Clearly, field monitored data is needed in this regard.
4These qualitative assessment terms are also affected by the density of the waste; well-compacted
 waste produces less differential settlement than poorly-compacted waste.
6.5.4  Landfill Fires
While the incidence of landfill fires in MSW landfills has greatly diminished since the
days of the "open dump", they still sometimes occur. Air-to-methane mixture ratios of
20 to 50% have given rise to at least one fire, which damaged a geosynthetic final cover
system.  The vulnerable time frame of a facility with respect to landfill fires appears to
be after the GM is seamed and before cover soil is placed. Wind uplift of the GM  can
draw air  in through vents providing the oxygen necessary to create ignitable conditions.

Fires at depth within a waste mass may occasionally occur. The origin of such fires is
apparently spontaneous combustion and an air source is required for sustenance. The
key to preventing such a fire is to block air entry.  Identifying and blocking all potential
sources of air entry can sometimes be difficult.
                                    6-25

-------
6.5.5 Fate of Gas Management Systems
For large regionalized landfills where energy is utilized there is an incentive to maintain
the gas management system in good working order. When the energy conversion
becomes inefficient, however, the wells or vents may be decoupled from their external
piping systems and be allowed to vent to the atmosphere.  It is important, at that time, to
show the amount of gas being vented is below regulatory limits and does not present a
health or an environmental hazard.  It is also important to show that gas emissions
through the final cover system in the vicinity of the decoupled well or vent are below
regulatory limits.

6.6  Long-Term Landfill Management
The performance data for operating landfills presented in this report demonstrate that
landfills can be designed, constructed, and operated/maintained to achieve very high
levels of leachate and landfill gas containment and collection. The report has also
demonstrated that design, construction, and operation/maintenance issues and
problems persist at many landfills.  In the preceding  part of this chapter,  the authors
have attempted to provide guidance to design engineers on how to avoid the most
significant issues and problems that may typically arise.  Information on the anticipated
service lives of the various engineered components  of a landfill waste containment
system was also given.

The ultimate degradation of any individual waste containment system component of a
landfill after the completion of that component's useful service life may or may not lead
to a release of leachate or gas and contamination of groundwater.  Furthermore, a
release may, or may not, result in a significant environmental impact. In evaluating the
consequences of ultimate degradation,  the design engineer must consider a wide range
of factors including: the climatological and hydrogeologic setting; the composition, age,
and level of degradation of the waste; the potential for leachate and gas generation after
the component has completed its service life; the potential to maintain, rehabilitate, or
install other systems to achieve leachate and gas containment;  and collection, cost, and
social and institutional factors. These various factors should be considered within an
overall decision-making framework that addresses long-term landfill management.
Long-term landfill management strategies are discussed in Appendix G.

6.7  References
Anderson, D.G. and Kavazanjian, E., Jr. (1995), "Performance of Landfills Under
  Seismic Loading," Proceedings of the 3rd International Conference on Recent
  Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis,
  Missouri, pp. 1557-1587.
Ajaz, A. and Parry,  R. H.  G. (1975a), "Stress-Strain  Behavior of Two Compacted Clays
  in Tension and Compression," Geotechnique, Vol. 25, No. 3, pp. 495-512.
                                   6-26

-------
Ajaz, A. and Parry, R. H. G. (1975b), "Analysis of Bending Stresses in Soil Beams,"
   Geotechnique, Vol. 25, No. 3, pp. 586-591.
Ajaz, A. and Parry, R. H. G. (1976). "Bending Test for Compacted Clays," Journal of
   the Geotechnical Engineering Division, Vol. 102, No. 9, pp. 929-943.
Bonaparte, R. and Gross, B.A. (1990), "Field Behavior of Double-Liner Systems,"
   Proceedings of the Symposium on Waste Containment Systems, ASCE Geotechnical
   Special Publication No. 26, pp. 52-83.
Bonaparte, R., Gross, B. A., Daniel, D. E., Koerner, R. M., and Dwyer, S. F. (2002),
   "Technical Guidance for RCRA/CERCLA Final Covers," U.S. EPA Office of Solid
   Waste and  Emergency Response, Washington, D.C. (in final review)
Bonaparte, R. and Othman, M. A. (1995), "Characteristics of Modern MSW Landfill
   Performance," Geotechnical News, Vol. 13, No. 1, pp. 25-30.
Breitenbach, A. J. and Swan, R. H.  (1999), "Influence of High Load Deformations on
   Geomembrane Liner Interface Strengths," Geosynthetics '99 Conference
   Proceedings, Industrial Fabrics Association International, St. Paul,  MN, Vol. 1, pp.
   517-529
Brown, K. W., Thomas,  J. C., Lytton, R. L, Jayawickarama, P. and Bahrt, S. C. (1987),
   "Quantification of Leak Rates Through Holes in Landfill Liners," report for cooperative
   agreement  CR-810940, EPA Risk Reduction Engineering Research Laboratory,
   Cincinnati, 147 p.
Byrne,  J. (1994),  "Design Issues with Strain-Softening Interfaces in Landfill  Liners,"
   Proceedings, Waste Tech '94, National Solid Waste Management Association,
   Charleston, 26 p.
Byrne,  R. J., Kendall, J. and Brown, S, (1992), "Cause and Mechanism of Failure,
   Kettleman Hills Landfill B-19, Unit 1A," Stability and Performance of Slopes and
   Embankments-ll, ASCE Geotechnical Special Publication No. 31,  R. B. Seed and R.
   W. Boulanger, eds.,  pp. 1188-1215.
Dove, J. E., Frost, D. J., Bachus, R. C. and Han, J. (1997), "The Influence of
   Geomembrane Surface Roughness on Interface Strength," Proceedings,
   Geosynthetics '97 Conference, Long  Beach, CA, Mar., Vol. 2, pp. 863-876.
Edgers, L., Noble, J. J. and Williams, E. (1990), "A Biologic Model for Long Term
   Settlement in Landfills, " Tufts University, Medford, MA.
Eid, H. T. and Stark, T. D. (1997), "Shear Behavior of an Unreinforced Geosynthetic
   Clay Liner," Geosynthetics International, Vol. 4, No. 6, pp. 645-659.
EPA (1987), Proposed Rulemaking, 40 CFR Parts 260, 264, 265, 270 and 271, "Liners
   and  Leak Detection for Hazardous Waste Land Disposal Units," Federal Register,
   Vol.  52, No. 103, pp.  20218-20311.
Gilbert, R. B.,  Scranton, H. D. and Daniel, D. E. (1997), "Shear Strength Testing for
   Geosynthetic Clay Liners," Testing and Acceptance Criteria for Geosynthetic Clay
   Liners, ASTM STP 1308, Larry W. Well, Ed., American Society for Testing and
   Materials.
                                  6-27

-------
Giroud, J. P., Bachus, R. C. and Bonaparte, R. (1995), "Influence of Water Flow on the
  Stability of Geosynthetic-Soil Layered Systems on Slopes," Geosynthetics
  International, Vol. 2, No. 6, pp. 1149-1180.
Giroud, J. P., Badu-Tweneboah, K. and Soderman, K. L. (1997), "Comparison of
  Leachate Flow Through Compacted Clay and Geosynthetic Clay Liners in Landfill
  Liner Systems," Geosynthetics International, Vol. 4, Nos. 3 and 4, September, pp.
  391-431.
Giroud, J. P. and Beech, J. F. (1989), "Stability of Soil Layers on Geosynthetic Lining
  Systems," Proceedings, Geosynthetics '89 Conference, Vol. 1, San Diego, pp. 35-46.
Giroud, J. P. and Bonaparte, R. (1989a), "Leakage Through Liners Constructed with
  Geomembranes, Part I:  Geomembrane Liners," Geotextiles and Geomembranes,
  Vol.8, No. 1, pp. 26-67.
Giroud, J. P. and Bonaparte, R. (1989b), "Leakage Through Liners Constructed with
  Geomembranes, Part II:  Composite Liners," Geotextiles and Geomembranes,  Vol.
  8, No. 2, pp. 77-111.
Hsuan, Y. G. and Koerner, R. M. (1998), "Antioxidant Depletion Lifetime in High Density
  Polyethylene Geomembranes," Journal of Geotechnical and Geoenvironmental
  Engineering, ASCE, Vol. 124, No. 6, pp. 532-541.
Kavazanjian, E., Jr. (1998), "Current Issues in Seismic Design of Geosynthetic Cover
  Systems," Proceedings of the 6th International Conference on Geosynthetics,
  Atlanta, pp. 219-226.
Kavazanjian, E., Jr., Matasovic, N., Bonaparte, R. and Schmertmann,  G. R. (1995),
  "Evaluation of MSW Properties for Seismic Analyses," Proceedings of the ASCE
  Specialty Conference Geoenvironment 2000, ASCE Geotechnical Special Publication
  No. 46, Y.B. Acarand D.E. Daniel, eds., Vol. 2, pp. 1126-1141.
Koerner, G. R., Koerner, R. M. and Martin, J. P. (1994), "Geotextile Filters Used for
  Leachate Collection Systems:  Testing, Design of Field Behavior", Journal of
  Geotechnical Engineering, ASCE, Vol. 120, No. 10, pp. 1792-1803.
Koerner, R. M. and Daniel, D. E. (1997), "Final Covers for Solid Waste Landfill and
  Abandoned Dumps," ASCE Press, 256 pgs.
Koerner, R. M. and Hwu, B. L. (1991), "Stability and Tension Considerations Regarding
  Cover Soils in Geomembrane Lined Slopes," Geotextiles and Geomembranes, Vol.
  10, No. 4, pp. 335-355.
Koerner, R. M. and Soong, T. Y. (1998), "Analysis and Design of Veneer Cover Soils,"
  Proceedings, Sixth International Conference on Geosynthetics," Industrial Fabrics
  Association International, St. Paul, MN, Vol.  1, pp. 1-26.
Konig, D., Kockel, R. and Jessberger, H. L. (1996), "Zur Beurteilung der Standsicherhert
  und zur Prognose  der Setzungen von Mischabfalldeponien," Proceedings 12th
  Nurnberg Deponieseminar, Vol. 75, Eigenverlag LGA, Nurnberg, Germany,  pp. 95-
  117.
Leonards, G. A. and Narain, J. (1963), "Flexibility of Clay and Cracking of Earth Dams,"
  Journal of the Soil Mechanics and Foundations Division, Vol. 89, No. 2, pp.  47-98.
                                  6-28

-------
Matasovic, N., Kavazanjian, E., Jr., Augello, A. J., Bray, J. D. and Seed, R. B. (1995),
  "Solid Waste Landfill Drainage Caused by 17 January 1994 Northridge Earthquake,"
  The Northridge, California Earthquake of 17 January 1994, California Department of
  Conservation, Division of Mines and Geology Special Publication No. 116, M.C.
  Woods and R.W. Sieple, eds., pp. 221-229.
Matasovic, N. and Kavazanjian, E., Jr. (1996), "Observations of the Performance of
  Solid Waste Landfills During Earthquakes," Proceedings of the 11th World
  Conference on Earthquake Engineering, Elsevier Science Ltd. Paper No. 341, 8 p.
  (on CD ROM).
Othman, M. A., Benson, C. H., Chamberlain, E. J., and Zimmie, T. F. (1994),
  "Laboratory Testing to Evaluate Changes in Hydraulic Conductivity of Compacted
  Clays Caused by Freeze-Thaw:  State-of-the-Art," Hydraulic Conductivity and Waste
  Containment Transport in Soils, STP 1142, D.E. Daniel and S.J. Trautwein (eds.),
  American Society for Testing and Materials, Philadelphia, PA, pp. 227-254.
Reynolds, R. T. (1991), "Geotechnical Field Techniques Used in Monitoring Slope
  Stability at a Landfill," Proceedings Field Measurements in Geotechnics, ed. G.
  Sorum, Rotterdam:  A. A. Balkema, pp. 883-891.
Richardson, G. N., Kavazanjian, E., Jr. and Matasovic, N. (1995), "RCRA Subtitle D
  (258) Seismic Design Guidance for Municipal Solid Waste Landfill Facilities," U.S.
  Environmental Protection Agency Report No. 600/R-95/051, 143 p.
Rowe, R. K. (1998), "Geosynthetics and the Minimization of Contaminant Migration
  through Barrier Systems Beneath Solid Waste," Proceedings, 6th International
  Conference on Geosynthetics, Atlanta, pp. 27-102.
Sabatini, P. J., Schmertmann,  G. R. and Swan,  R. H. (1998), "Issues in
  Clay/Geomembrane Interface Testing," Proceedings of the 6th International
  Conference on Geosynthetics, Atlanta, pp. 423-426.
Sabatini, P. J., Griffin, L. M., Bonaparte, R.,  Espinoza, R. D., Giroud, J. P. (2001),
  "Reliability of State-of-Practice for Selection of Shear Strength Parameters for Waste
  Containment System Stability Analysis,"  Proceedings, GRI-15 Conference on Hot
  Topics in Geosynthetics - II (Peak/Residual; RECMs; Installation Concerns),
  Geosynthetic Research Institute, pp 86-109.
Seed, R. B. and Bonaparte, R. (1992), "Seismic Analysis and Design of Lined Waste
  Fills: Current Practice," Stability and Performance of Slopes and Embankments - II,
  ASCE Geotechnical Special Publication No. 31, pp. 1521-1545.
Sharma, H. D., Hullings, D. E., and Greguras, F. D. (1997), "Interface Strength Tests
  and Application to Landfill Design, " Geosynthetics '97 Conference Proceedings,
  Industrial Fabrics Association International, St.  Paul, MN, Vol. 2, pp.  913-926.
Spikula, D. (1996), "Subsidence Performance of Landfills: A 7-Year Review,"
  Proceedings GRI-10 Conference on Field Performance of Geosynthetics and
  Geosynthetic Related Systems, Geosynthetic Research Institute, Philadelphia, pp.
  237-244.
                                   6-29

-------
Stark, T. D. and Poeppel, A. R. (1994), "Landfill Liner Interface Strengths from
  Torsional-Ring Shear Tests," Journal of Geotechnical Engineering, Vol. 120, No. 3,
  pp. 597-615.
                                   6-30

-------
               Appendix A
Behavior of Waves in High Density
   Polyethylene Geomembranes
                      by

             Robert M. Koerner, Ph.D., P.E.
                  Drexel University
               Philadelphia, PA 19104
                  performed under

           EPA Cooperative Agreement Number
                  CR-821448-01-0
                  Project Officer

                 Mr. David A. Carson
        United States Environmental Protection Agency
           Office of Research and Development
        National Risk Management Research Laboratory
                Cincinnati, OH 45268

-------
                               Appendix A
                           Behavior of Waves in
               High Density Polyethylene Geomembranes
A-1  Overview and Focus
Geomembranes (GMs) form the essential material component in many liner systems
which require a liquid or vapor barrier. Such applications are landfill liners, landfill
covers, liquid impoundment liners, and other waste pile liners. The usual assumption in
the placement of such liners is that they lay flat on the subgrade beneath them, e.g., on
the underlying compacted clay liner, geosynthetic clay liner, etc.  This is sometimes not
the case. Waves, or wrinkles, of different sizes can occur in the as-placed and seamed
GMs, see Figures A-1  and A-2.  These waves have given design engineers a certain
amount of concern as to the behavior of GMs after soil backfilling or covering. The
research study described in this appendix was developed to shed insight into the issue
of GM wrinkles.
Figure A-1. Relatively small waves, or wrinkles, in a field deployed GM.
                                    A-1

-------
Figure A-2.  Relatively large waves, or wrinkles, in a field deployed GM.
The approach to this study of the behavior of GM waves involved an extensive series of
laboratory tasks.  It is important to note that the purpose of the tests was to evaluate the
behavior of GM waves under field stresses. The tests were not designed to try to
quantify the effects of waves on hydraulic containment performance. The scope of the
laboratory testing program involved an assessment of the effects of the following four
variables on wave behavior:

      (a) normal stress;
      (b) original wave height;
      (c) thickness of GM; and
      (d) temperature.

Due to its common use in a variety of waste containment systems, high density
polyethylene (HOPE) GMs were used throughout the study.  In particular, one
manufacturer's commercially available GM was used. The only GM variable considered
was thickness. In all other cases, the thickness was maintained at 1.5 mm, which  is a
commonly used HOPE GM thickness in many applications.

GM waves, such as seen in Figures A-1  and A-2, can be classified as two different
types: thermally-induced GM waves and construction induced GM waves.
                                      A-2

-------
Thermally-induced waves in GMs are created due to the thermal expansion
characteristics of GMs after they are seamed together and before backfilling occurs.
These types of waves have been observed in GMs for many years (Schultz and Miklas,
1980). The height and/or width of the wave depends on the GM type (e.g.,  modulus,
thickness, surface texture, surface color), temperature difference after seaming and
before backfilling, and distance between points of fixity, e.g., previously backfilled
locations.

As an illustration of a thermally induced wave, a 30 m long section of 1.5 mm thick
HOPE GM (with a thermal expansion coefficient of 15 x 10~5/°C) undergoing a
temperature change from 15°C installation temperature to 50°C (sheet surface) under a
summer sun, will expand the following amount:

            AL  =AT(a)L = (50-15)(15x 10'5) (30) = 0.158 m = 158mm

Obviously, such thermally induced GM waves can be created in the field via the local
ambient conditions.

Alternatively, construction induced GM waves are sometimes created purposely.  In
North America, an adequate amount of slack is sometimes left in the GM liner to
compensate for the coldest temperatures envisioned (EPA, 1993). The philosophy is
that the majority of the slack will be removed when the GM is covered and the sheet
temperature is reduced.  Ultimately, when the envisioned coldest temperature is
reached, the rest of the built-in slack will be completely removed, therefore,  intimate
contact to the underlying soil will be achieved.

In order to estimate the wave dimensions that can be created by a given amount of
slack in GMs, Figure A-3 was developed. In the figure, the slack in the GM  which
results in the creation of waves with various height-to-width ratios is plotted  as a
function of wave height.

As seen in Figure A-3, a  slack of 158 mm, as calculated  in the earlier example, can
create the following different wave patterns:

      • a 120 mm high wave with height-to-width ratio of 1.0;
      • a 165 mm high wave with height-to-width ratio of 0.5;
      • a 215 mm high wave with height-to-width ratio of 0.33; or
      • a 265 mm high wave with height-to-width ratio of 0.2
                                      A-3

-------
            500.
            400.
   Slack in
Geomembrane
    (mm)
300.
            200.
            100.
              0
                                                   Height-to-Width
                                                        Ratio
                                                         1.0
                                                         0.5

                                                         0.33
                                                         0.2
                 0
             50      100      150     200
                     Height of Wave (mm)
                     250
          300
Figure A-3. Slack in GM resulting in the creation of waves with various height-to-
width ratios.
Alternatively, if the GM is relatively flexible or thinner than in the previous example, two
or even more smaller waves can be created within the same slack in the GM.  Table
A-1  summarizes the types of multiple waves that can be produced by a 158 mm
expansion of a GM.  In an actual facility, an expansion of this magnitude can certainly
create waves of the type seen in the photographs of Figures A-1 and A-2.  While this
example is based on the coefficient of thermal expansion/contraction of 1.5 mm thick
HOPE GMs,  it should be noted that all types of GMs currently used in the waste
containment  industry have similar values of coefficient of thermal expansion/contraction
(Koerner, 1998).

Table A-1. Types of Waves Produced by a Slack of 158 mm in a GM in a Distance
of 30 m with a Temperature Difference of 35°C
  Height-to-Width
   Ratio (H/W)
      Single Wave Height
             (mm)
Two Waves Height
      (mm)
Three Waves Height
       (mm)
1.00
0.50
0.33
0.20
120
165
215
265
60
80
105
130
40
55
70
90
                                     A-4

-------
However, such an ideal situation of a perfectly flat GM is very difficult to achieve.  The
reasons are as follows:

   • it is very difficult to quantify the actual difference between the two extreme
     temperatures, i.e., the installation temperature and the coldest temperature
     envisioned;
   • the focus of concern is the sheet temperature, not the ambient temperature;
   • the sheet temperature is a complicated function of ambient temperature,  surface
     color and texture, incidence of sun, weather, etc.;
   • accurate measurement of the coefficient of thermal expansion is difficult;
   • frictional forces mobilized between the interfaces can retard, or even constrain, the
     reduction of slack; and
   • it is extremely difficult to build into an installed and seamed GM, a prescribed
     amount of slack.

As a result, relatively large waves of the type seen in Figure A-2 are commonly seen in
the field.

It has been observed both in the field and in the laboratory that the installed wave
greatly distorts from its original shape under increasing normal stress. However, the
deformation pattern depends on the GM type  (e.g., thickness, modulus, flexural rigidity,
etc.), the original wave shape (e.g., height, height-to-width ratio, etc.), and the
surrounding environment (e.g., stress level, duration, temperature, etc.).

Figure A-4 illustrates some possible deformation scenarios.  Figure A-4a shows how the
wave distorts under relatively low normal stress. Figure A-4b shows that the profile of
the wave remains almost unchanged when  higher normal stress is applied.  This could
possibly be the case for waves in relatively thick and/or stiff GMs. When normal stress
is applied nonuniformly (e.g., with a horizontal component), the waves may roll over
towards one side as seen in Figure A-4c. When normal stress is applied to waves in
relatively thinner or more flexible GMs, they may become vertically flattened as seen in
Figure A-4d.  For extremely flexible GMs, they may even be flattened in a pancake
manner as seen in Figure A-4e.  Note that conditions as shown in Figure A-4d and e are
also possible when the service temperature is relatively high.

The concern as to the ultimate fate of GM waves should certainly receive attention as to
a rigorous understanding of the problem.  However, to date, all analyses and
investigations into GM waves have been semi-qualitative,  see Giroud and Morel (1992)
and Giroud (1995). Quantitative approaches which evaluate the  ultimate fate of GM
waves in a more rigorous manner are needed. For instance, there could be a  maximum
wave height, for a given set of conditions, where the GM wave will eventually
                                      A-5

-------
                            (a) under low normal stress
                       r\
                (b) after initial distortion            (C) one sided roll-over
                 (d) vertically flattened           (e) horizontally flattened

Figure A-4. Possible deformation scenarios in GM waves.
lay flat on the soil.  Given a maximum wave height, which could be specified in the
installation contract, the optimal fate of GM waves might be as follows:

   1.  The installer seams the GM with waves up to a maximum specified amount.
   2.  Typically a geotextile (GT) will cover the GM and the temperature of the GM will
      decrease. Thus, the wave(s) will decrease in size (Koerner and Koerner, 1995).
   3.  Upon backfilling over the GT covered  GM, the waves are fixed in position and
      contained by friction from further size  reduction stemming from future decreasing
      temperature.
   4.  Under increasing normal stress, due to soil, solid waste or liquids, the wave
      distorts from its original shape. As seen in Figure A-5, from results of this study,
      the wave becomes narrower in width at its base and only marginally shorter in its
      height. Thus, the wave's height-to-width ratio  is actually accentuated from  its
      initial condition (from approximately 0.33 to 0.44 for this particular GM).
   5.  Over time, creep and/or stress relaxation in the polymer structure occurs and the
      wave height decreases in size thus reducing the  H/W ratio.
   6.  Ultimately, it is hoped that the wave flattens to a H/W ratio of zero, so as to
      achieve contact with the underlying soil subgrade.

From the above description it is suggested that creep and stress relaxation play a key
(and essentially unknown) role in the ultimate elimination of GM waves.  Furthermore,
by knowing the characteristics of the "entombed" wave,  one can  possibly back-calculate
to the originally allowable maximum wave height.
                                       A-6

-------
Figure A-5.  Wave distortion under increasing normal stress from large-scale
laboratory experiments conducted in this study.
These issues then frame the essence of this study.  It is focused completely on GM
waves which exist in the GM at the point of backfilling and are caused by elevated
temperature above that which existed when the GM rolls were seamed.  In this study,
only GMs made from HOPE are evaluated. This is felt to be justified since HOPE
represents approximately 70% of the landfill liner market in North America. In other
countries, e.g., Germany, it is the only type of GM that is allowed.

A-2 Experimental Setup and Monitoring
A large-size experimental test box was constructed in the laboratory for the evaluation
of the behavior of HOPE GM waves.  Initially, the test box was utilized to conduct
preliminary tests to gain a better understanding of the problem to be investigated.  It
was then used for the justification of performing smaller scale experiments.  Finally, it
was designated for conducting a 10,000-hour control test. Details regarding each of  the
these items will  be presented after a description of the test box.

A photograph and schematic illustration of the test box is shown in Figure A-6.  The
basic components of the setup include a rigid box and a data acquisition system.  The
box has dimensions  of 1.8 m long by 1.0 m wide by 1.0 m high.  On the front panel,
there  is a 0.5 m-wide plexiglass window for the  purpose of visual observations. An air
bag which provides a uniform normal pressure up to 70 kPa is placed on top of the soil
and the reaction is transmitted through a 25 mm thick wooden board to five steel
reaction cross beams connected at the top of the box.  Also,  a  number of electrical
resistance strain gages are bonded on the test specimen at various locations with wires
extended out of the box and connected to a data acquisition system.
                                      A-7

-------
                                          1.8m
   1.0m
                                                                     0.6 m
        Void,



Sand above--
                                                                     0.4m
                                                                 — *"
                                                           To arata acqusition
Figure A-6. Photograph and schematic illustration of the large-scale experimental

test box used in this study.
                                     A-8

-------
The experimental monitoring of the behavior of HOPE GM waves includes two parts:
profile-tracing of the actual wave and strain gage monitoring. The profile-tracing
provides the opportunity of visual observation and recording the distortion of GM waves
under various experimental conditions. Important information such as the final
configuration, the final height-to-width ratio, and the locations of stress concentrations
can be obtained using this type of monitoring. Tracing the profile of GM waves is done
via the window on the front panel of the test box. An example of profile-tracing was
shown in Figure A-5. This type of monitoring was also performed routinely on trial runs
before the actual experiments began to determine the layout pattern of the other type of
experimental monitoring, i.e., the strain gage monitoring.

Strain gage monitoring quantifies the actual strain induced at different locations of the
GM wave under various experimental conditions.  When used in  conjunction with a data
acquisition system, this type of monitoring provides reliable information on the
experiment over the duration of the test. The strain gages used in this study are
electrical resistance (foil-type) strain gages having resistance of 120-ohms and gage
length of 12.7 mm. With proper configuration, this particular type of gage measures
strain within the range of ±5%.  The installation procedure recommended by the gage
manufacturer was precisely followed.  The surface cleaning and preparation was
considered most critical in this regard.  The photograph of an installed strain gage is
shown in Figure A-7. Note that a bondable terminal along with two curved "jumper
wires" are also used in the gage installation to prevent the gage from being  subjected to
any unexpected stresses.
                                     C onnection
                                      Terminal
Figure A-7. Strain gage with soldered connection installed on GM specimen.
                                       A-9

-------
Sets of preliminary tests were conducted using the large-scale experimental setup.
These tests were performed at an early stage of the task and were designed to gain
further understanding of the GM wave, as well as to evaluate the possibility of
transferring the large-scale tests to a small-scale experimental setup.  The material
used in these tests was a 1.5-mm thick smooth  HOPE GM.  Details of the preliminary
tests are presented as follows.

The first series of preliminary tests consisted of three separate experiments.  Namely, a
1.5-mm thick GM with relatively large, moderate, and relatively small waves.  The
waves were created by using specimens longer than the inner length of the test box.
After each specimen was placed  in the box, sand backfilling was started from the end to
the center of the  box in a symmetrical manner.  Consequently, the "slack" of specimen
was "pushed" toward the center and, as a result, a wave was formed.  The "original"
configuration of waves was defined as the wave profile under approximately 100 mm of
sand backfill. Using profile-tracing as previously described,  the shape of the original
wave was recorded. The same type of monitoring was repeated at various stages of
the backfilling and progressed until the maximum normal pressure provided by the
experimental setup, i.e., 70 kPa, was reached.

The results of the profile-tracing monitoring of these tests are shown in Figure A-8.  In
the figure, the outermost curves of all tests represent the original wave configuration
and the innermost curves correspond to the final wave profiles under 70 kPa.  As seen
in the figure, under increasing normal stress, the waves greatly distort from their original
shapes.  The waves become narrower in width at the base but only marginally shorter in
height. A quantitative parameter was devised by calculating a ratio of the wave  height
to its base width, i.e., a H/W ratio. For purposes of gaining perspective with field
installations, a somewhat accepted rule-of-thumb in  the field deployment of GMs is that
the height-to-width  (H/W) ratio should not be greater than 0.5. This being the case, the
"relatively large" and "moderate" waves in this study were already marginal from the
outset. The accentuated H/W ratios upon backfilling, 2.0, 0.9 and 0.4  as seen in Figure
A-8, were already considered as a valuable finding in the course of this study.

Even further, with respect to the empirical field guide, the drastic increase in the H/W
ratio for the "relatively large" and  "moderate" waves  indicated locations of high curvature
and therefore the possibility of high stress concentrations under even higher normal
stresses. Such waves should clearly be removed before the placing of backfill.  As a
result, the rest of this task focused on waves with an original height smaller than the
height of the relatively small wave shown in Figure A-8.

Also seen in Figure A-8 are the reference marks located at various portions of the
waves. These marks are very helpful in tracking the critical  locations of a wave  under
normal pressure with respect to the undeformed test specimen. Therefore, the
information will be used to establish the layout pattern of strain gage installation.
                                      A-10

-------
 (a) Relatively Large Wave
^\
Original
Final
Ht.
240 mm
125 mm
H/W
0.5
2.0
250 mm
                                                                       200 mm
                                                                       150 mm
                                                                       100 mm
                                                                       50 mm
                                                                       0 mm
(b) Moderate Wave
^\
Original
Final
Ht.
130 mm
70 mm
H/W
0.4
0.9
250 mm
                                                                       200 mm
                                                                       150 mm
                                                                       100 mm
                                                                       50 mm
                                                                       '0 mm
 (c) Relatively Small Wave
^\
Original
Final
Ht.
80 mm
35 mm
H/W
0.25
0.4
250 mm
                                                                       200 mm
                                                                       150 mm
                                                                       100 mm
                                                                       50 mm
                                                                       0 mm
Figure A-8.  Results of the profile-tracing monitoring of three preliminary tests.
                                     A-11

-------
As mentioned earlier, the pressurizing mechanism (i.e., the air bag and reaction beams)
in the large-scale test box can only provide a uniform normal pressure up to 70 kPa.  If
the average unit weight of typical solid waste is assumed as 12 kN/m3, such a normal
pressure is approximately equivalent to solid waste of 6 m in height.  This is relatively
low for a typical landfill. In order to evaluate the behavior of GM waves under high
normal pressures, e.g., greater than 1,000 kPa, transferring the experiments to smaller
setups which allow the application of higher normal pressures is necessary. Moreover,
smaller setups which can be housed in a environmental room will be especially
beneficial since the effect of temperature on the behavior of GM waves can then be
investigated.  However, such smaller tests must be justified on the basis  of this larger
test setup.

A small-scale setup justification test was designed and conducted to examine the
behavior of the wave itself. A wave, identical to the relatively small wave shown in
Figure A-8, was created in the large-scale test box. However,  instead of being
supported by the side walls of the test box, both ends of the specimen were held by
metal sticks 50 mm away from the walls of the box. In addition, both ends of the test
specimen were covered by 75 mm-wide smooth HOPE GM strips acting  as protective
slip-sheets. The experimental setup is shown in Figure A-9.
1 mm HOPE
strips on bot
                 50 mm
          c
                                          Air bag
                        GM
                         sides
                                             Supporting sticks
                                             removed after box
                                             is filled with sand
                                       reference marks
            u
Figure A-9.  Justification experiment for small-scale experimental setups.
                                      A-12

-------
Before the backfilling process was started, a 300 mm by 300 mm square region was
marked on the window of the test box. It was used as a virtual image of a smaller test
box in which the GM wave could be housed.  Two reference marks, immediately
adjacent to the square region, were made on the front edge of the wave, as seen in
Figure A-9. With the supporting sticks on both ends of the test specimen, backfilling
was carefully carried out until the test box was filled.  The supporting sticks were then
removed, leaving two horizontal spaces of 50 mm each on both ends of the specimen
(protected by the slip sheets) for possible lateral movement.

The GM wave was then pressurized using the air bag against the reaction beams. It
was observed that under a normal pressure of 70 kPa, the wave distorted in a manner
exactly like the relatively small wave shown in Figure A-8.  Moreover, the two reference
marks  remained completely stationary, i.e., there was no lateral movement of the GM.
This observation suggests that the frictional forces, mobilized  between the GM
specimen and the adjacent sand fill, were sufficient to restrict the horizontal portions of
the GM from any lateral movement and decrease in wave height.

In other words, the mobilized friction forces on the horizontal extensions of the wave
offered the same reaction as would a smaller test box simulated by the 300 mm by 300
mm square region. This important finding not only provided the justification of using a
smaller scale test  box, it also justified the use of both experimental setups, large and
small scale, to simulate situations in the field where the HOPE GMs waves are normally
much further apart.

Based on the above findings, four rigid boxes having dimensions of 300 mm long by
300 mm wide by 300 mm high were built.  Along with steel reaction frames and a
hydraulic pressurizing system, these boxes allow a application of normal pressure
higher than 1,500  kPa.  This is equivalent to a solid waste landfill of approximately 125
m in height, i.e., a so-called "megafill". In addition, all four boxes can be simultaneously
housed in a environmental room where constant environmental conditions can  be
maintained within  ranges of 0 to 55°C temperature and 0 to 98%  relatively humidity.
Photographs of one of four identical small scale test boxes and the environmental room
used in this task are shown in Figure A-10. As seen in the figure, data acquisition is
also available for strain gage measuring.

One of the objectives of the experimental  part of the task is to investigate the behavior
of HOPE GM waves under various conditions. As discussed earlier, the four small-
scale test boxes in conjunction with the environmental room are ideal in this regard.
The other objective of the experimental part of this study is to  obtain actual long-term
experimental data so that the validity of using rheologic models for the purpose of long-
term prediction can be evaluated.
                                      A-13

-------
Figure A-10.  Photographs of the small scale test box and the environmental room
used in this study.
Four sets of 1,000 hour experiments, utilizing the small scale test boxes just described
within an environmental room, were designed and conducted to evaluate the effect of
four experimental parameters on the behavior of HOPE GM waves.  These parameters
were the normal stress, original height of wave, thickness of GM, and testing
temperature. Table A-2 presents the experimental design of these tests. As seen, the
effects of different variables were evaluated by varying the particular one under
investigation while holding the others constant.  In all cases, smooth HOPE GMs were
used and strain gages were attached to the wave specimens at different locations with
continuous readout over the duration of the tests.  Note that all of the waves in the
experiments listed in Table A-2 were created with an original height-to-width ratio of
approximately 0.33.  Such a ratio was found typical for most of the naturally formed
HOPE GM waves in the laboratory covered by little-to-no backfill.

The large-scale test box was reserved and used for conducting a single long-term
(10,000 hours) control experiment.  A 1.5-mm thick smooth HOPE GM wave with
original height of 60 mm and a original height-to-width ratio of 0.33 was created and it
was subjected to a constant normal stress of 70 kPa at a temperature of 23±2°C. This
test is considered to be the control test for subsequent comparison of the results of the
small-scale tests.
                                     A-14

-------
Table A-2.  Experiments Conducted Using Small Scale Test Boxes	
 Experimental  	Experimental Conditions	
   Parameter     Normal Stress   Original Height       GM         Temperature
   Evaluated        (kPa)          of Wave     Thickness  (mm)       (°C)
                                    (mm)

Normal
Stress

180
360
700
1,100

60



1.5 23


 Original Height
    of Wave
      GM
   Thickness
    Testing
 Temperature
700
700
700
14
20
40
60
80
60
14
20
40
60
1.5
1.0
1.5
2.0
2.5
1.5
23
23
23
42
55
A-3  Experimental Results -1,000 hour Tests
The results of all twenty five of the 1,000 hour tests, as listed in Table A-2, will be
presented in this section. They will be given on a variable-by-variable basis.  Both
original and final (after 1,000 hours) shapes of the GM waves along with the
corresponding heights and height-to-width ratios will be shown.  Also, if applicable, a
comparison among results generated under different test conditions will be made to
evaluate the effect of that particular experimental variable.

As listed in Table A-2, four 1.5 mm thick HOPE GM wave specimens, having original
heights of 60 mm, were subjected to four different normal stresses, namely, 180, 360,
700,  and  1,100 kPa. The temperature was maintained at 23°C for all experiments over
the entire duration of the experiments, i.e., 1,000 hours.  The original (same for all
specimens) and the final shapes of all test specimens, obtained via profile-tracing
monitoring, are shown in Figure A-11.

Six strain gages,  numbered  from G1 to G6, were originally bonded at  the locations
shown in Figure A-11  for all specimens.  Note that gages G4 to G6 (shown as darker
circles in  Figure A-11) were bonded on the lower side of the GM since the gages which
                                     A-15

-------
Figure A-11. Original and final shapes of HOPE GM waves under various normal
stresses (grid lines have dimensions of 10 mm by 10 mm).

were used respond more accurately under tension than compression. As a result of
different normal stresses, these gages measured the strains corresponding to various
locations on the GM test specimens. A typical result of the test conducted under a
normal stress of 700 kPa is shown in Figure A-12 where the measured strains are
plotted against time. By viewing Figures A-11 and A-12 simultaneously, it is seen that
the upper portion of this particular wave specimen experienced measurable strain with a
maximum tensile strain of 3.4% recorded near the crest of wave.
   Strain (%)
               0
200
800
1000
                                    400        600
                                     Time (hours)

Figure A-12. Strain measurement results of experiment conducted at 700 kPa.
                                    A-16

-------
By investigating the results generated from both parts of the experimental monitoring,
i.e., the profile-tracing illustrated in Figure A-11  and the strain gage measuring
illustrated in Figure A-12, information such as final wave height, final height-to-width
ratio, maximum strain recorded, and the locations of high stress concentrations were
obtained.  Table A-3 summarizes such information obtained from the first series of
1,000 hour experiments.

As shown in Table A-3, the final wave height decreases with increasing normal stress.
However,  the height-to-width ratio increases with increasing normal stress even more
significantly.  It was seen that the effect on the height-to-width ratio is essentially
doubled in comparison with the effect on  the final wave height.  For example, a normal
stress of 700 kPa resulted in a 37% reduction in the wave height compared to its
original configuration.  However, the same normal stress caused a 76% increase in the
height-to-width ratio. Since high height-to-width ratios generally indicate large
curvatures and locations of high stress concentration, the overall effect of high normal
stress is obviously unfavorable.

Table A-3. Summarized Results of Test Series No.1 - Effect of Normal Stress
Normal
Stress
(kPa)
0
(original)
180
360
700
1,100
Final Wave
Ht.
(mm)
60
(original)
47
42
38
34
Final
H/W
Ratio
0.33
(original)
0.47
0.51
0.58
0.62
Max.
Strain
(%)
+ 1.7
(original)
+ 1.8
+ 2.0
+ 3.0
+ 3.2
Actual Location(s) of Highest
Stress Concentration
(Strain Gage Location)
Crest of wave (G1)
Crest of wave (G1)
Crest of wave (G1)
Crest of wave (G1)
Upper portion of wave
(G1, G2andG3)
Upper portion and base of wave
(G2 and G5)
The strain recorded in each experiment shows that tensile strain increases as normal
stress increases.  This is expected since the H/W values increase significantly with
greater curvature.  Nevertheless, the GM is tensioned significantly less than its yield
point. (Note that the tensile yield strain for this GM is in the range of 15 to 25%
depending on the temperature.) Therefore, tensile yield is not expected. However, the
general design objective is to place the GM with as little stress as possible.  This
concern will be re-examined later where the actual stresses induced will be quantified
using various rheologic models.

The second series of 1,000 hour experiments was designed to evaluate the  effect of the
original wave height on the behavior of HOPE GM waves. Five tests using 1.5 mm-thick
                                      A-17

-------
HOPE GM wave specimens were conducted. The original heights of the waves were
14, 20, 40, 60, and 80 mm, respectively. All specimens were subjected to a constant
normal stress of 700 kPa and maintained at a constant temperature of 23°C over the
entire duration of the experiment. The original and final  (after 1,000 hours) shapes of
the test specimens are shown in Figure A-13. Again, reference marks which identify the
locations and movement of the bonded strain gages are also shown in Figure A-13.

By summarizing the results generated from both parts of the monitoring, Table A-4 was
established.
                                     -G1--G2
                    (a) GM wave with original height of 14 mm
                    (b) GM wave with original height of 20 mm

Figure A-13.  Original and final shapes of HOPE GM waves with various original
wave heights (grid lines have dimensions of 10 mm by 10 mm).
                                     A-18

-------
                    (c) GM wave with original height of 40 mm
                                         G2
                                                 G5
                                                   G6
                                                  \
                    (d) GM wave with original height of 60 mm
                    (e) GM wave with original height of 80 mm

Figure A-13 (cont.).  Original and final shapes of HOPE GM waves with various
original wave heights (grid lines have dimensions of 10 mm by 10 mm).
                                    A-19

-------
Table A-4.  Summarized Results of Test Series No.2 - Effect of Original Wave Height
 Original   Original    Final     Final    Max.      Actual Location(s) of Highest
WaveHt.    H/W    Wave Ht.   H/W    Strain         Stress Concentration
  (mm)     Ratio      (mm)     Ratio     (%)	(Strain Gage Location)
14
20
40

60
0.
0.
0.

0.
.17
.15
.27

.33
8
12
25

38
0.
0.
0.

0.
14
.18
.38

.58
+ 0.
+ 1.
+ 2.

+ 3.
2
2
.4

.0
Negligible
Base of wave
Upper portion
(G2 and G4)
Upper portion

(G3)
and base

of wave


of wave


(G1, G2andG3)
80

0.

.33

47

0.

.65

+ 3.

.4

Upper portion
(G2 and G4)
and base

of wave

As seen in Table A-4, there was an approximate 40% reduction in height after 1,000
hours for all waves.  As to the final H/W ratio, it increases with increasing original wave
height. Note that for waves originally higher than 60 mm, the final H/W ratios exceeded
a value of 0.5.  With regard to the maximum strain recorded, an increasing trend is also
seen with increasing original height.  Moreover, there was no sign of achieving intimate
contact between the specimen and the underlying subgrade after 1,000 hours,  even for
the wave with the smallest original height, i.e., the 14 mm wave.

The third series of 1,000 hour experiments was designed to evaluate the effect of GM
thickness on the behavior of HOPE GM waves. Four tests using HOPE GM wave
specimens, with thicknesses of 1.0, 1.5, 2.0, and 2.5 mm, were conducted.  The original
heights of all wave specimens were approximately 60 mm. Owing to the various
stiffnesses of the GMs having different thicknesses, a constant value of original H/W
ratio could not be maintained, see Table A-5. All specimens were subjected to a
constant  normal stress of 700 kPa and maintained at a constant temperature of 23°C
over the entire duration of the experiments.  The original and final (after 1,000 hours)
shapes of the test specimens, along with reference marks which indicate the location
and movement of the strain gages, are shown in Figure A-14.

As shown in Table A-5, with the only exception being the 1.0-mm-thick GM wave, the
following observations are made.  First, the thickness of GM has very little effect on the
final height of GM waves.  There was an approximate 40% reduction in height after
1,000 hours for all waves. In other words, the original height essentially determined the
final height of GM waves.  Second, the GM thickness did show a significant effect on
the final H/W ratio of the waves.  That is to say, the final H/W ratio decreases with
increasing GM thickness.  The latter observation can be interpreted in an alternative
manner.  That is, for waves with the same original height, thicker GMs resulted in wider
voids  beneath the wave.  Third, the maximum strain recorded in each experiment shows
that tensile strain slightly increases as the thickness of GM increases.
                                      A-20

-------
Table A-5. Summarized Results of Test Series No.3 - Effect of GM Thickness
GM
Thickness
(mm)
1.0
1.5

2.0

2.5

Original
H/W
Ratio
0.24
0.34

0.18

0.21

Final
Wave Ht.
(mm)
27
38

33

38

Final
H/W
Ratio
0.52
0.56

0.34

0.32

Max.
Strain
(%)
+ 2.5
+ 3.0

+ 3.1

+ 3.3

Actual Location(s) of Highest
Stress Concentration
(Strain Gage Location)
Base of wave (G5)
Upper portion and base of
wave (G1, G2 and G3)
Upper portion and base of
wave(G1, G2, G4 and G5)
Upper portion and base of
wave (G2, G3, G4 and G5)
Note: Original heights of all wave specimens were approximately 60 mm
                                                 G5
                     (a) GM wave with thickness of 1.0 mm
                     (b) GM wave with thickness of 1.5 mm

Figure A-14. Original and final shapes of HOPE GM waves with various
thicknesses (grid lines have dimensions of 10 mm by 10 mm).
                                    A-21

-------
                      (c) GM wave with thickness of 2.0 mm
                      (d) GM wave with thickness of 2.5 mm

Figure A-14 (cont.). Original and final shapes of HOPE GM waves with various
thicknesses (grid lines have dimensions of 10 mm by 10 mm).
The fourth series of 1,000 hour experiments were designed to evaluate the effect of
temperature on the behavior of HOPE GM waves. Three sets of experiments, each
consisting of 1.5 mm thick HOPE GM waves with original heights of 14, 20, 40, and 60
mm, were conducted at temperatures of 23, 42 and 55°C.

The original shapes of all wave specimens were formed at 23°C with approximately 100
mm of sand backfill over them.  Temperature was then increased, as necessary, to the
desired value. This was meant to replicate field situations where the exposed GMs
experience an increase in temperature after placement and seaming.  The test boxes
were then filled with sand, followed by a decrease in temperature back to 23°C, to
simulate the decreasing in the sheet temperature of the field deployed  GMs after the
protection and drainage layers are placed.  After approximately 24 hours, a constant
normal stress of 700 kPa was applied. After another hour, temperature was  increased
                                     A-22

-------
from 23°C to the desired value and maintained for the remainder of the experiment.
The last step was intended to simulate a possible increase in the sheet temperature
over the entire lifetime of landfills.

The original and final shapes of the test specimens at the three temperatures, along
with reference marks which indicate the location and movement of the strain gages, are
shown in Figure A-15.

A typical strain measurement result of this series of experiments is shown in Figure
A-16. This particular test was conducted at a temperature of 42°C using a wave
specimen with an original height of 20 mm. As seen in the figure, temperature was
increased from 23 to 42°C one hour after the normal stress was applied. For this
particular experiment, a trend of increasing strain with increasing temperature was
observed in all measurements.  This is due to a combined effect of both thermal
expansion and material softening with increasing temperature.  Although such a trend is
seen in most of the other measurements, a decreasing trend was also observed in
some cases.  This suggests that the change of shapes due to the material softening
with increasing temperature can sometimes cause portions of the GM waves to  undergo
compressive stresses. When such an effect is more significant than the effect of
thermal expansion, a decreasing strain with increasing temperature is seen.

The summarized results generated from this test series of the monitoring is presented in
Table A-6. Note that the values of maximum strain listed in the table are corresponding
to the maximum final (after 1,000 hours) strain.

A-4 Experimental Results -10,000 hour Tests
The strain gage measurement results of the 10,000 hour test are presented graphically
in Figure A-17.  Experimental data up to 1,000 hours was used to establish the first set
of Kelvin-Chain models for predictions out to 10,000 hours.  The calculated curves
using these models are shown in dashed lines. As seen in Figure A-17, they agree with
the actual data measured between 1,000 and 10,000 hours very well.  This encouraging
finding is felt to justify the use of the Kelvin-Chain model for the  purpose of  long-term
prediction, Soong (1996). With  this in mind, the second set of Kelvin-Chain models was
developed using the entire  experimental strain gage measurements and another order
of extrapolation, i.e., predictions out to 100,000 hours, was made.  The resulting curves
are also shown in Figure A-17 as solid lines.

A-5 Analysis of Test Results
The experimental results, including the profile-tracing of the actual waves and the strain
gage monitoring, of the 1,000 hour tests were summarized and briefly discussed in the
previous section. Complete results of all of the strain gage monitoring,  along with the
                                      A-23

-------
                    (a) GM wave with original height of 14 mm
                    (b) GM wave with original height of 20 mm
                    (c) GM wave with original height of 40 mm
                                     G1
                                         G2
                                              G4
                                                 G5
                         23°
                         42'
3
                         >5°C
                    (d) GM wave with original height of 60 mm

Figure A-15. Original and final shapes of HOPE GM waves at various
temperatures (grid lines have dimensions of 10 mm by 10 mm).
                                    A-24

-------
       Strain (%)
               -1


•

+
0
t


•
• •

+ +
o o
23°C

•
» • ii ••
« i - » **
T
o
1 + + 1 *
, o i o°
23 to 42°
	 1 	
1
*f* • •<
t. « »<
*
I
K "
iii 1 1
| + +H

C 4

•%~.
«»«4*«
/rvpococr

-H-k-j... .

2°C

1

* G1
, + G2
o G3
• G4




                  -3
                10
-2
-1
                            10
                            10
10
                        10     10      10      10
                 I700 kPa stress applied     Tjme (hours)

Figure A-16.  Strain measurement results of test conducted on wave specimen
with an original height of 20 mm and at a temperature of 42°C.
Table A-6.  Summarized Results of Test Series No. 4 - Effect of Temperature
Original
Height (mm)/
H/W Ratio

14/0.17


20/0.15



40/0.27



60/0.33



Temp.
(°C)
23
42
55
23
42
55
23

42
55
23

42

55

Final Wave
Height
(mm)
8
10
5
12
14
12
25

25
25
38

30

28

Final
H/W
Ratio
0.14
0.19
0.20
0.18
0.21
0.30
0.38

0.42
0.40
0.58

0.52

0.55

Max.
Strain
(%)
+ 0.2*
+ 0.6
+ 1.3
+ 1.2
+ 1.6
+2.1
+ 2.4

+ 3.2
+ 2.1
+ 3.0

+ 4.9

+ 4.9

Actual Location(s) of Highest
Stress Concentration
(Strain Gage Locations)
Negligible
Negligible
Base of wave (G2)
Base of wave (G3)
Base of wave (G4)
Base of wave (G4)
Upper portion and base of
wave (G2 and G3)
Base of wave (G3)
Crest of wave (G1)
Upper portion and base of
wave (G1, G2, G3 and G5)
Upper portion and base of
wave (G1, G2, G3 and G5)
Upper portion and base of
wave(G1, G2 and G5)
Note:  "+" strain=tension
      "-" strain=compression
                                     A-25

-------
     Strain (%)
             1
                       10
10
10
10
10
10
10
                                     Time (hours)
Figure A-17. Experimental and modeled results of the 10,000 hour control test.

predicted behavior up to 10,000 hours, can be found in Soong (1996).  In this section,
the previous test results will be analyzed further.  Various aspects of the test results,
including final wave height, final height-to-width ratio, and the maximum strain at the
end of 1,000 hour experiments, will be utilized to quantify the effect of different
experimental variables on the behavior of HOPE GM waves.

In this section, the height of HOPE GM wave specimens at the end of the 1,000 hour
experiments as previously described are plotted against the relevant experimental
variables.  These variables include normal stress, original height of wave, thickness of
GM and testing temperature.  The results are shown in Figures A-18 through A-21.
Some observations are made and summarized in Table A-7.

Additionally, the height-to-width (H/W) ratio of the HOPE GM wave specimens at the
end of the experiments are plotted against various experimental variables, as shown in
Figures A-22 through A-25. Some observations are made and summarized in Table
A-8.

Lastly, the maximum tensile strain of the HOPE GM wave specimens at the end of the
experiments (irrespective of their locations) are plotted against various experimental
variables, as shown in  Figures A-26 through A-29. Some observations are made and
summarized in Table A-9.
                                     A-26

-------
             70
             60

       Final
       Wave
       Height 50
       (mm)
             40
             30
                0      200      400      600     800     1000     1200

                                 Normal Stress (kPa)

Figure A-18. Effect of normal stress on the final height of HOPE GM waves.
             50
             40

       Final
      Wave  30
      Height
       (mm)
             20
             10
                0
20        40        60        80

     Original Height of Wave (mm)
100
Figure A-19. Effect of original height of wave on the final height of HOPE GM
waves.
                                    A-27

-------
       Final
      Wave
      Height
       (mm)
             60
             50
40
             30
             20
                                                 o
                0.5
             1.0        1.5        2.0        2.5

               Thickness of Geomembrane (mm)
             3.0
Figure A-20.  Effect of GM thickness on the final height of HOPE GM waves.
              50
              40
       Final
      Wave
      Height
       (mm)
30.
              20
                                                •  Original
                                                O  Original
                                                   original
                                                   Original
                                            leight = 60 mm
                                            leight = 40 mm
                                            leignt = '2(1 mm
                                            leight = 14 mm
                 20
                30            40

                   Testing Temperature (°C)
50
60
Figure A-21.  Effect of testing temperature on the final height of HOPE GM waves
having various original heights.
                                     A-28

-------
Table A-7.  Effects of Different Experimental Variables on the Final Height of
HOPE GM Waves
    Experimental Variable
               Observations
Normal Stress
Original Height of Wave
Thickness of GM
Testing Temperature
Final wave height decreases with increasing
normal stress
% reduction in height = 27 log 
-------
07
06 -
0 5 -
Final
H/W
Poti-i 04-
Katio w-^
03 -
02 -
0 1 -
0.0 -





/<>
/



(
/
>


(
X
/
-S



./ c
) .X





)






               0
              20         40          60
                  Original Height of Wave (mm)
80
100
Figure A-23.  Effect of original height of wave on the final height-to-width ratio of
HOPE GM waves.

           0.7
      Final
      H/W
      Ratio
            0.6
0.5
            0.4
            0.3
            0.2
               0.5
              1.0        1.5        2.0        2.5

                Thickness of Geomembrane (mm)
          3.0
Figure A-24.  Effect of GM thickness on the final height-to-width ratio of HOPE GM
waves.
                                     A-30

-------
      Final
      H/W  o.5
      Ratio
            0.4
                                                •  Original Height = 60 mm
                                                O  Original Height = 40 mm
                                                °  Original Height = 20 mm
                                                   Original Height = 14 mm
            0.0
                20
30            40            50

   Testing Temperature (°C)
60
Figure A-25. Effect of testing temperature on the final height-to-width ratio of
HOPE GM waves having various original heights.
Table A-8. Effects of Different Experimental Variables on the Final Height-to-Width
Ratio of HOPE GM Waves
    Experimental Variable
                  Observations
Normal Stress
Original Height of Wave
Thickness of GM
Testing Temperature
  Final H/W ratio increases with increasing normal
  stress
  % reduction in height = 59 log 
-------
               4.0


               3.5



       Maximum
        Strain
         (%)   2.5


               2.0


               1.5


               1.0
                          200     400     600     800

                                   Normal Stress (kPa)
                                             1000
1200
Figure A-26.  Effect of normal stress on the maximum strain measured at the end
of experiments of HOPE GM waves.
      Maximum
       Strain
        to,
4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0
                      A
                           20        40         60

                               Original Height of Wave (mm)
                                             80
  100
Figure A-27.  Effect of original height of wave on the maximum strain measured at
the end of experiments of HOPE GM waves.
                                    A-32

-------
             4.0
             3.5
      Maximum
       Strain
             3.0
             2.5
             2.0
                0.5
1.0         1.5        2.0        2.5
   Thickness of Geomembrane (mm)
 3.0
Figure A-28.  Effect of GM thickness on the maximum strain measured at the end
of experiments of HOPE GM waves.
                 7

                 6

                 5

       Maximum
        Tensile   4
         Strain
          (%)     3
                      • Original Height = 60 mm
                      O Original Height = 40 mm
                      D Original Height = 20 mm
                      • Original Height = 14 mm
                  20
    30           40          50
       Testing Temperature (°C)
60
Figure A-29.  Effect of testing temperature on the maximum strain measured at
the end of experiments of HOPE GM waves having various original heights.
                                     A-33

-------
Table A-9.  Effects of Different Experimental Variables on the Maximum Strain
Measured at the end of Experiments of HDPE GM Waves	
   Experimental Variable
                   Observations
Normal Stress
Original Height of Wave
Thickness of GM
Testing Temperature
• Maximum strain increases approximately
 linearly with increasing normal stress
• Max. % Strain = 0.0015 an + 1.6
 Where an = normal stress in kPa

• Maximum strain increases logarithmically with
 increasing original wave height
• Max. % Strain =4.1  log (OH) - 4.34
 Where OH = original height in mm

• Maximum strain increases approximately
 linearly with increasing GM thickness
• Max. % Strain =0.5t +2.1
 Where t = thickness of GM in mm

• Maximum strain increases with increasing
 temperature for waves originally shorter than 40 mm
• Maximum strain showed no clear trend with
 increasing temperature for waves originally higher
 than 40 mm
The Maxwell-Weichert model was seen to successfully predict the stress relaxation
behavior of HDPE GMs over the temperature range of -10 to 70°C (Soong et al., 1994;
Soong, 1995, 1996). This covers the range of interest in this study.  Moreover, the
effects of strain rate on the stress/strain relationships and the initial modulus of HDPE
GMs were also successfully described by the same model. As a result, the initial
modulus values of HDPE GMs at various temperatures, which are suitable for the use of
design and stress analysis, were quantified. Values of the initial modulus of HDPE
GMs, which will be used in the stress analysis to follow, have been assembled and
summarized in Table A-10.

Table A-10.  Modulus of HDPE GMs at Various Temperatures to be Used in the
Stress Analysis  to Follow
Temperature
(°C)
23
42
55
Initial Modulus
(MPa)
230
140
90
Using the modulus values presented in Table A-10, the stress induced in the GM can be
calculated with any known strain at a given temperature. However, such stresses will
relax over time.  The stress relaxation  behavior of the tested HDPE GM is dependent
                                     A-34

-------
upon temperature only.  In addition, master curves generated via time-temperature
superposition can be used for the prediction of long-term stress relaxation behavior.

As shown in Figure A-30, a normalized master curve for a 1.5 mm thick HOPE GM is
plotted against time at various temperatures. Via proper curve fitting, these normalized
master curves can be described using numerical expressions.  The resulting
expressions from the above procedure are given in Equations A-1, A-2 and A-3 for
temperatures of 10, 30 and 50°C, respectively.  Note that the "time" terms in these
equations are in the units of hours.

             120
             100

   Normalized
    Relaxation 80
     Modulus
              60
              40
                            V
                   -4    -3    -2   " -1  '0     1
                 10   10   10   10   10   10
                                                            50<
             10
  2   '3   ~ 4   '5     6    7
10   10   10   10   10   10
                                       Time (hours)

Figure A-30.  Normalized master curves of the long-term stress relaxation
behavior of HOPE GM at various temperatures.
Normalized stress relaxation behavior of HOPE GM at 10°C:

      (% Relaxation) = 51.4 + 8.9 log (time) -1.0 (log (time))2 + 0.05 (log (time))3  (A-1)
Normalized stress relaxation behavior of HOPE GM at 30°C:

      (% Relaxation) = 53.0 + 8.4 log (time) -1.2 (log (time))2 + 0.07 (log (time))3  (A-2)
Normalized stress relaxation behavior of HOPE GM at 50°C:

      (% Relaxation) = 48.0 + 5.3 log (time) -1.2 (log (time))2 + 0.19 (log (time))3  (A-3)
                                     A-35

-------
A procedure for analyzing the stress induced in the GM wave is proposed as follows.
Note that a worksheet, as shown in Table A-11, will be utilized to illustrate the
procedure conceptually. Also note that the numerical expression for the stress
relaxation behavior at 30°C, i.e., Equation A-2, will be used to analyze the results of
experiments conducted at 23°C.  As to the experiments conducted at 42 and 55°C, they
will be analyzed using the expression for the behavior at 50°C, i.e., Equation A-3.

Table A-11.  Elements of the Worksheet for the Stress Analysis of the
Experimental Results	
Time  Strain   Stress Induced     Relaxation       Relaxation
         s\         During       Behavior of ai0   Behavior of ai1
                                                                         Residual
                                                                        Stress, ar
  t
         so
                    = E"l"
                        x SQ
                                      Summation of
                                         stress
                                       (horizontally)
  t1      s1      ai2=Ex(e1-e0)    (1-Eqn*(ti-to)§)xai0

  t2      s2      ai2 = Ex(s2-s-|)     (1-Eqn(t2-to))xCTj0   (1-Eqn(t2-ti))xaj-|
 tn-1    sn-1
         sn     aj2 = Ex(en-en-i)    (1-Eqn(tn-to))xajQ   (1-Eqn(tn-ti))xan
 tf-1
 t.
 final
        sf-1
         sf
aif=
(1-Eqn(tfto))xai0    (1-Eqn(tf-ti))xaji
Notes:  t   Appropriate initial modulus value listed in Table A-10
          Equation A-3 for experiments conducted at 23°C
          Equation A-4 for experiments conducted at 42 and 55°C
       §   Replace the "time" terms in equations by the difference between the considered time and the
          corresponding stress induction time.

As seen  in Table A-11, the stress  induced between  any two adjacent instants of time is
determined via multiplying the differences in their corresponding strains by an
appropriate initial modulus value,  i.e., the values listed in Table A-10.  Immediately after
a stress  is induced, the GM will start to relax according to the appropriate modeled
behavior as expressed in Equations A-1, A-2 and A-3, depending on the temperature.
This concept is illustrated in the fourth and subsequent columns of Table A-11.   Finally,
as seen  in the last column of Table A-11, the instantaneous residual stress in the GM is
                                        A-36

-------
calculated by summing the remainder of all the discretized stresses corresponding to
that particular time instant.

Three example calculations which illustrate the above stress analysis procedure are as
follows. They are corresponding to the most critical strain measurements of three
different 1,000 hour experiments and their extrapolations.  Detailed information
regarding these three experiments is summarized in Table A-12.

Table A-12. Example Used to Illustrate the  Use of the Stress Analysis Procedure
Example   Thickness of    Original   Normal   Temperature    Location where
                GM          Wave     Stress                    Strain is the
                            Height                               Maximum
               (mm)	(mm)     (kPa)	(°C)
1 1.5
2 1.0
3 1.5


20
60
60


700
700
700


23
23
55


Near the base
of the wave where
the wave curvature
changes to
accommodate the
horizontal subgrade
Example A-1
As shown in Table A-12, this particular experiment was conducted at 23°C. Hence, an
initial modulus of 230 MPa and a relaxation behavior as expressed in Equation A-2 is
used in this particular stress analysis.  By inserting the strain data, along with the
appropriate constant and expression, into a preestablished spreadsheet, the strain data
is converted to stresses. The results are shown in Figure A-31, where strain and stress
are plotted against time. Note that the incorrect "modulus times strain" curve is also
shown in the figure to demonstrate the amount of stress relaxed over the entire duration
of time.

As seen in Figure A-31, a stress of 3,700 kPa was induced immediately after the full
load was applied to the wave specimen.  Subsequently, through the phenomenon of
stress relaxation along with the decreasing actual strain, the residual stress decreased
approximately 2,000 kPa to 750 kPa after 10,000 hours.

Example A-2
As shown in Table A-12, this particular experiment was also conducted at 23°C. Hence,
an initial modulus of 230 MPa and a relaxation behavior as expressed in Equation A-2
was used in this particular stress analysis.  A similar procedure to that used in Example
A-1 was carried out and the  results are shown in Figure A-32.
                                      A-37

-------
        Strain
                                                             5000
                                                             4000

                                                                 Stress
                                                                 (kPa)
                                                             3000
             0.5
             0.0
                  -3-2-101234
                 10    10   10    10    10   10   10    10
 5V
10
                                 Time (hours)


Figure A-31.  Results of the stress analysis of example 1.
           10    10     10    10    10    10    10
                                Time (hours)


Figure A-32.  Results of the stress analysis of example 2.
                                     A-38

-------
As seen in Figure A-32, a stress as high as 5,500 kPa was induced immediately after
the full load was applied to the wave specimen. Although there was only a slight
decrease in strain over the entire duration of time, a significant amount of stress was
still relaxed via the general stress relaxation phenomenon. As shown in the figure, the
residual stress decreased approximately 4,000 kPa to 1,500 kPa after 10,000 hours.

Example A-3
This experiment was started at 23°C and maintained at that temperature for one hour.
The temperature was then increased from 23°C to 55°C.  It took approximately nine
hours for the entire experimental setup to reach equilibrium at 55°C. Hence, for
analyzing strain data recorded during the initial one hour,  an initial  modulus of 230 MPa
and a relaxation behavior as expressed in Equation A-2 was used.  As for analyzing the
strain data recorded at twelve hours and beyond, an  initial modulus of 90 MPa and a
relaxation behavior as expressed in Equation A-3 was used.  Again, a similar procedure
as that used in the previous two examples was carried out and the results are shown in
Figure A-33.
      Strain
           3.0
           2.0
           1.0
                                                               10000
                                                               8000
                                                               6000
                                                                   Stress
                                                                    (kPa)
                                                              4000
                                                             Hint
                                                             ress
                                                          related
                                                              2000
                -3     -2     -1    0
               10    10    10    10   1!
                                      10    102    10   10    10
                                Time (hours)

Figure A-33.  Results of the stress analysis of example 3.
                                                               0
As seen in Figure A-33, a stress more than 9,000 kPa was induced immediately after
the full load was applied to the wave specimen. During the initial one hour of the test,
the stress relaxed to a residual value of approximately 4,500 kPa (i.e., 50% relaxation in
                                      A-39

-------
one hour). The effect of the subsequent increasing in temperature is clearly shown in
both curves between one and twelve hours. Finally, at a relatively high temperature of
55°C, the residual stress decreased approximately 2,900 kPa to 1,600 kPa after 10,000
hours.

The same procedure as illustrated in these three examples was carried out for all
twenty-five of the 1,000 hour experiments conducted in this study.  Again, only the most
critical strain measurement for each experiment was analyzed.  The complete results all
of analyses can be found in Soong (1996).

The residual stresses after 10,000 hours were also compared to the yield stress at the
particular temperature of the respective test.  The values of yield stress were obtained
via tensile tests conducted at the appropriate corresponding temperatures. The test
specimens were 1.5 mm thick HOPE GMs with a height of 50 mm and a width of 100
mm. The rate of extension used to conduct these tests was 12.7 mm/min (25%/min).
The short-term, but temperature corrected, yield stresses of HOPE GMs were evaluated
and are listed in Table A-13.

Table A-13.  Yield stresses of HOPE GMs at various temperatures to be used in
calculating the percent residual stresses to follow.
Temperature
(°C)
23
42
55
Yield Stress
(kPa)
15000
12000
9400
The entire procedure for obtaining the residual stress as a percentage of the yield stress
is summarized in a flow chart format as shown in Figure A-34.

The results of the stress analysis, in terms of the residual stress after 10,000 hours, are
summarized in Table A-14. Both the actual residual stress values and the percent of
the yield stress are presented. Some observations are made and summarized in Table
A-15.

A-6  Summary and  Conclusions
In this appendix,  the characteristics, fate, and behavior of waves of the type that are
seen in field deployed GMs were evaluated. The entire task was laboratory oriented.
However, full size waves were created,  thus it is believed that scale effects did not
significantly influence the test results. Due to their widespread use, the study focused
on HOPE GMs. The effects of four important experimental variables on the different
aspects of the  behavior of the waves were evaluated.  The variables are normal stress,
original wave height, GM thickness, and temperature.
                                     A-40

-------
                     Stress Analysis Procedure to Quantify
                        Residual Stresses in GM Waves
                           as % of the Yield Stress
   Experiment*
    Tensile Strains
    1,000 hou
 lly Measure
     out to
 • Duration
  Parameters [Evaluated:
  • normal str ;
  • original wa
  • thickness
  • temperaturs
        Extra
  1,000 hr. date
  using Kelvin
 !SS
   height
dfGM
        Perform
Stress Relaxalion
 Function of Time
 )olate
  to 10,000 hr.
 Chain Model
                                                            Tests as a
                                                            & Temp.
                   Use Maxwell-\fVeichert Model
                         To Del ermine
                    Instantaneois Modulus at
                       Creep Strain Rates
                                                                 Use Tim
                                                            Superposition
                                                             Stress Relax*
                           3-Temp.
                           for Generalized
                           ition Modulus
                         Convert all Me*
                                to Stre
                          (Instantaneou:
                        Normalized Residual
                             as a perc
                             oftheYiek
                          sured Strains
                          sses
                           & Relaxed)
                             Stresses
                          ;ntage
                           Stress
                                                           Experimentally Measure
                                                               Yield Stress at
                                                            Various Temperatures
Figure A-34.  Flow chart for the procedure of obtaining residual stresses in terms
of percent yield stress.
The experimental design for this task represented 25 separate tests each conducted for
1,000 hours.  In addition, a single control test was maintained for 10,000 hours (1.1
years).  Each of the tests utilized HOPE GMs with strain gages attached at a number of
critical locations.  This enabled extensional strain to be monitored for the duration of the
experiments.  The results of the strain gage measurements on the 1,000 hour tests
were then modeled and extrapolated one order of magnitude to 10,000  hours using the
Kelvin-chain model. The applicability of using the Kelvin-chain model was established
                                         A-41

-------
on the basis of the experimental results of the 10,000 hour control test. The other
important rheologic model presented in this study is the Maxwell-Weichert model.  It is
an analytic model that was calibrated using the results of large-scale stress relaxation
experiments. The Maxwell-Weichert model was  used to predict the stress relaxation
behavior of the modeled material at a range of temperatures.  In addition, the initial
portion of the stress/strain relationships of the modeled material at slow strain rates was
also predicted. As a result of combining both predictions, the design modulus of HOPE
GMs at various temperatures was determined. By incorporating the generalized stress
relaxation  behavior with such design modulus values, the measured strains were
converted  into tensile stresses.  These stresses were then expresses as a percent of
the tensile yield stress of the GM.

Table A-14.  Residual stress (after 10,000 hours) in the HOPE GM specimens of
experiments conducted in this study.
Experimental Parameter and Variables Residual Stress
(kPa)
Normal Stress 180kPa
360 kPa
700 kPa
1100kPa
Original Height of Wave 14mm
20 mm
40 mm
60 mm
80 mm
Thickness of GM 1.0mm
1.5 mm
2.0 mm
2.5 mm
Testing Temperature 23°C
14mm-42°C
55°C
23°C
20 mm - 42°C
55°C
23°C
40 mm - 42°C
55°C
23°C
60 mm - 42°C
55°C
1200
1300
2000
2100
130
740
1500
2000
2300
1600
2000
1600
1800
130
250
440
740
850
750
1500
1600
690
2000
2600
1600
Residual Stress
(% of Yield)
7.9
8.8
13.2
13.8
0.8
4.9
9.5
13.2
14.9
10.3
13.2
10.6
11.5
0.8
2.1
4.5
4.9
7.3
8.0
9.5
13.7
7.4
13.2
22.0
17.5
                                      A-42

-------
 Table A-15.  Effects of the Variables Evaluated in this Study on Residual Stress
 After 10,000 hours of HDPE GM Wave Experiments	
     Experimental Variable	Observations	
 Normal Stress                  •  Residual stress after 10,000 hours increases
                                 approximately linearly with increasing normal
                                 stress
                               •  Residual stress (% of yield) = 6.8 + 7 an
                                 where an = normal stress in MPa

 Original Height of Wave         •  Residual stress after 10,000 hours increases
                                 logarithmically with increasing original wave
                                 height
                               •  Residual stress (% of yield) = 18 log (OH) - 30
                                 where OH = original wave height in mm

 Thickness of GM               •  Thickness of GMs has no effect on
                                 variation of the residual stress

 Testing Temperature            •  Residual stress increases approximately linearly
                                 with increasing temperature - for waves originally
                                 shorter than 40  mm
                               •  Residual stress shows no clear trend with
                                 increasing temperature - for waves originally
	higher than 40 mm	
 The completed laboratory tests and the associated extrapolated results for 10,000 hours
 were evaluated and a number of observations were developed. These observations are
 subdivided according to the physical manifestation of the wave and its long-term stress
 condition.

 Regarding the original wave heights (which varied from 14 to 80 mm):
       • wave height decreased with increasing normal stress;
       • an average reduction in wave heights of 40% was observed after 1,000 hours;
       • GM thickness had a negligible effect on the decrease in wave height with
        normal stress;
       • there was a  slight decrease in wave height with increasing temperature;
       • final wave heights varied from 5 to 47 mm after 1,000 hours; and
       • intimate contact with the soil subgrade was not achieved after 1,000 hours,
        even for the smallest wave (14 mm) at the highest testing temperature.

 Regarding the original H/W values for the waves (which varied from 0.17 to 0.33):
       • H/W increased with increasing normal stress;
       • H/W increased approximately linearly with increasing original wave height;
       • H/W decreased approximately linearly with increasing GM thickness;
                                      A-43

-------
      • H/W decreased slightly with increasing temperature; and
      • final H/W values recorded from all experiments varied from 0.14 to 0.65 after
        1,000 hours.

Regarding the tensile strains measured at the end of the 1,000 hour experiments along
the top of the GM near the crest of the wave and the bottom of the GM near the
inflection points of the wave at its sides:
      • strains at the maximum point of curvature of the waves increased
        approximately linearly with increasing normal stress;
      • strains at the maximum point of curvature of the waves increased
        logarithmically with  increasing original wave height of the waves;
      • strains at the maximum point of curvature of the waves increased linearly with
        increasing GM thickness;
      • strains at the maximum point of curvature of the waves increased with
        increasing testing temperatures for waves originally shorter than 40 mm;
      • strains at the maximum point of curvature of the waves showed no clear trend
        with increasing testing temperatures for waves originally higher than 40  mm;
      • maximum recorded from all experiments varied from 3.2% to approximately
        4.9% after 1,000 hours.

Regarding the residual tensile stresses after the 1,000 hour experiments which were
then extrapolated to 10,000  hours:
      • residual tensile stress at the points of maximum curvature increased with
        increasing normal stress;
      • residual tensile stress at the points of maximum curvature increased with
        increasing original wave height;
      • thickness of the GM had essentially no effect on the residual tensile stresses;
      • residual tensile stresses increased with increasing testing temperature for
        waves originally shorter than 40 mm;
      • residual tensile stresses showed no clear trend with increasing testing
        temperature for waves originally higher than 40 mm; and
      • residual tensile stresses recorded from all experiments varied from 130  kPa
        (approximately 1 % of the yield stress) to 2,600 kPa (approximately 22% of the
        yield stress).

Based on the test results and the observations given above the following conclusions
are provided:

     • GM waves, which are induced in the field during placement and seaming  of GMs,
      distort upon the application of even a small normal stress. The distortion
      typically increases the height-to-width ratio of the wave.
     • The maximum tensile strain measured in this series of twenty-five 1,000-hour
      tests was approximately 5%.  Note that yield of HOPE GMs is in the range of 15
                                      A-44

-------
      to 25% strain (depending on the temperature), thus yielding of the GM was not
      observed in the tests.
     • The maximum  tensile stresses occur at locations of maximum tensile strain.
      These locations are on the side of the GM that undergoes extension, i.e., along
      the upper surface of the wave near its crest and along the lower surface where
      the wave curvature changes to accommodate the horizontal subgrade beneath
      the wave.
     • Based on an extrapolation to 10,000 hours to account for polymer stress
      relaxation, residual tensile stresses in the GM waves varied from 1% to 22% of
      the GM short-term tensile yield stress.
     • Over the 1,000-hour experimental time of stress application for the main series of
      tests,  the waves did not appear to significantly decrease, much less disappear.
     • It is important to note  that this study did not address the potential effects of the
      waves on liquid flow in lateral drainage layers above the GM, on liquid migration
      through the GM, or on the estimated GM service life.

A-7  Recommendations for the Field Placement of GMs
As illustrated in the Section A-1, the current practice of field placement of GMs in North
America is to install the GM with a certain amount of slack.  The concept is that the
majority of the slack will be removed when the GM is covered and the temperature of
the GM is reduced from its exposed temperature during installation and seaming.  The
goal  is that when the long-term steady-state temperature is reached during the GM's in-
situ service life, the slack will be completely removed as  a result of thermal contraction
and,  therefore, intimate contact by the GM with the subgrade will be achieved.  Many
construction  quality assurance (CQA) documents  in current practice include statements
referring to slack  in the GM.  For example, in EPA (1993), it states "The GM shall have
adequate slack such that it does not lift up off the  subgrade or substrate material at any
location within the facility, i.e., no "trampolining" of the GM shall be allowed to occur at
any time."

As a result of such statements, informal rules have been developed by some for the
deployment of HOPE  GMs.  One such informal rule is that the height of GM wave must
be such that it does not fold  over on itself during backfilling; another informal rule is that
the height-to-width ratio of the installed GM wave  should not be greater than 0.5. The
implicit assumption in allowing such waves is that the subsequent decrease in
temperature, along with the creep and stress relaxation inherent in the GM, will
eventually remove the waves and reduce residual stresses to negligible levels.

However, the experimental and analytic work presented  in this study brings into
question the acceptability of  these informal rules.  It was shown in this study that the
dissipation of waves that typically occur in GM liners under current installation
procedures is only nominal and much of the original wave remains over time.  The
implication is that contact with the subgrade material should not be expected to be
                                     A-45

-------
achieved, even with relatively small waves having an original height of 14 mm. This
was the smallest wave evaluated in this study.

The results of this study show that if waves are to be avoided, the GM must be
essentially flat on the underlying subgrade before backfilling. Waves having small
heights, e.g., less than 14 mm, might be acceptable for wet clay subgrades, providing
the underlying clay is soft enough so the normal stress can "deform" the adjacent wet
clay into the void that is created beneath the wave.  Further study in this regard is
needed. Based on this task, however, the size of such waves is likely to be very small,
e.g., 5 mm or less.

Even after accounting for the stress  relaxation that occurs over 10,000 hours, a
significant amount of tensile stress still remains in GM waves. Such tensile stress could
shorten the service life of a GM in comparison to GMs that are installed flat on the
subgrade.  As already noted, this issue was not evaluated as part of the current study.

One possible GM installation option to mitigate the potential negative consequences of
GM waves is to deploy and seam the GM without slack.  This installation procedure has
found increasing application in Germany. With this procedure, as the liner cools during
the night, it develops tensile stress due to restrained thermal contraction. The following
day, the temperature again rises and the GM is covered with soil at approximately the
same temperature that it was seamed.  In this way, contact with the subgrade is
achieved with only nominal tensile stress in the GM.  Unfortunately, subsequently
induced thermal stresses, if any, will not be dissipated through the phenomenon of
stress relaxation. This was shown by Lord et al. (1995).  Moreover, experiments
showed that going from high installation temperature, e.g., 40°C, to low final service
temperature, e.g., 25°C, can induce tensile stresses as high as 1,000 kPa, see Soong
(1996) for details.

It is suggested that a balance must be achieved so as to achieve contact with the
subgrade while only inducing a nominal amount of tensile stress in the GM. This
nominal amount of tensile stress is subjective at this time, Hsuan et al (1993).  Studies
are ongoing in this regard.  This balance may require some, or all, of the following
changes in the current practice of field deployment and seaming of GMs used in landfill
liner applications.

     1. GMs having light colored (e.g., white) surfaces can be used to advantage in
        decreasing the surface temperature of the GMs while exposed, hence the
        height of the waves will be smaller (Koerner and Koerner, 1995).
     2. GMs should be deployed and seamed without intentional slack.  However,
        installation should be carried out at a temperature as close to the coolest part
        of the day as possible. After the covering GT is placed, if one is required, the
        periphery of the seamed area can be ballasted with cover soil.
                                      A-46

-------
     3.  If a GT covering is not required, placement of an overlying light colored
        temporary GT may be necessary. This can prevent the GM from being
        exposed to direct sunlight before backfilling occurs.
     4.  Backfilling should be performed only in the coolest part of the day.  Quite
        possibly, it might have to be placed  at night.

The above procedures will help considerably  in gaining contact between the GM and the
underlying subgrade.  Since the GMs should  only experience small decreases in
temperature between installation, backfilling,  and in-situ service conditions, the induced
tensile stresses should be able to be accommodated with a properly selected stress-
crack resistant GM.

A-8  References

EPA (1993), Technical Guidance Document,  "Quality Assurance and  Quality Control for
   Waste Containment Facilities", EPA/600/R-93/182, September.
Giroud, J.P. and Morel, N. (1992), "Analysis of Geomembrane Wrinkles", Journal of
   Geotextiles and Geomembranes, Vol. 11,  No. 3, pp. 255-276 (Erratum: 1993, Vol.
   12, No. 4, p378).
Giroud, J.P. (1995), "Wrinkle Management for Polyethylene Geomembranes Requires
   Active Approach", Geotechnical Fabrics Report, Vol. 13, No. 3, pp. 14-17.
Hsuan, Y.G., Koerner, R.M. and Lord, A.E. Jr. (1993), "Notched Constant Tensile Load
   Test (NCTL) for High Density Polyethylene Geomembranes", Geotechnical Testing
   Journal, GTJODJ, Vol. 16, No. 4, December, pp. 450-457.
Koerner, R.M. (1998), Designing with Geosynthetics, 4th ed. New Jersey: Prentice Hall
   Inc.
Koerner G.R. and Koerner R.M. (1995), "Temperature Behavior of Field Deployed
   HOPE Geomembranes" Proceedings Conference on Geosynthetics, Nashville, TN,
   IFAI, pp. 921-937.
Lord, A.E., Jr., Soong T.-Y.  and Koerner, R.M. (1995), "Relaxation Behavior of
   Thermally-Induced Stress in HOPE Geomembranes", Geosynthetics International,
   Vol. 2, No. 3, pp. 626-634.
Schultz, D.W. and Miklas, M.P. Jr., (1980), Proceedings Disposal of Hazardous Waste,
   EPA-600/9-80-010, March, pp. 135-159.
Soong, T.-Y., Lord, A.E., Jr. and Koerner, R.M. (1994), " Stress Relaxation Behavior of
   HOPE Geomembranes", Proceedings 5th International Conference on Geotextiles,
   Geomembranes and Related Products, Singapore, pp. 1121-1124.
Soong T.-Y. (1995), "Effects of Four Experimental Variables on the Stress Relaxation
   Behavior of HOPE Geomembranes" Proceedings Conference on Geosynthetics,
   Nashville, TN, IFAI, pp. 1139-1147.
Soong, T.-Y. (1996), "Behavior of Waves in HOPE Geomembranes,"  Ph.D. Thesis,
   Drexel University, Philadelphia, PA.
                                     A-47

-------
                 Appendix B
  Antioxidant Depletion Time in High
Density Polyethylene Geomembranes
                         by

    Robert M. Koerner, Ph.D., P.E.       Grace Hsuan, Ph.D.
    Drexel University                Geosynthetic Research Institute
    Philadelphia, PA 19104           Philadelphia, PA 19104
                    performed under

              EPA Cooperative Agreement Number
                    CR-821448-01-0
                     Project Officer

                   Mr. David A. Carson
          United States Environmental Protection Agency
              Office of Research and Development
          National Risk Management Research Laboratory
                  Cincinnati, OH 45268

-------
                                 Appendix B
                       Antioxidant Depletion Time in
               High Density Polyethylene Geomembranes
B-1 Introduction
High density polyethylene (HOPE) geomembranes (GMs) have been used extensively
as barrier materials in waste containment applications, e.g., landfills, surface
impoundments, and waste piles.  The required service lifetime of such GMs varies
according to the type of waste, the sensitivity of the local environment, the stipulated
regulations (if any), and other factors. Service timeframes that have been considered
for landfills have typically fallen into the following  ranges:

•  regulatory minimum (post closure)         =  30 years
•  typical nonhazardous waste              =  50 - 200 years
•  hazardous/low level radioactive waste      =  200 -1000 years

Ideally, the service life of a GM should be at least equal to the service life of the landfill
structure. Thus, it is important to be able to quantify the anticipated service lifetime of
GMs used in waste containment applications.

The most direct way to assess service lifetime is to use information obtained from GMs
that have been installed at actual landfills.  However, the first generation of HOPE GM
lined waste facilities is only about 20 to 25 years old.  The available information
suggests that 20-year old HOPE GMs continue to perform in a manner consistent with
their as-installed properties.  An alternative approach  is needed to estimate GM service
life beyond the 20 to 25 year timeframe. In this appendix, the results of a set of
laboratory tests are presented and described.  The results are used to develop
estimates of the service lifetime of HOPE GMs.

The laboratory testing described herein involves aging the GM samples under an
environment that is designed to simulate actual field conditions. The reaction rate that
causes the degradation of the samples under such test conditions is accelerated by
incubating the samples at elevated test temperatures.  This results in an aging of the
samples in a relatively short  period of time, i.e., a few years under accelerated
conditions in comparison to perhaps hundreds of years under actual site conditions.
The degradation data from such elevated temperature testing can then be extrapolated
to predict the lifetime at a site specific ambient temperature by using the Arrhenius
method.
                                      B-1

-------
It should be emphasized that this appendix focuses on GMs that are covered or
backfilled in a "timely manner". Covering with another geosynthetic material or
backfilling with soil is necessary to protect the GMs from ultraviolet (UV) degradation
which is not considered in this task. Furthermore, the surface temperatures of GMs that
are exposed to sunlight are invariably much higher than the applications to which this
study on covered GMs is directed.

Note that this report focuses on lifetime of the antioxidants which are part of a HOPE
GM formulation. Subsequent stages of the total lifetime of the GM are induction time
and the onset of physical/mechanical property degradation. Due to the long term nature
of the incubation processes (up to 10 years), only the first stage of antioxidant depletion
time is reported in this appendix.  An example of the entire sequence of the three stages
of the long-term aging process of GMs was given in Section 2.5 of the main report.

B-2  Formulation, Compounding and Fabrication of HOPE GMs
Before going into a discussion on the long term aging mechanisms of HOPE GMs, the
various steps of producing HOPE GMs will be explained.  The components to  be
formulated, their compounding, and finally the manufacturing process are described in
this section.

The components of an HOPE GM consist of 96 to 97.5%  polyethylene (PE) resin, 2 to 3
% carbon black, and 0.5 to 1.0% antioxidants.  It should be recognized that HOPE GMs
are actually manufactured using PE resin with a density between 0.932 and 0.940 glee.
This resin density is classified as medium density PE according to ASTM D 883.  The
addition of carbon black and antioxidants, however, increases the formulated density of
the product to a range between 0.941  and 0.950 g/cc which is defined as HOPE in
ASTM  D 883. Therefore, the conventional term used in the industry of "HOPE" will be
used.

•   PE  - The resin used for HOPE GMs is a linear copolymer which  is produced by using
   ethylene and a-olefin as comonomer under low pressure and the appropriate type of
   catalyst.  The amount of a-olefin has a direct effect on the density of the resin; a
   greater amount of a-olefin added in the polymerization yields a  lower density PE
   polymer.

•   Carbon black - Carbon black is added into a HOPE GM formulation mainly for UV
   light stabilization. The loading range of carbon black in GMs is typically  2 to 3% by
   weight per ASTM D 1603.  Up to the level of opacity, the higher the loading of
   carbon black, the greater is the degree of UV light stability.  However, the addition of
   carbon black above the opacity level (which is around 3%) will not further improve
   UV resistance (Accorsi and Romero, 1995).
•   Antioxidants - Antioxidants are introduced to an  HOPE GM formulation for the
                                      B-2

-------
   purposes of oxidation prevention during high temperature extrusion and to improve
   the product long-term service life. There are a number of types of antioxidants used
   in GM manufacture and each of them has unique functional characteristics. Usually,
   synergistic mixtures of antioxidants of more than one type are used. Although the
   total amount of antioxidants in the GM is relatively small, less than 1%, their
   presence is vital to achieving the desired product service life. Note that this aspect
   of antioxidant depletion, and the corresponding time to depletion,  is the subject of
   this appendix.

The compounding methods that are used to mix the three components (PE resin,
carbon black and antioxidants) vary from manufacturer to manufacturer.  Three different
methods can be utilized. They are as follows:

•  GM Manufacturers Perform Their Own Mixing:
   GM manufacturers can purchase pure PE resin that contains no carbon black nor
   antioxidants from resin producers.  They then purchase carbon black powder and
   antioxidants from their respective suppliers. The appropriate amounts of these three
   ingredients are mixed in an extruder, forming pellets that consist of the proper
   proportion of each component.  These stabilized pellets are then transferred to
   another extruder for GM production.

•  Let-down From Concentrated Carbon Black Pellets:
   GM manufacturers can purchase PE resin that contains antioxidants only.
   Separately, they then purchase concentrated carbon black pellets consisting of
   approximately 25% carbon black in a PE resin carrier which is the same generic type
   as the parent PE resin.  During the production of the GMs, the exact proportion of
   PE  resin/antioxidant pellets and concentrated carbon black pellets are added to the
   extruder, resulting a product with the proper proportion of each component.

•  Completely Formulated Pellets:
   GM manufacturers can purchase pellets that consist of the proper proportion of PE
   resin, carbon black and antioxidants.  The completely formulated pellets go directly
   to the extruder for GM production.

Upon using a large extruder to mix, melt and filter the resin pellets into a flowing viscous
mass, there are two major processes used for manufacturing HOPE GMs. Their
differences are at the exit section of the  extruder which is some type  of die. One
process is flat sheet extrusion wherein a flat die (or "coathanger die") is utilized.  The
other process is blown film wherein a circular die is used.  Struve (1994) explains the
details of the two processes.
•  Flat die extrusion:
   Flat dies used  in the GM extrusion process are configured with adjustable lips from
   where the polymer sheet exits.  By adjusting the die lips (manually or automatically),
                                       B-3

-------
   the thickness of the GM can be accurately controlled.  Figure B-1 (a) is a schematic
   diagram of center fed flat die. The molten polymer from the extruder enters centrally
   into the die and spreads horizontally in both directions. On exiting, the somewhat
   cooled polymer sheet is deposited onto a series of chilled rolls.  For the production
   of a very wide sheet, two side by side coathanger dies can be joined together, see
   Figure B-1 (b).  The molten polymer in each side of the die is supplied from separate
   extruders. The two melt streams commingle together within the die. Again, the
   somewhat cooled polymer sheet is deposited onto a series of chilled rolls. After
   further cooling, the GM sheet is rolled onto a core for shipment and placement.
Figure B-1 (a). Flat die extrusion process to manufacture GM (Struve, 1994).
Figure B-1(b).
1994).
Dual flat die extruders used to manufacture wide GMs (Struve,
   Blow film extrusion:
   Circular dies are also utilized in the extrusion process of manufacturing PE GMs.
   They are oriented such that the polymer exits the die vertically. The molten polymer
   supplied from the extruder enters into an annular chamber through a number of
   symmetrically radial feed ports.  As the somewhat cooled polymer exits the die, a
   large cylinder of GM is formed, as can be seen in Figure B-2.  The cylinder is closed
   at the top where it passes between a set of nip rollers which draws the GM up and
   away from the die.  The dimensional stability of the cylinder is provided by internal
   and external air pressure. After the material passes through the nip rollers, the
   collapsed cylinder is cut longitudinally, opened to form a full width of GM sheet and
   rolled onto a core for shipment and placement.
                                      B-4

-------
          Nip rollers
          Extruder
                                 Cut here and
                                 unfolded
                                    Film
                                    bubble
Circular
Die
Figure B-2. Blow film extrusion process used to manufacture PE GMs.

B-3 Stages of Degradation in HOPE GMs
After proper placement of the GM sheets and seaming into a liner system, the GM will
hopefully serve as a barrier for many years. During the service period, aging takes
place in the GM.  The aging process of HOPE GMs can be considered to be a
combination of: physical aging and  chemical aging.  Both aging mechanisms take place
simultaneously. Physical aging implies a slow processes in which the material attempts
to establish equilibrium from its as-manufactured nonequilibrium state.  For semi-
crystalline polymers like HOPE, the process involves changes in the crystallinity of the
material (Petermann et al.,  1976).  Under this definition of physical aging there are no
primary (covalent) bonds broken.

On the other hand, chemical aging  indicates some type of degradation involving the
breaking of covalent bonds, e.g., thermal-oxidation, radioactive-degradation, etc.,
(Struik, 1978).  This process eventually leads to a reduction in engineering properties.
Therefore, from an applications point of view, chemical aging is the important
degradation mechanism and should be studied in great detail.  In the following sections
the different stages of chemical aging in HOPE GMs are described.

Conceptually, the chemical aging process of a HOPE GM can be considered to consist
of three distinct stages. They can be seen in Figure B-3.  These three  stages are
designated as  (a) depletion time of antioxidants, (b) induction time to the onset of
polymer degradation and (c) degradation of the polymer to decrease some engineering
property(s) to an arbitrary level, e.g., to 50% of its original value.
                                      B-5

-------
 T3
 CD
 C
 '.2
 CD
    100
     50
 CD
 Q.
 O
A 1 B 1 C

•* ^i^ ^i^ ^
1 1 \\
1 1 \
1 1 -^
1 1
i i
1 1
1 1 1



1
\
\


A - depletion time of
antioxidants
B = induction time to onset
of polymer degradation
C = time to reach 50%
degradation of a particular
property

                     Aging Time (log scale)

Figure B-3. The three conceptual stages in chemical aging of HOPE GMs.

B-3.1 Depletion of Antioxidants
The purpose of antioxidants in a HOPE GM formulation is to prevent degradation during
processing and to prevent oxidation reactions taking place during the first stage of
service life. However, there is only a limited amount of antioxidants in the formulation.
Hence, the lifetime for this stage is also limited. Once the antioxidants are completely
depleted, oxygen will begin to attack the polymer, leading to the induction time and
subsequently the deterioration of performance properties.  The duration of this
antioxidant depletion  stage depends strongly on the type of selected antioxidants.
Since many different  antioxidants can be selected, depletion time can vary from
formulation to formulation, subsequently affecting the lifetime of the GM.  Proper
selection, however, will be seen to contribute greatly to the overall lifetime of the GM.

The depletion of antioxidants may be consequence of two processes: chemical
reactions of the antioxidants, and physical loss of the antioxidants from the polymers.  In
addition, the rate of depletion is related to the type of antioxidants, to the service
temperature, and to the nature of the site specific environment. Regarding the chemical
reactions of antioxidants,  two main functions are involved: the scavenging of free
radicals, converting them  into stable molecules, and the reaction with unstable
hydroperoxide (ROOH) forming a more stable  substance.  Regarding their physical loss,
the process involves  the distribution of antioxidants in the GM and their volatility and
extractability.  Since antioxidants are the  main subject of this appendix, a detailed
investigation of these two processes will be presented in Section B-5.

B-3.2 Induction Time
In a pure PE resin, i.e., one with no carbon black and antioxidants, oxidation occurs
extremely slow at the beginning; often immeasurably slow. However, at the end of this
period acceleration occurs more rapidly.  Eventually, the reaction decelerates and once
                                       B-6

-------
again becomes very slow. This progression is illustrated by the curve in Figure B-4(a).
The initial portion of the curve (before measurable degradation takes place) is called the
induction period (or induction time) of the polymer.

In a stabilized polymer such as one with antioxidants, the acceleration stage takes a
considerably longer time to reach. The antioxidants create an additional depletion time
stage prior to the onset of the induction time, as shown in Figure B-4(b).
                                  Induction|  Acceleration   Deceleration
                                   period       period         period
                              c
                           c o
                           CD ~
                           o> 9-
                           >* o
                              X\J
                           s^ W
                           O .a
 Antioxidant  Induction  Acceleration  I  Deceleration
depletion time1   period'     period
                  c
               c  o
               CD  ~
                  w
               O .a
                                                               period
                                                               (b)
                                       Aging Time

Figure B-4. Curves illustrating the various stage of oxidation:  (a) unstabilized PE,
(b) stabilized PE.

Regarding the chemical process, the first step of oxidation in an unstabilized PE is the
formation of free radicals.  The free radicals subsequently react with oxygen and start
chain reactions.  The reactions are described in Eqs. B-1 to B-6 (Grassie and Scott,
1985).
Initiation stage:

     RH -» R» + H» (under energy or catalyst residues)

     R« + 02 -> ROO
                                                        (B-1)

                                                        (B-2)
                                       B-7

-------
Propagation stage:

     ROO + RH -» ROOM + R-                                              (B-3)

Acceleration stage:

     ROOM -» RO' + OH' (under energy)                                      (B-4)

     RO + RH -» ROM + R'                                                  (B-5)

     OH- + RH -» H20 + R-                                                  (B-6)

(where: RH represents the PE polymer chains, the symbol "•" represents free radicals
which are highly reactive.)

In the induction period, little hydroperoxide (ROOH) is present and when formed it does
not decompose.  Thus, the acceleration stage of the oxidation cannot be achieved. As
oxidation propagates slowly, additional ROOH molecules are formed. Once the
concentration of ROOH reaches a critical level, decomposition of ROOH begins and
accelerated chain reactions begin, signifying the end of the induction period (Rapoport
and Zaikov, 1986). This indicates that the concentration of ROOH has a major effect on
the duration of the induction period.

Viebke et al. (1994) have studied the induction time of an unstabilized medium-density
PE pipe. The pipes were internally pressure tested with stagnant water and externally
by circulating air at temperatures ranging from 70 to 105 °C.  They found the activation
energy of oxidation in the induction period to be 75 KJ/mol. Using their experimental
values, an induction time of 12 years can be extrapolated at a temperature of 25°C for
the material evaluated.

B-3.3 Material Property Degradation
The end of the induction period signifies  the onset of relatively rapid oxidation. This is
because the free radicals increase significantly via the decomposition of ROOH, as
indicated in Eqs.  B-4 to B-6.  One of the  free radicals is an alkyl radical (R») which
represents polymer chains that contain a free radical.  In the early stage of acceleration,
cross-linking occurs  in these alkyl radicals due to oxygen deficiency.  The reactions
involved are expressed by Eqs. B-7 and  B-8. The physical and mechanical properties
of the material subsequently respond to such molecular changes. The most noticeable
change is in the melt index, since it relates to the molecular weight of the polymer.  In
this stage, a lower melt index value is detected.  In contrast, the mechanical properties
do not seem to be very sensitive to cross-linking.  The tensile properties generally
remain unchanged or are unable to be detected.
                                      B-8

-------
                                — CH ~ - CR /| - CH _ —
CH9-CR1-CH9—
                                            _CH9-CRrCH9—
As oxidation proceeds further, and abundance of oxygen becomes available, the
reactions of alkyl radicals change to chain scission.  This causes a reduction in
molecular weight, as shown in Eqs. B-9 and B-10.  In this stage, the physical and
mechanical properties of the material change according to the extent of the chain
scission.  The melt index value reverses from the previous low value to a value higher
than the original starting value signifying a decrease in molecular weight. As for tensile
properties, break stress and break strain decrease.  Tensile modulus and yield stress
increase and yield strain decrease, although to a lesser extent. Eventually, the GM
becomes so brittle that all tensile properties change significantly and the engineering
performance is jeopardized. This signifies the end of the so-called "service life" of the
GM.

                       02&RH
— CH2-CRrCH2 — - ^ — CH2-CR1-CH2 —

                                              OOH
                                -CH2-CRi-CH2-  +    OH

                                         °*                               (B-9)

— CH2-CR, -Chi,  —	^   — CH2-CR,-0  +  -CH2 —

         °*                                                              (B-10)

Although quite arbitrary, the limit of service life of a GM is often selected as a 50%
reduction in a specific design  property.  This is commonly referred to as the half-lifetime,
or simply the halflife. The specific property could be tensile modulus, break stress,
                                      B-9

-------
break strain, impact strength, etc.  It should be noted that even at halflife the GM still
exists and can function albeit at a decreased performance level.

Hence, the lifetime of a GM will be equal to the depletion time of antioxidants, plus
induction time of the polymer, plus the time to reach a 50% reduction in a specific
engineering property. Graphically this was shown in Figure B-3 as the sum of "A", "B"
and "C".

B-4  Major Influences on Oxidation Behavior
There are many aspects of the polymer resin, its formulation, the ambient environment
and its service conditions that can effect the oxidation behavior of HOPE GMs.  This
section describes several of them placed in two categories: internal material effects and
external environmental/service effects.

B-4.7 Internal Material Effects
The chemical and physical structure of the polymer has a strong influence on the rate of
oxidation. This structure controls the formation of free radicals and the diffusion of
oxygen into the polymer.  Three major factors will be discussed:  branch density,
crystallinity and transition metals.

The medium density PE used to manufacture HOPE GMs is a copolymer. Apart from
the dominant ethylene monomer, a comonomer is added to the polymerization. The
comonomer is some type of a-olefin such as butene, hexene, methyl pentene, or octene
(Chu and Hsieh, 1992).  The comonomer forms short chain branches along the
backbone of the PE chain. Two examples are given in Figure B-5. The concentration
of the short chains varies from 5 to 8 per 1000 carbon atoms. The particular carbon
atom where the  branch attaches is surrounded by three other carbon atoms and is
defined as the tertiary carbon. The hydrogen atom attached to the tertiary carbon
possess a lower dissociation energy than other hydrogen atoms, thus free radicals are
most likely to occur at these locations. This is illustrated by Eq. B-11.  In other words,
PE with greater branch density concentration will generate more free radicals than
those with less branches under the same conditions.
                                      B-10

-------
          — CH2-CH2-CH - CH2 - CH2—         butene as the comonomer

                        CH2

                        CH3

          — CH 2- CH2 - CH - CH2 - CH2—        hexene as the comonomer

                        CH2
                        CH2

                        CH2
                        CH3
Figure B-5.  PE with butene and hexene as the comonomer.
   ,H   ,H    H   H   H              H   H   .    H   H
 -C-C-C-C-C	^-C-C-C-C-C   +   H.           (B-11)
   HH|HH              H   H   I     H   H
            CH2                            CH2
            CH3                            CH3

It is established (Michaels and Bixler, 1961) that the crystalline regions in PE are
sufficiently dense to severely limit oxygen penetration. The result of this impermeability
is that the diffusion of oxygen in the polymer is essentially controlled by the amorphous
region. Hence, the diffusion coefficient increases as crystallinity decreases.  During the
initial stage of the oxidation, alkyl radicals are probably produced in both the crystalline
and amorphous regions.  As oxygen gradually diffuses into the amorphous region, it
converts the radicals to alkylperoxy radicals, i.e., ROO, starting the chain oxidation
reactions.  On the other hand, those alkyl radicals that are trapped in the crystalline
matrix are unable to progress further (Billingham and Calvert, 1986).  In addition,
crystallinity  relates closely to the branch density of the polymer. This is because chain
branches interrupt the folding of the polymer chains, reducing the total amount of
crystallinity  in the polymer.  Therefore, as branch density increases, the crystallinity
decreases and the rate of oxidation increases.

The oxidation reaction of PE can be  increased in the presence of transition metals, e.g.,
Co, Mn, Cu, Al and Fe (Osawa and Ishizuka, 1973). The source of these elements
usually comes from residual catalyst used to polymerize the resin.  Although the
concentration of these elements is very low, they still can be a concern regarding the
long term durability of the polymer.  The transition metals break down hydroperoxides
                                      B-11

-------
via "redox" reactions, creating an additional amount of free radicals, as demonstrated in
Eqs. B-12 and B-13.

ROOM + Mn+ -> RO- + M(n+1) + OH-                                         (B-12)

ROOH + M(n+1) -> ROO' + Mn+  + H+                                         (B-13)
B-4.2 External Environmental Effects
The oxidation reaction in PE is rather sensitive to the surrounding ambient environment.
Any conditions that provide oxygen and accelerate the formation of free radicals,
particularly the decomposition of hydroperoxide, increase the rate of oxidation. Three
considerations are described: energy level, oxygen concentration and adjacent
materials.

Sunlight, heat and radiation are three types of energy which should be considered. For
an exposed GM, sunlight is the major concern. Coupled with heat there is a great
potential for free radical formation. Covering in a timely manner, however, avoids
photodegradation and greatly diminishes the heat from direct sunlight exposure.  As
mentioned previously, this study does not address sunlight exposed  GMs. Heat,
however, can come from other sources than direct sunlight. All  other things  being
equal, a GM will degrade faster at higher temperature as opposed to lower temperature.
In predicting lifetime,  it is essential to accurately estimate the service temperature of the
buried GM. For buried wastes that are radioactive there is a potential for the GM to be
exposed to high energy levels depending on the type of waste.  It is expected that low
level radioactive (LLR) and low level radioactive mixed (LLRM) wastes are orders of
magnitude too low to  produce energy levels that could cause degradation. Conversely,
high  level radioactive (HLR) and transuranic (TRU) wastes must be assessed
accordingly.  They are not within the scope of this study

The concentration of available oxygen is an obvious essential component to  any
oxidation reaction.  For exposed GMs, the availability of oxygen  is high and the oxygen
concentration is at its maximum.  Contrary, for the liner beneath a landfill, the available
oxygen will be extremely limited.  In the case of a liner for municipal solid waste landfill,
biodegradation of the waste will probably consume most of the available oxygen.
(Poland and Harper, 1986) showed that the biodegradation of solid waste changes from
aerobic and anaerobic after approximately 3 to 5 years).  Under this  situation, if
degradation occurs in PE, it may lead to crosslinking rather than chain scission, as
shown  in Eqs.  B-7 and B-8.  In surface impoundment applications, the portion of GM
that is covered by liquid is only exposed to approximately one eighth of the oxygen in
comparison to that exposed in air. Unfortunately, in unsaturated soil, the percentage of
oxygen present is very difficult to be defined since it is affected by the type of soil and
the moisture content.  Table B-1 lists an approximate ranking of GM  exposure to oxygen
                                      B-12

-------
on the bases of various applications. This leads directly to different experimental
incubation possibilities, e.g., immersed in liquid, liquid on top/air on bottom, and
completely in air.

Table B-1. Oxygen Availability to GMs in Several Common Applications.
Application
surface
impoundment
Liners

landfill liners

final covers

Location
top of slope
top of slope
base of slope
base of slope
beneath waste

above waste

GM
Surface
top
bottom
top
bottom
top
bottom
top
bottom
Oxygen
Availability
high
moderate
low
moderate
very low
low
high to moderate
very low
The type of material (soil or liquid) that makes direct contact with the GM has an
influence on the oxidation rate. If the adjacent soil contains a large amount of transition
metals (the amount is very subjective) and there is moisture or liquid present, the
transition metals can diffuse into the GM. This action catalyzes the oxidation by
accelerating the decomposition of ROOM, as explained in Eqs. B-12 and B-13.
Furthermore, if the liquid that is contained by the GM consists of a relatively large
amount of organic solvent, the amorphous phase of the GM can swell, increasing the
oxygen diffusion coefficient and accelerating the oxidation.

B-4.3  Commentary on Various Influences
This section described some of the  major influences (internal and external) on the rate
of oxidation of HOPE GMs.  In devising this particular long term experimental program
of approximately 10 years duration,  only one type of HOPE GM was selected. Thus the
internal effects from  the material itself such as  branch density, crystallinity and transition
metals are fixed to that particular GM type.
Additionally, in developing the incubation procedures, various external effects had to be
considered. Three separate incubation environments were evaluated. They are water
immersed, landfill simulated (water above and  air below) and air immersed.  It is
suggested that these different types of incubation procedures cover the range of typical
end uses illustrated in Table B-1.

B-5 Overview of Antioxidants
Since the subject of this appendix is the depletion of antioxidants in HOPE GMs, it is
essential to explain the performance of antioxidants during their depletion period.  Three
properties will be discussed in this section; the function of antioxidant, the types of
                                      B-13

-------
antioxidants along with their individual characteristics, and the antioxidant depletion
mechanisms.
B-5.1 Function of Antioxidants
The sequence of oxidation reactions in HOPE GMs indicated by Eqs. B-1 to B-6 can
also be interconnected by cycles "A" and "B", as illustrated in Figure B-6. There are four
important links in these two cycles, designated as (a) to (d).  If any of the links are
broken, the rate of oxidation of the polymer will be retarded.  If all four links are broken,
then oxidation will be stopped. The purpose of antioxidants in the polymer is to break
such links.
                               0.

                    RH
      ROH&H20
                                                        Numbers 1 to 6
                                                        represent the
                                                        Equations 1 to 6
                                                        in the text.
            (c)  X (4)
                          ROOM
Figure B-6. Oxidation cycles in PE (Grassie and Scott, 1985).

Since the involved molecular species in each of the four cycle links are not the same,
different types of antioxidants are designed to accommodate various requirements.
Antioxidants can be divided into two categories; primary and secondary.

Primary Antioxidants - They provide stabilization by trapping or deactivating free radical
species after they are formed, i.e.,  breaking links (a), (b) and (d). The antioxidants
which intercept the links (b) and (d) function in that they donate an electron. The
electrons react with free radicals ROO, RO and »OH converting them to ROOM, ROM
                                      B-14

-------
 and H20, respectively. The types of antioxidants that break the link (a) are electron
 acceptors. They convert the alkyl free radical (R») to form a stable polymer chain.

 Secondary Antioxidants - They are designed to intercept the link (c) in the "B" cycle.
 Their function is to decompose hydroperoxides (ROOM), preventing them from
 becoming free radicals.  The chemical reactions change the ROOM to a stable alcohol
 (ROM).

 B-5.2 Types and Characteristics of Antioxidants
 Apart from the two categories just described, antioxidants can be further classified into
 four large chemical types within which many different types are included. Table B-2 lists
 the chemical type and some of the commercial available antioxidants that can be used
 in PE GMs. To ensure long term durability, a manufacturer will use two or more types
 of antioxidants; typically one from each category.

 Table B-2. Types of Antioxidants (after Fay and King, 1994)	
      Category          Chemical Type            Examples of Commercially
	Available Antioxidants	
 Primary            Hindered Phenol         Irganox® 1076, lrganox®1010,
                                            Santowhite Crystals
                    Hindered Amines         Tinuvin® 622, Chimassorb® 922
                    (HALS*)

 Secondary          Phophites                lrgafos®168
                    Sulfur Compounds        distearyl thiodipropionate (DSTDP)
                    (Thiosynergists)          dilauryl thiodipropionate (DLTDP),
                    Hindered Amines         Irganox® 1076, Irganox® 1010,
	(HALS*)	Santowhite Crystals	
 * HALS = hindered amine light stabilizers

 There is another issue that needs to  be considered during the selection of antioxidants.
 That is the effective temperature range for each of the selected antioxidants. The
 antioxidant formulation or "package"  should protect the product at both the high
 temperature of the extrusion process and the significant lower temperature during its
 lifetime.  Thus the functioning temperature range for each type of antioxidant should be
 recognized. For the four chemical types listed above, the  effective temperature ranges
 are given in Figure B-7.  The graph shows that phosphites have an effective
 temperature range above 150°C.  They are considered to be process stabilizers.  Either
 thiosynergists or hindered amines will be added to the formulation to accommodate the
 low temperature service protection. On the other hand, for a formulation consisting of
 hindered phenols, a wide range of temperatures are covered; from ambient to process
 temperatures. However, hindered phenols are only primary stabilizers.  A secondary
 antioxidant is also required which can be either thiosynergists or hindered amines.
                                      B-15

-------
           Phosphites

     Hindered Phenols

       Thiosynergists

     Hindered Amines
                      0       50     100     150     200
                                         Temperature °C
250
300
Figure B-7. Effective temperature ranges of the four antioxidant types
(Fay and King, 1994).

In the mixing of the various types of antioxidants, one must beware of the possible
antagonistic effects between them. For products that require long term thermal stability
and light stability, a combination of phenolic and thiosynergist for thermal stability, and
hindered amine for light stability could be used. Unfortunately, the oxidation product of
the sulfur compound can be acidic which reacts with hindered amine, preventing its
interacting with free radicals (Kikkawa et al., 1987).  In HOPE GMs, the carbon black
can also influence the stability of the material, in particular, the thermal stability.
Materials containing carbon black absorb more heat than those without carbon black.
While this discussion is seemingly complicated, it should be recognized that the polymer
industry is quite advanced in the selection of antioxidants. There are many custom
designed packages for each product, including GMs, in order to accommodate a wide
range of processing and service requirements.

One last item to conclude this subsection is the issue of antioxidant cost.  Antioxidants
are comparatively much more expensive than the polymer resin. Thus cost of the final
product is weighted heavily by the amount and type of antioxidants used in the
formulation. A careful balance must be drawn between the required performance and
economy of the final product.

fi-5.3 Antioxidant Depletion Mechanisms
The amount of antioxidant in a HOPE GM decreases gradually as aging progresses.
The depletion can be caused by two mechanisms; chemical reactions of the
antioxidants and physical loss of antioxidants from the polymers by leaching. These
mechanisms can occur simultaneously.
                                     B-16

-------
Chemical reactions: As discussed previously, the antioxidants are consumed by free
radicals and alkylperoxides present in the material. The rate of consumption which
progresses from the surfaces of the GM inward depends on the concentration of these
two species.  Since phenolic and phosphite types of stabilizers are utilized in the
processing stage, the antioxidants that remain in GMs for longevity protection are
probably a combination of residue phenolic types along with thiosynergists or hindered
amines. Of these three types of antioxidants, hindered amines have a unique reactive
behavior. They can be cyclical and regenerative, both leading to a long functioning
time.  Only undesirable side reactions can terminate their efficiency (Fay and King,
1994).

Physical loss:  The two major concerns with respect to the physical stability of
antioxidants in the polymer are their volatility and extractability (Luston,  1986).
Research has indicated that the distribution of antioxidants in semicrystalline polymers
is not uniform, owing to the presence of crystalline and amorphous phases. It appears
that a greater concentration of antioxidants is found in the amorphous region which is
fortunate because the amorphous region is also the most sensitive to degradation.
Hence the mobility of antioxidants in the amorphous phase controls these two physical
processes.

The volatility of antioxidants is a thermally activated process and temperature changes
effect not only the evaporation of the stabilizers from the surface of the polymer but also
their diffusion from the interior to the surface layer. For HOPE GMs, the typical
operating temperature is well below 60°C. Hence volatility is probably not a major
concern. Because of this, one must avoid inducing such a mechanism in accelerated
laboratory aging tests. Very high testing  temperatures should not be utilized.  However,
elevated temperature is necessarily to  accelerate the laboratory aging study. Therefore
a careful balance is required  in the design of the experimental incubation setup.  As
noted previously this task is proposed to  be a 10-year effort.  The reason for such long
time is that the selected test temperatures are relatively low so as to minimize volatility
of antioxidants from occurring.

Extractability of antioxidants plays a part  wherever the GMs comes into contact with
liquids such as water or leachate.  The rate of extraction is controlled by the dissolution
of antioxidants from the surface and the diffusion from the interior structure to the
surface. However, dissolution is faster  than evaporation. Smith et al. (1992) performed
an aging study on a medium  density PE pipe material that was exposed to water
internally and air externally. They monitored the antioxidant depletion across the
thickness of the pipe via oxidative induction time (OIT). They found that the
consumption of antioxidants was three times faster in water than in air at temperatures
of 105, 95and80°C.
                                      B-17

-------
Thus in the physical loss of antioxidants, extraction takes a central role in lifetime
predictions.  Clearly, this is a concern if the GM contacts liquid during its service life.
Unfortunately, there is no data available regarding the effect of humidity on antioxidant
loss.

B-6 Experimental Design
As indicated in Table B-1, HOPE GMs can experience different levels of oxygen
concentration depending on the application, its site specific location, and the materials
in contact with the upper and lower surfaces.  It is important that the laboratory aging
tests simulate the site conditions as close as possible.  In this regard, four different
laboratory incubation protocols have been developed. They are described in Table B-3.
A detailed description of each incubation method is presented in this section.

Table B-3. Incubation Method of HOPE GMs in this Task
Incubation Incubation Applied Simulated GM Application
Series Method Stress
I
water none
(both sides)
surface impoundments below
liquid level
IV
air
(both sides)

water above/air
beneath

water
(both sides)
                               none
260 kPa
(compression)

30% yield stress
(tension)	
landfill covers and waste pile
covers

landfills liners beneath waste
surface impoundments along
side slopes below liquid level
The incubation method of Series I is designed to simulate GMs which are exposed to
liquids (water or leachate) on both sides and are essentially nonstressed, e.g., shallow
surface impoundments. HOPE GM samples are fully immersed in four water baths
maintained at constant temperatures of 55, 65, 75 and 85°C.  The dimensions of the
incubated samples are 150 mm by 150 mm. Samples are retrieved at various time
intervals and evaluated by a number of tests for their physical, chemical and mechanical
properties.

The incubation method of Series II is designed to simulate GMs which are exposed to
air on both of their surfaces, e.g., landfill covers and waste pile covers. GMs in these
applications will be exposed to air from above  and perhaps from beneath as well.  The
exact oxygen concentration that the GM will be subjected to is very difficult to define.
The incubation represents the extreme condition of full oxygen exposure. Hence the
lifetime predicted from this experiment will be a conservative value. In this series, GM
samples are exposed to a continuous flow of air.  HOPE GM samples are suspended in
                                      B-18

-------
five forced air ovens maintained at temperatures of 55, 65, 75, 95°C and 115°C.  The
dimensions of the samples are 100 mm by 150 mm.  Samples are retrieved at various
time intervals and are evaluated by a number of tests for their physical, chemical and
mechanical properties.

The incubation method of Series III is intended to simulate GMs situated beneath solid
waste landfills.  Circular shaped samples of 200 mm diameter are placed  in incubation
columns similar to those suggested by  Mitchell and Spanner (1985), as shown in Figure
B-8. Twenty (20) identical units of this type are used  in this incubation series. A static
compressive stress of 260 kPa is applied to each sample. This is approximately
equivalent to a 30 m high solid waste landfill. Above each sample is a layer of 100 mm
thick sand with 300 mm of water head.   Beneath each sample is dry soil with a limited
amount of air.  Four test temperatures of 55, 65, 75, and 85°C are being utilized.
Samples are retrieved at various time intervals and evaluated by a number of tests for
their physical, chemical and mechanical properties.
                  Insulation
                   heat tape
               thermocouple
                                 MINIM
                                  Sand
                                                   10

                                              Piezometer
                  ]
                                   Sand
B-
                                                         Load
. perforated steel
 loading plate

geomembrane
sample under
compression
Figure B-8.  A schematic diagram of the compression column for incubation
series III.

The incubation method of Series IV is intended to simulate GMs located on side slopes
of surface impoundments where tensile stress may be generated.  GM samples of
dimensions 38 mm by 150 mm are subjected to a constant tensile stress equal to 30%
of their room temperature yield stress.  The tensioned samples are completely
immersed in four water baths maintained at constant temperatures of 55, 65, 75 and
                                     B-19

-------
 85°C.  Samples are retrieved at various time intervals and evaluated by a number of
 tests for their physical, chemical and mechanical properties.

 The four incubation conditions used in this study are attempts at replicating the most
 commonly encountered site situations where GMs are being used.  How the resulting
 predicted lifetimes compare to one another is obviously a major focus of this task.  From
 a comparison of the results it is hoped that one can deduce effects on oxidation on GMs
 that are not directly studied, e.g., different types of stresses, tension effects, water
 extraction effects, etc.

 B-7 Evaluation Tests on Incubated Samples
 The HOPE GM samples in the four incubation series just described are retrieved after
 predetermined  lengths of time.  The times depend on the change in property behavior
 which cannot be estimated apriori. Hence the retrieval time is adjusted for each series.
 The progression of the aging process in each series is monitored by the results of a set
 of physical, chemical and mechanical tests. Table B-4 shows the tests that are used to
 track the behavior of the incubated GM samples. Most of tests are standard test
 methods commonly performed on HOPE  GMs test specimens. The only necessary
 commentary has to do with the different types of OIT tests, since these are the test
 results presented and analyzed in this appendix.

 Table B-4. Tests Used to Evaluate Incubated Samples	
	GM Property	ASTM Test Methods	Test Description	
 crystallinity                 D 1505                    density

 antioxidant amount (total)    D 3895 and D5885         Standard OIT and
                                                     high pressure OIT

 molecular weight (indirect)   D1238                     melt flow index

 mechanical properties       D 638                     tensile properties

 stress crack resistance      05397-appendix            single point notched
	constant tensile load	

 In this study, the total amount of antioxidant remaining in the incubated  GM samples is
 evaluated using two slightly different OIT tests.  They are the standard method and the
 high pressure method. Although OIT cannot identify the individual type or exact amount
 of each antioxidant present in the formulation, it does quantify the effectiveness of the
 antioxidants.  OIT is the time required for the GM test specimen to be oxidized under a
 specific pressure and temperature.  Since antioxidants protect the GM from oxidation,
 the length of OIT (in minutes) indicates the amount of antioxidants present in the test
 specimen. Howard (1973) showed that OIT is proportional to the antioxidant
                                      B-20

-------
concentration in the same formulation package. However, for different antioxidant
packages, direct comparison between two single OIT values can be very misleading
and caution must be expressed.  Since we are only investigating a single GM type and
its antioxidant package, this caution is not a concern in this task.

B-7.1 Standard Oxidative Induction Time (Std-OIT) Test
The Std-OIT test is performed according to ASTM D3895. The test uses a differential
scanning calorimeter (DSC) with a specimen testing cell that can sustain a 35 kPa
gauge pressure. A 5 mg GM specimen is heated from room temperature to 200°C at a
heating rate of 20°C/min under a nitrogen atmosphere. The gas flow rate is maintained
at 50 cc/min.  When  200°C is reached, the cell is maintained in an isothermal condition
for 5 minutes. The gas is then changed from nitrogen to oxygen. The pressure and
flow rate of oxygen are 35 kPa gauge pressure and 50 cc/min, respectively.  The test is
terminated after an exothermal peak is detected. The exothermal peak results from the
oxidation of the  GM specimen. An example thermal curve with its identified OIT value is
shown in Figure B-9.
                                   Oxidation Exotherm
                                       OIT
        ml
      elting peak
        I
nitrogen i
                                                oxygen
                                    Time (min.)
Figure B-9. Thermal curve of a standard OIT test.

B-7.2 High Pressure Oxidative Induction Time (HP-OIT)Test
The HP-OIT test is also performed using a DSC except now with a cell that can sustain
a pressure of 5500 kPa. This type of cell is called a high pressure cell and
consequently the test is called high pressure OIT (HP-OIT).  It is performed according to
ASTM D5885.

Tikuisis et al. (1993) have performed a detailed study on the effect of pressure and
temperature on  HP-OIT values.  A series of such tests was evaluated using 8 different
isothermal temperatures ranging from 150°C to 200°C under 8 different pressures, from
                                     B-21

-------
690 kPa to 5500 kPa.  They found an Arrhenius relationship between temperature and
HP-OIT values.  Pressure had very little influence on the HP-OIT values at temperatures
above 170 °C.  At 150 °C isothermal temperature, pressure greater than 3500 kPa
resulted in little change HP-OIT values.  As a result of their study, the generally agreed
upon pressure and isothermal temperature can be selected.  In the draft ASTM
standard, these values are 3500 kPa and 150°C, respectively.

The test protocol of the HP-OIT test used in this study is that a 5 mg GM specimen is
heated from room temperature to 150°C at a heating rate of 20°C/min under a nitrogen
atmosphere. The pressure of the cell in this nitrogen stage is maintained at 35 kPa
gauge pressure. The gas flow rate is not monitored. When 150°C is reached, the cell is
maintained in an isothermal condition for 5  minutes.  The gas is then changed from
nitrogen to oxygen.  The oxygen pressure in the cell is gradually increased to 3500 kPa
within 1 minute. The test is terminated after an exothermal peak is detected.  The
exothermal peak results from the oxidation  of the GM specimen. The resulting thermal
curve is similar to that shown in Figure B-9, except that the HP-OIT value is much
longer than the Std-OIT value for the same material. This is due to the lower testing
temperature.

B-7.3 Commentary on the Different OIT Tests
The major differences between the two OIT tests are oxygen pressure and isothermal
temperature. For the standard  OIT test, a low pressure and high temperature are used.
For the HP-OIT test, a high pressure and low temperature are utilized. Their differences
create somewhat of a dilemma insofar as the selection of a preferred test method for
OIT.  Table B-5 summaries the advantages and disadvantages.

The main reason behind developing the HP-OIT test is that the 200°C testing
temperature used in Std-OIT test is unable  to bring out the stabilization effect of
thiosynergists and hindered amine types of antioxidants.  As shown in Figure B-7, the
maximum effective temperature of both of these antioxidants is below 150°C. At 200°C,
both types of antioxidants rapidly volatilize from the GM thus losing their apparent
effect.  As a result, GMs with these types of antioxidants will exhibit a shorter OIT value
than those without.  Yet the long term performance of these GMs may be very similar,
or even better than those without these types of antioxidants.  In the HP-OIT test, the
test temperature is lowered to 150°C. Note that 150°C is the minimum temperature to
ensure complete melting of the HOPE GM specimen. The low testing temperature,
however, results in an extremely long test at the standard pressure of 35 kPa, making
the test somewhat unpractical. Hence a high pressure is applied.  At a higher oxygen
pressure, the concentration gradient of oxygen atoms across the specimen's  surface
becomes greater. This increases the number of oxygen atoms diffusing into the  molten
specimen, thereby accelerating the oxidation and reducing the testing time.
                                     B-22

-------
 Table B-5.  Differences Between the Standard and High Pressure PIT Tests	
	Test	Advantages	Disadvantages	
 Std-OIT            •  existing ASTM test protocol     •  high temperature may bias
 (200°C, 35 kPa)     •  short testing time (~ 100 min)     the test results for certain
                    •  standard test apparatus          types of antioxidants

 HP-OIT             •  existing ASTM test protocol     •  long testing time (>300
 (150°C, 3500 kPa)   •  able to distinguish the            min.)
                       stabilization effect of different   •  special testing cell and set
                       types of antioxidants in the       up are required
                       GM
                    •  lower temperature relates
	closer to service conditions	

 B-8 Data Extrapolation Method
 It is well established that chemical reactions of all types proceed more rapidly at higher
 temperatures than at lower temperatures. The relationship between chemical reaction
 rate (Rr) and temperature is usually expressed by the Arrhenius equation, Eq. B-14:

 Rr =Cexp(-Q/RT)                                                       (B-14)

 where:

 Rr =  reaction rate
 C  =  constant (independent of temperature)
 Q  =  activation energy of the reaction
 R  =  gas constant
 T  =  absolute temperature

 Taking the natural logarithm of both sides of Eq.  B-14, Eq. B-15 is obtained.

 ln(Rr) = lnC-Q/RT                                                        (B-15)

 If the log reaction rate is plotted against inverse temperature as shown in Figure B-10,
 the slope of the line will be  -Q/R and the intercept on the vertical axis will be the
 constant "C". The plot in Figure B-10 is called the "Arrhenius Plot" from which reaction
 rates at other temperatures (typically lower temperatures) can be extrapolated.  In order
 to produce a reliable extrapolation, there should  be a minimum of three data points,  i.e.,
 data from three different incubation temperatures, so that the experimental portion of
 the line can be reasonably  established. In addition, the test temperatures cannot  be so
 high that changes in material occur, thereby altering the nature of the reaction.
                                       B-23

-------
             CD
             -CD
c
o
t;
CD
CD
             O)
             O
                            experimentally obtained
                                portion of curve
                                                extrapolated portion
                                                     of curve
(e
high temperature
g. laboratory tests)
low temperature" x
(e.g. site specific condition)
                            Inverse Temperature (1/T)

Figure B-10.  Generalized Arrhenius plot for low-temperature reaction rate
predictions from high-temperature laboratory experimental data.

In this aging study, the four selected testing temperatures are 55, 65, 75 and 85°C.  The
reaction rate being evaluated in this particular report will be the antioxidant depletion
rate. Data obtained from the experimental portion will be extrapolated to a lower site
specific temperature.  Hence, the potential lifetime of the antioxidants in the HOPE GM
can be assessed. In Figure B-3, this  is Stage "A" of the overall predicted lifetime of the
HOPE GM being evaluated.

B-9  Results and Data Analysis on Antioxidant Depletion
This appendix presents the results to date on antioxidant depletion rates.  Incubation
Series I and III are presented.  Series II and IV are ongoing.  As shown in Table B-4,
GM test samples in Series I are incubated in water media under nonstressed conditions
whereas GM test samples in Series III are exposed to water above/air beneath and
under compressive stress.   The antioxidant depletion rate is measured using both the
Std-OIT and HP-OIT tests as explained previously.

B-9.7 Preparation of OIT Test Specimens
The incubated samples from each test series were retrieved after varying incubation
periods. The retrieved samples were equilibrated at room conditions for 24 hours.
They were then cleaned with tap water to remove surface contaminants. The cleaned
samples were placed in a plastic bag and stored inside a cabinet until testing.

OIT test specimens weighing 5 mg each were taken from the incubated samples. They
were cut from surface to surface across the thickness of the GM  near the center portion
                                      B-24

-------
of the sample.  Therefore, the resulting OIT values represent the average amount of
antioxidants across the thickness of the test specimens.

For the Std-OIT tests, three replicates were performed on each incubated sample and
the average value was used in the analysis.  For the HP-OIT tests, a single test was
performed on most of the incubated samples.  Some samples were tested twice to
verify the reproducibility of the test.

B-9.2 Results and Data Analysis of Incubation Series I
HOPE GM samples in incubation Series I were completely immersed in water at
temperatures of 55, 65, 75 and 85°C. The average OIT value for each incubated
sample was evaluated by Std-OIT and HP-OIT tests. In this subsection, the results of
both tests are presented together with the step-by-step data analysis which leads to the
prediction of antioxidant depletion.

The Std-OIT and HP-OIT test results are shown in Table B-6. The OIT values are
presented graphically by plotting OIT value against incubation time. Figures B-11 and
B-12 show Std-OIT and HP-OIT values, respectively.  The curves in both figures
indicate an exponential decrease in the amount of antioxidant present as incubation
time increases.  The curves also exhibit that the decrease in  OIT values  is greater for
the higher incubation temperatures than for the lower temperatures.

Since OIT values decrease exponentially as incubation time increases in both tests, a
linear relationship can be obtained between  In(OIT) and incubation time. Figures B-13
and B-14 show the In(OIT) versus incubation time plots for Std-OIT and HP-OIT,
respectively. The generalized equation for each of the straight lines is expressed by Eq.
B-16.

ln(OIT)=ln(P) + (S)*(t)                                                   (B-16)
                                      B-25

-------
Table B-6. OIT Test Results of Incubation Series I
Incubation
Time
(months)
0.1
1.0
3
9
12
18
30
55°C
Std-OIT
(min)
80.5
79.5
77.0
59.0
45.3
n/a
19.1

HP-OIT
(min)
210*
201
196
173
160
n/a
111
65°C
Std-OIT
(min)
80.5
78.2
74.0
40.2
24.2
17.0
10.8

HP-OIT
(min)
210*
204
157
135
120
109
87
Incubation
Time
(months)
0.1
1.0
3
9
12
18
30
75°C
Std-OIT
(min)
80.5
75.2
69.5
15.1
9.7
10.3
2.1

HP-OIT
(min.)
210*
172
154
82
87
76
38*
85°C
Std-OIT
(min)
80.5
70.5
63.4
12.9
6.2
3.4
0.5

HP-OIT
(min)
210*
181
127
72
50*
38
28*
Notes:
All Std-OIT values are the average of three replicate tests.
All HP-OIT values are from a single test with the exception of those marked with an
asterisk which are the average of two replicate tests.

where

OIT  = OIT time
S    = slope of the lines (i.e., OIT depletion rate)
t     = incubation time
P    = constant (the original value of OIT time in either the Std-OIT or HP-OIT tests)

Table  B-7  lists the depletion rates  that are obtained from  both Std-OIT  and HP-OIT
tests.
                                       B-26

-------
                           10      15     20      25     30      35
                             Incubation Time (month)
  Figure B-11.  Standard OIT versus incubation time plot for incubation
  Series I.
         250
             0      5      10      15     20      25

                             Incubation Time (month)

Figure B-12. HP-OIT versus incubation time plot for incubation Series I.
                                      B-27

-------
g
E,
H
o
                                                      30     35
        0       5      10      15     20      25
                        Incubation Time (month)
Figure B-13. Ln(OIT) versus incubation time plot for incubation
Series I  using Standard OIT tests.
     5.5
<§
H
O
       5  -
     4.5  -
       4  -
     3.5  -
         0      5      10      15     20      25      30     35

                         Incubation Time (month)
   Figure B-14. Ln(OIT) versus incubation time plot for incubation
   series I using HP-OIT test.
                                  B-28

-------
 Table B-7. Antioxidant Depletion Rates of Incubation Series I (i.e., the slopes of
 the straight lines in Figures B-13 and B-14).	
     Test Temperature	Std-OIT	HP-PIT	
           55°C                     -0.0467                    -0.0215
           65°C                     -0.0749                    -0.0342
           75°C                     -0.1280                    -0.0615
	85°C	-0.1765	-0.0822	

 The next step in the analysis is to extrapolate the OIT depletion rate to a lower
 temperature,  such as site specific temperature.  This is performed utilizing the
 Arrhenius equation, as described in Eqs. B-17 and B-18.

 S = A*Exp(-E/RT)                                                         (B-17)

 ln(S) = ln(A) + (-E/R)*(1/T)                                                  (B-18)

 where

 S = OIT depletion rate (slope of the lines listed in Table B-7)
 E = activation energy of the antioxidant depletion mechanism (KJ/mol)
 R = gas constant (8.31 J/mol°K)
 T = test temperature in absolute value (°K)
 A = constant

 Thus  a linear relationship can be established between ln(S) and inverse temperature,
 as indicated in Figure B-15. The activation energy deduced from the slope of the lines is
 43 KJ/mol for Std-OIT and 44 KJ/mol for HP-OIT. These two values are seen to be
 extremely close to one another. The corresponding Arrhenius Equation for Std-OIT and
 HP-OIT are expressed in Eqs. B-19 and B-20.

 ln(S) = 12.839 - 5210.2/T   for Std-OIT                                      (B-19)

 ln(S) = 12.372- 5311.8/T   for HP-OIT                                      (B-20)

 Using Eqs.  B-19 and B-20, the OIT depletion rates at a site specific temperature can be
 obtained. Koerner and Koerner (1995) and Yazadini et al.  (1995) found that the
 temperatures at the base of landfills in Pennsylvania and California, USA are around
 25°C. Thus 25°C is used to demonstrate the extrapolation calculation. OIT depletion
 rates of both tests are as follows:

 S = - 0.0096    for Std-OIT
 S = - 0.0043    for HP-OIT
                                       B-29

-------
        -1
    o:
    Q.
    Q
    H
    O
        -2-
        -3-
                                         y = 12.839-5210.2x RA2 = 0.995
                                              372-5311.8X RA2 = 0.990
        0.0027
0.0028
0.0029
0.0030
0.0031
Figure B-15. Arrhenius plot for incubation Series I (water immersion-
nonstressed).

In order to predict the aging time required to deplete the antioxidants in the HOPE GM
evaluated, Eq. B-16 is utilized. The calculation procedure is as follows:

•  For Std-OIT tests:
  The Std-OIT value for a pure unstabilized (i.e., no antioxidants) HOPE fluff was found
  to be  0.5 minutes. Thus 0.5 minutes is taken to be the OIT value when essentially all
  of antioxidants in the incubated HOPE GMs are consumed.  The calculation to find
  the time for this depletion at a service temperature of 25°C is as follows:

  In (OIT) = ln(P) + (S) *(t)                                                  (B-16)
  In (0.5) = In (80.5) + (-0.0096) (t)
  -0.69  = 4.39 - 0.0096 (t)
  t  = 529 months (44 years)

•  For HP-OIT tests:
  The OIT value for a  pure unstabilized HOPE fluff was found to be 25 minutes. (This
  relatively high value is due to the low isothermal temperature and switching nitrogen
  to oxygen in the test method). Thus 25 minutes is taken to be the OIT value when
  essentially all of the antioxidants in the incubated HOPE GM consumed.
                                      B-30

-------
  In (OIT) = ln(P) + (S)*(t)                                                   (B-16)
  In (25) = In (210) + (-0.0043)*(t)
  3.22 = 5.35 - 0.0043*(t)
  t = 495 months (41 years)

Thus it is seen that the predicted antioxidant lifetime at a service temperature of 25°C is
approximately 40 years for this particular HOPE GM formulation under this set of
immersion conditions.

B-9.3 Results and Data Analysis of Incubation Series III
HOPE GM samples in incubation Series III were exposed to water on top and air
beneath and a compressive stress of 260 kPa. The incubation temperatures are 55, 65,
75 and 85°C. The average amount of antioxidants in each aged sample was evaluated
by both Std-OIT and HP-OIT tests. In this subsection, the results of both tests are
presented together with the step-by-step data analysis which leads to the prediction of
antioxidant depletion.

The Std-OIT and HP-OIT values are shown in Table B-8. Also, the OIT values are
presented graphically by plotting OIT value against incubation time.  Figures B-16 and
B-17 show the Std-OIT and HP-OIT values, respectively.  Similar to Series I, the
depletion of OIT decreases exponentially as incubation time increases. The curves also
exhibit that the decrease in OIT values is greater for the higher incubation temperatures
than for the lower temperatures.

Since OIT values in this incubation series also decrease exponentially as incubation
time increases,  the data extrapolation steps will follow those used in Series I.  Based
on Eq. B-16, a straight line can be formed by plotting In(OIT) versus incubation time, as
indicated in Figures B-18 and B-19 for Std-OIT and HP-OIT, respectively.  The slope of
the lines represent the OIT depletion  rate at each particular temperature. Table B-9
lists the depletion rates that are obtained from both Std-OIT and HP-OIT tests.

Table B-8. OIT Test Results of Incubation Series III
Incubation
Time
(months)
0.1
3
9
12
18
24
55°C
Std-OIT
(min)
80.5
74.3
55.5
54.1
57.0
52.9

HP-OIT
(min)
210*
221
181*
175*
186
167
65°C
Std-OIT
(min)
80.5
77.9
50.5
36.8
19.0
25.9

HP-OIT
(min)
210*
189
164
135
105*
125
                                      B-31

-------
Table B-8 (cont.). PIT Test Results of Incubation Series
Incubation
Time
(months)
0.1
3
9
12
18
24
75°C
Std-OIT
(min)
80.5
66.2
45.3
27.9
17.5
12.6

HP-OIT
(min)
210*
192
143
113
103
92
85°C
Std-OIT
(min)
80.5
55.0
23.5
12.6
4.3
4.0

HP-OIT
(min)
210*
181
113
94
76
38
Notes:
All Std-OIT values are the average of three replicate tests.
All HP-OIT values are from a single test with the exception of those marked with an
asterisk which are the average of two replicate tests.
Table B-9. Antioxidant Depletion Rates of Incubation Series III (i.e., the slopes of
the straight lines in Figures B-18 and B-19).
Test Temperature
55°C
65°C
75°C
85°C
Std-OIT
-0.0217
-0.0589
-0.0798
-0.1404
HP-OIT
-0.0097
-0.0284
-0.0387
-0.0661
The next step in the analysis is to extrapolate the OIT depletion rate to 25°C using Eq.
B-18. Figure B-20 shows the Arrhenius plot for both OIT tests.  The activation energy
deduced from the slope of the lines is 56 KJ/mol for Std-OIT and 58 KJ/mol for HP-OIT.
Again these two values are seen to be extremely similar to one another.  However, they
are both slightly higher than those obtained in incubation Series I. This indicates that
the reaction mechanism for antioxidant depletion in Series III requires more energy than
that in Series I.  In other words, the OIT depletion rate in  Series III is slower compared
to Series I and the correspondingly lifetime prediction will be longer. The corresponding
Arrhenius Equations for Std-OIT and HP-OIT are expressed in Eqs. B-21 and B-22.

ln(S) = 16.885-6738.9/T      forStd-OIT                                   (B-21)

ln(S) = 16.856-6991.3/T      for HP-OIT                                   (B-22)
                                      B-32

-------
           100
                                 10        15
                             Incubation Time (month)
Figure B-16. Standard OIT versus incubation time plot for incubation Series
          250
          200  -
                                 10        15       20
                             Incubation Time (month)
Figure B-17. HP-OIT versus incubation time plot for incubation Series
                                     B-33

-------
       0
                                 10        15        20
                             Incubation Time (month)
Figure B-18.  Ln (OIT) versus incubation time plot for incubation Series III using
Standard OIT tests.
           5.5
             5  -
       E
       H
       0
4.5  -
             4  -
            5.5
                        5        10        15        20

                             Incubation Time (month)
                                                   25
Figure B-19.  Ln (OIT) versus incubation time plot for incubation Series III using
HP-OIT tests.
                                     B-34

-------
       -1
       -2-
CD
-t->
CO
Q:
c
o
    Q- _
    CD
    Q
    o
    =  -4H
       -5
       0.0027
                        Standard OIT
                        HP-OIT
                                   y= 17.045 -6798.2x RA2 = 0.953
                                   y= 16.856 -6991.3x RA2 = 0.943
                 0.0028
0.0029
0.0030
0.0031
Figure B-20. Arrhenius plot for incubation Series III (water top/air beneath-
compression stress).

Using Eqs. B-21 and B-22, the OIT depletion rates can be obtained.
S = - 0.0033
S =-0.0014
            for Std-OIT
            for HP-OIT
In order to predict the aging time that is required to deplete the antioxidants in the
HOPE GMs, Eq. B-16 is utilized. The calculation procedures to obtain the depletion
times at 25°C are  as follows:

•  For Std-OIT tests:
  The OIT value for a pure unstabilized HOPE fluff was found to be 0.5 minutes. Thus
  0.5 minutes is taken to be the OIT value when the antioxidants are consumed.
  In (OIT) = ln(P) + (S) *(t)
  In (0.5) = In (80.5) + (-0.0033)*(t)
  t = 1539 months (128 years)
                                                                      (B-16)
                                      B-35

-------
•  For HP-OIT tests:
  The OIT value for a pure unstabilized HOPE fluff was found to be 25 minutes.  Thus
  25 minutes is taken to be the OIT value when the antioxidants are consumed.

  In (OIT) = ln(P) + (S)*(t)                                                  (B-16)
  ln(25) = ln(210) + (-0.0014)*(t)
  t = 1521 months (126 years)

Thus it is seen that the predicted antioxidant lifetime at a service temperature of 25°C is
approximately 120 years for this particular HOPE GM formulation under this set of
simulated conditions.

B-9.4 Status  of Incubation Series II and IV
Both incubation Series II and  IV were started in June 1998.  Thus there is only a small
amount of data available and  it  is insufficient to perform an analysis and to generate
lifetime predictions.

However, incubation Series II does require a brief discussion.  Instead of using a single
HOPE GM in the evaluation, eight different HOPE GMs are being evaluated.  The GMs
were supplied by five different manufacturers.  Thus different antioxidants were most
likely to have  be used in the GM formulation packages.  In addition, higher
temperatures  are being applied to this particular incubation series.   The test
temperatures  are 55°, 65°C, 75°C, 95°C and 115°C.  The purpose of experimental
design is to investigate the highest possible incubation temperature for HOPE GMs
without changing the antioxidant depletion mechanism. The incubation samples from
the three highest test temperatures are  being retrieved every month in order to monitor
the rate of depletion of antioxidant content.

B-10  Summary
The three distinct stages of aging of HOPE GMs are described in this report. These
stages are (A) depletion time  of antioxidants, (B) induction time to the onset of polymer
degradation and (C) the time  to reach 50% degradation of a particular property.  The
lifetime of the GM is equal to the summation of these three stages.   The focus of this
task, however, is on the depletion time of antioxidants.

Four different incubation conditions were designed to simulate various field applications
of HOPE GMs.  The incubation  environments involve a combination of air, water,
compressive stress, and tensile stress.  In addition, the aging mechanisms in each
incubation condition were accelerated by elevated test temperatures which were set at
55, 65, 75 and 85°C.  In this appendix, only data from two of the incubation series,
Series I and Series III were presented and analyzed.  Samples in Series I were
completely immersed in water without any applied stress. Series III involved samples
                                      B-36

-------
that were exposed to water on top and air beneath, and a compressive stress of 260
kPa. Samples from  both series were retrieved after specified periods of time for
property evaluation.  The antioxidant depletion of the incubated samples was monitored
using both the Std-OIT test and HP-OIT test.

Data obtained from the elevated temperatures tests were then extrapolated to a site
specific (lower) temperature using the Arrhenius model.  For a site specific temperature
of 25°C, the time to consume the antioxidants in this particular HOPE GM formulation
will take 40 years under incubation Series I conditions.  On the other hand, it will take
120 years under incubation Series III conditions.  The shorter depletion time in Series I
is probably  due to the extraction rate of antioxidant which is higher in Series I than in
Series  III. The samples in  Series I were exposed to moving water on all of their
surfaces, whereas samples in Series III were exposed to static water on only one
surface. It  is known that moving water as in the Series I tests actually causes leaching
of antioxidants. Hence, the depletion time for incubation Series I is likely conservative
in comparison to most field situations since it is not common for both sides of the GM to
be exposed to moving liquids. In this regard, the results of the Series  III tests may
better represent HOPE GM in-service  conditions.  It is noted, however, that the Series
III tests were conducted with water rather than leachate. Certain strong leachates may
increase the antioxidant depletion rate. Additional research of this effect is needed.

Regarding the effect of compressive stress on the antioxidant depletion rate in the
Series  III incubation, a definitive result has not yet been obtained. On a preliminary
basis it appears that compressive stress may reduce the depletion rate, since the
depletion time for Series III samples is three times greater than that  of the Series I
samples. One possible hypothesis  is that the compressive stress may increase the
density of the amorphous phase of the HOPE material, consequently reducing the
diffusion rates of both antioxidants out of, and oxygen into,  the GM.

Finally, it should again be emphasized that the antioxidant depletion time represents
only the initial step in a three-step GM aging process, i.e., it is Stage A in Figure B-3.  At
the end of the antioxidant depletion time,  the physical and mechanical properties of the
HOPE  GM still remain essentially unchanged.  In order to establish the service life (i.e.,
the half lifetime) of an HOPE GM, the induction time plus the time to reach 50%
reduction in the relevant mechanical property must be obtained.  This will take longer
than the current incubation time of three years, thus the time frame of this study is
estimated to be ten years.  The second and third parts of the study will be presented in
due course.

B-11 Conclusion
Since this is only the first part in a series of three stages on the topic of HOPE lifetime
prediction, this conclusion will necessarily be preliminary.  Clearly though, this study
establishes that the depletion of antioxidants in the HOPE GM  under investigation is
                                      B-37

-------
quite long. Depending on the incubation method, the time for antioxidant depletion at
25°C is between 40 to 120 years.

These values, in and of themselves, are powerful indicators that HOPE GMs should last
well beyond the 30-year post closure period required in many environmental regulations
without any measurable degradation of mechanical properties.  Clearly, a service
lifetime measured in at least hundreds of years appears to be achievable.  It is hoped
that the results of the ongoing study will allow even better estimates of GM service
lifetime in the near future.

B-12 References
Accorsi, J. and Romero, E., (1995) "Special Carbon Blacks for Plastics" Plastics
   Engineering, April  1995, pp. 29-32.
ASTM D638 - Test Method for Tensile Properties of Plastics.
ASTM D883 - Terminology Relating to Plastics.
ASTM D1238 - Test Method for Flow Rates of Thermoplastics by Extrusion Plastomer.
ASTM D1505 - Test Method for Density of Plastics by the Density-Gradient Technique.
ASTM D1603 - Test Method for Carbon Black in Olefin Plastics.
ASTM D3895 - Test Method for Oxidative Induction Time of Polyolefins by Differential
   Scanning Calorimetry.
ASTM D5397 - Test Method for Evaluation of Stress Crack Resistance of Polyolefin
   Geomembranes Using Notched Constant Tensile Load Test
ASTM D5885 - Test Method for Oxidative Induction Time for Polyolefin Geosynthetics
   by High Pressure Differential Calorimetry.
Billingham, N.C. and Calvert, P.O., (1986), "The Physical Chemistry of Oxidation and
   Stabilization of Polyolefins", Developments in Polymer Stabilization - 3,  Chapter 5,
   edited by Scott, G., Published by Applied Science Publishers Ltd, London, pp. 139-
   190.
Chu, P.P. and Hsieh, E. T.,  (1992), "13C and 1H Nuclear Magnetic Resonance
   Methodologies in Industrial Polymer Research and Production", MQC/MQA and
   CQC/CQA of Geosynthetics, Edited by Koerner, R.M. and Hsuan, Y.G.,  GRI
   Conference Series, Published by IFAI, St Paul, MN., pp. 234-243.
Fay, J. J. and King R. E., (1994),  "Antioxidants for Geosynthetic Resins and
   Applications",  Geosynthetic Resins, Formulations and Manufacturing,  Edited by
   Hsuan, Y.G. and Koerner,  R.M., GRI Conference Series, Published by IFAI, St Paul,
   MN., U.S.A., pp. 77-96.
Grassie, N. and Scott, G., (1985), Polymer Degradation and Stabilization, Published by
   Cambridge  University Press, New York, U.S.A.
Howard, J.B., (1973),  "Data for Control of Stability in Polyolefin Wire and Cable
   Compounds",  Polymer Engineering Science, Vol. 13, No. 6, pp. 429-434.
Koerner, G.R. and Koerner, R.M., (1995), "Temperature Behavior of Field Deployed
   HOPE Geomembranes",  Geosynthetics '95 Conference Proceedings, Nashville,
   TN., pp. 921-938.
Kikkawa, K., Nakahara, Y., and Ohkatsu, Y., (1987), "Antagonism Between Hindered
   Amine Light Stabilizers and Sulfur-Containing Compounds", Polymer Degradation
   Stabilization, Vol. 18, pp. 237-245.
                                     B-38

-------
Luston, J., (1986), "Physical Loss of Stabilizers from Polymers", Developments in
   Polymer Stabilization - 2, Chapter 5, edited by Scott, G., Published by Applied
   Science Publishers Ltd, London, pp. 185-240.
Mitchell, D.H. and Spanner, G.E., (1985), Field Performance Assessment of Synthetic
   Liners for Uranium Tailings Ponds, Status Report, Battelle PNL, U.S. NRC,
   NUREG/CR-4023, PNL-5005.
Michaels, A.S. and Bixler,  H.J., (1961), "Solubility of Gases in Polyethylene", Journal of
   Polymer Science, 50, Vol. L, pp. 393-412.
Osawa, Z. and Ishizuka, T., (1973), "Catalytic Action of Metal Salts in Autoxidation and
   Polymerization. (X) - The Effect of Various Methyl Stearates on the Thermal
   Oxidation of 2,6,10,14-Tetramethylpentadecane", Journal of Applied Polymer
   Science,  Vol.  17, pp. 2897-2907.
Petermann, J., Miles, M. and Gleiter, H., (1976), "Growth of Polymer Crystals During
   Annealing", Journal of Macromolecular Science - Physics, B12(3), pp. 393-404.
Poland, F.G. and Harper, S.R., (1986), "Critical Review and Summary of Leachate and
   Gas Production from Landfills - Final Report", U.S. EPA Cooperative Agreement,
   CR 908 997.
Rapoport,  N.Ya. and Zaikov, G.E., (1986), "Kinetics and Mechanism of the Oxidation of
   Stressed Polymer", Developments in Polymer Stabilization - 4, Chapter 6, edited by
   Scott, G., Published by Applied Science Publishers Ltd, London, pp. 207-258.
Smith, G.D., Karlsoon, K. and Gedde, U.W., (1992), "Modeling of Antioxidant Loss From
   Polyolefins in  Hot-Water Applications. I; Model and Application to Medium Density
   Polyethylene Pipes", Polymer Engineering and Science, Vol. 32, No. 10, pp. 658-667.
Struik, L.C.E., (1978), Physical Aging in Amorphous Polymers and Other Materials,
   Elsevier Scientific Publishing Company, Amsterdam, The Netherlands.
Struve, F., (1994), "Extrusion of Geomembranes", Geosynthetic Resins, Formulations
   and Manufacturing, Edited by Hsuan, Y.G. and Koerner, R.M., GRI Conference
   Series, Published by IFAI, St  Paul, MN., U.S.A., pp. 77-96.
Tikuisis, T., Lam, P., and Cossar, M., (1993), "High Pressure Oxidative Induction Time
   Analysis by Differential Scanning Calorimetry", MQC/MQA and CQC/CQA of
   Geosynthetics, Edited by Koerner, R.M. and Hsuan, Y.G., GRI Conference Series,
   Published by IFAI, St Paul, MN., pp. 191-201.
Viebke, J., Elble,  E., Ifwarson, M. and Gedde, U.W. (1994), "Degradation of
   Unstabilized Medium-Density Polyethylene Pipes in Hot-Water Applications",
   Polymer Engineering and science, Vol. 34,  No. 17, pp. 1354-1361.
Yazadini, R., Campbell, J.L. and  Koerner, G.R., (1995), "Long-Term In-Situ Strain
   Measurements of a High Density Polyethylene Geomembrane in a Municipal Solid
   Waste  Landfill", Geosynthetics '95 Conference Proceedings, Nashville, TN., pp. 893-
   906.

B-13  Acknowledgments
Due to the nature of long-term research, which is typified by this 10-year study,
financing by a number of agencies and sources is necessary. This particular task has
commingled funds in the form of  partnering by different organizations.  The authors
express their sincere appreciation to all of the following:
                                     B-39

-------
1.  US Environmental Protection Agency, via its Risk Reduction Engineering Laboratory
   under Cooperative Agreement No. CR 821448.  Mr. Robert Landreth (retired) and
   Mr. David A. Carson are the Project Officers.

2.  National Science Foundation, via its Geomechanical, Geotechnical and Geo-
   Environmental Systems (G3S) program under Grant No. CMS-9312772.
   Dr. Priscilla P. Nelson is the Project Officer.

3.  The consortium of Geosynthetic Research Institute (GRI) member organization via a
   portion of their membership fees. A listing of the current members is as follows:

GSE Lining Systems, Inc.-  William W. Walling/Melody Adams
Earth Tech Consultants,  Inc. - Walt Studebaker/Charles P. Ballod
U.S. Environmental Protection Agency -David A. Carson
Polyfelt GmbH - Gernot Mannsbart/Philippe Delmas
Browning-Ferris Industries - Charles Rivette/Dan Spikula [BoD]
E.  I. duPont de Nemours & Co., Inc. - John L. Guglielmetti/Ronald J.  Winkler
Federal Highway Administration -Albert F. DiMillio/Jerry A. DiMaggio
Tensar Earth Technologies, Inc. - Peter J. Vanderzee/Donald G. Bright/Mark H. Wayne
National Seal Co. - Gary Kolbasuk [BoD]/George Zagorski
Poly-Flex, Inc. - James Nobert/George Yazdani
Akzo Nobel Geosynthetics  Co. - Wim Voskamp/Joseph Luna
Phillips Petroleum Co. - Rex L. Bobsein [BoD]
GeoSyntec Consultants Inc. - Jean-Pierre Giroud/James A. McKelvey Ill/Majdi Othman
NOVA Chemicals Ltd. - Nolan Edmunds
Tenax,  S.p.A. - Pietro Rimoldi [BoD]/Aigen Zhao
Amoco Fabrics and Fibers Co. - Gary Willibey
U.S. Bureau of Reclamation -Alice I. Comer/Jack Haynes
EMCON - Donald E. Hullings/Mark A. Swyka
Montell USA, Inc. - Robert G. Butala
TC Mirafi, Inc. - Michael M. Koutsourais/Dean Sandri
CETCO - James T. Olsta
Huesker,  Inc. - Thomas G.  Collins
Solvay Polymers - J. Michael Killough
Naue-Fasertechnik GmbH - Georg Heerten/Kent von Maubeuge
Synthetic Industries, Inc. - Marc S. Theisen/Deron N. Austin
Mobil Chemical Co. - PerK. Husby/Frank J. Velisek
BBA Nonwovens - John Matheny/Geoff Kempton
NTH Consultants, Ltd. - Jerome C. Neyer/Robert Sabanas
TRI/Environmental, Inc. - Sam R. Allen [BoD]/Richard Thomas
U.S. Army Corps of Engineers - David L.  Jams [BoD]
Chevron Chemical Co. - Pamela L. Maeger
                                    B-40

-------
Serrot Corp. - Robert A. Otto/Bill Torres [BoD]
Union Chemical Lab (ITRI) - Frank L. Chen
Haley and Aldrich, Inc. - Richard P. Stulgis [BoD]
Westinghouse-Savannah River - Michael Hasek
URS/Greiner/WCC - Pedro C. Repetto/John C.  Volk
S. D. Enterprise Co.,  Ltd. - David Eakin
Solmax Geosynthetiques - Robert Denis
EnviroSource Treatment & Disposal Services, Inc. - Patrick M. McNamara
Strata Systems, Inc. - John N. Paulson [BoD]
CARPI, Inc. -Alberto M. Scuero/JohnA. Wilkes
Rumpke Waste Service, Inc. - Bruce Schmucker
Civil & Environmental Consultants, Inc. - Richard J. Kenter
Firestone Building Products Co. - H. Joseph Kalbas
FITI (GSI-Korea) - Han-Yong Jeon
Waste Management Inc. - James R. (Ron) Jones
CETCO Europe, Ltd.  - Archie Filshill
                                    B-41

-------
                   Appendix C
Field  Performance  Data for Compacted
                   Clay  Liners
                          by

                  David E. Daniel, Ph. D., P.E.
                     University of Illinois
                     Urbana, IL 61801
                      performed under

               EPA Cooperative Agreement Number
                      CR-821448-01-0
                       Project Officer

                     Mr. David A. Carson
            United States Environmental Protection Agency
               Office of Research and Development
           National Risk Management Research Laboratory
                    Cincinnati, OH 45268

-------
                                 Appendix C
           Field Performance Data for Compacted Clay Liners

C-1 Introduction
The  performance of  compacted  clay  liners (CCLs)  constructed from  natural soil
materials and soil-bentonite blends was discussed  in Chapter 4.  A number of graphs
were presented correlating various parameters.

This appendix contains a summary of the data used in compiling the results presented
in  Chapter 4.   The  data are presented in  this appendix  in the form of tables  of
information.  The intent is to provide sufficient information so that future researchers can
add newly acquired data to the database and perform new analyses.  Also, data on
statistical variability  of certain parameters was  collected  and is  summarized  in this
appendix.

C-2 Data for Natural Soil Liner Materials
The data for natural soil liner materials are presented in four attached tables:

   •   Table C-1:  Material properties
   •   Table C-2:  Construction information
   •   Table C-3:  Quality assurance information
   •   Table C-4:  Hydraulic conductivity data

Each of the 89 sites is given a site number, which is shown  in column 1 of all the tables.
The symbols used are defined as follows:

    Clay Fraction = percent on a dry weight basis finer than 2 jim
    DF = maximum depth of penetration of wetting front into soil liner
    i = hydraulic gradient
    k = hydraulic conductivity
    L = thickness of soil liner
    LL = liquid limit of the soil
    MP = modified Proctor (ASTM D-1557)
    OWC = optimum water content
    Percent Fines = percent on a dry weight basis passing  the No. 200 sieve
    Percent Gravel = percent on a dry weight basis retained on the No. 4 sieve
    PI = plasticity index of the soil
    P0 = percent of (w,yd) points lying on or above the line of optimums
    RC = relative compaction (dry unit weight of compacted soil  divided by maximum
        dry unit weight from laboratory compaction test)
    RP =  reduced Proctor (less than the compactive effort from SP)
    ASj = degree of saturation of compacted soil minus degree of saturation on the line
        of optimum for the same dry unit weight
    SP = standard Proctor (ASTM D-698)
                                      c-1

-------
     TSB = two-stage borehole test
     w = water content as a percentage
     wopt = optimum water content
     Yd = dry unit weight
     Yd,max = maximum dry unit weight
     o' = effective stress in kPa
     \l/o = initial suction of soil liner

Some of the columns of data contain three data entries, one above the other, with the
following meaning:

   •  Upper number is the number of data points
   •  Middle number is the average  (geometric mean for hydraulic conductivity)
   •  Lower number is the standard  deviation

C-3  Data for Soil-Bentonite Admixed Liners
Data for soil-bentonite admixed liners are presented in tables as follows:

   •  Table C-4:  Material properties
   •  Table C-5:  Construction information
   •  Table C-6:  Quality assurance  data
   •  Table C-7:  Hydraulic conductivity data

The symbols are the same as those given in section C-2.
                                      c-2

-------
    Table C-1. Material Properties for Natural Clay Liner Materials in Database.
Site
No.
1


2

3




4
5

6

7


8

Location
and Date
SCA
Wilsonville, IL
Oct. 1992
Confidential

Keele Valley
Toronto, OT
1990
PADB
Livingston, LA
1987
Confidential

Confidential

Imperial, PA
Dec. 1990

Confidential

Source of Data
Benson et al.
1992

Benson & Boutwell
1992
Lahtietal. 1987
Readesetal. 1990

Johnson et al.
1990

Benson & Boutwell
1992
Benson & Boutwell
1992
GeoSyntec Report


Benson & Boutwell
1992
LL (%)
-
24
-
-
58
-
25
-
9
50
3
-
43
-
32
8
33
1
-
35
PI (%)
-
10
-
-
29
-
10
-
9
34
3
-
26
-
19
8
13
1
-
22
Percent
Gravel
-
4
-
-
;
-
2
-
-
0
-
-
3
3

8
5
2
-
1
Percent
Fines
-
65
-
-
85
-
85
-
-
95
-
-
87
-
88
8
77
6
-
75
Clay
Fraction
(%<2 urn)
-
37
-
-
50
-
22
-
-
47
-
-
32
-
35
8
27
2
-
45
wopt
(%)
10.2
9.0

26.8

12.3


17.9


14.3

13.5

14.1


14.5

Yd, max
(kN/m3)
20.1
21.3

14.6

19.0


16.8


18.6

19.5

18.6


18.8

Compactive
Effort
SP
MP

SP

SP


SP


MP

MP

SP


MP

o

-------
    Table C-1.  Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
9

10

11



12
13


14


15


16


17


Location
and Date
Sauk City, Wl
1988
Portage, Wl
1988
Marathon, Wl
1988
Marathon, Wl
1988

Imperial, PA
April 1991

Test Fill 2
July 1988

Confidential


Test Fill 1


Livingston, LA
Pad A
Oct. 1988
Source of Data
Gordon et al.
1989
Gordon et al.
1989
Gordon et al.
1989
Gordon et al.
1989

GeoSyntec Report


Mundell & Boos
1990

Benson & Boutwell
1992

Mundell & Boos
1990

Johnson et al.
1990

LL (%)
-
55
32
43
-
57
-
55
-
8
37
1
3
40
8
12
85
3
24
41
9
9
50
3
PI (%)
-
31
32
21
-
30
-
28
-
8
15
1
3
20
3
12
58
3
24
22
6
9
34
3
Percent
Gravel
-
4
-
1
-
-
-
-
-
8
2
1
6
0
-
1
0
-
20
0
-
-
0
-
Percent
Fines
-
_
-
~
-
-
-
-
-
8
78
5
6
70
8
1
99
-
20
77
-
-
95
-
Clay
(%<2 urn)
-
45
-
29
-
39
-
33
-
8
37
4
6
25
11
-
57
-
20
38
8
-
47
-
wopt
(%)
12.7

16.6

21.7

23.0


18.0


16.2


25.8


15.8


20.3


Yd, max
(kN/m3)
18.6

18.7

17.3

16.6


17.0


16.7


14.6


17.0


16.4


Compactive
Effort
MP

MP

MP

MP


SP


SP


SP


SP


SP


o

-------
    Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
18

19


20



21
22

23


24

25

Location
and Date
Tangipahoa Landfill
Amite, LA
March 1992
Confidential


Confidential

Confidential



1993
Green County, Wl
1987

Confidential

Confidential

Source of Data
Boutwell &
McManis
1995
Benson & Boutwell
1992

Personal Files

Trast
1993

Othman &
Luettich 1 994
Krantz & Bailey
1990

Trast
1993
Trast
1993
LL (%)
10
30
6
12
32
3
-
49
-
51
-
-
63
20
39
4.2
-
67
-
53
PI (%)
10
18
4
12
14
2
-
23
-
26
-
-
42
20
18
3.8
-
46
-
41
Percent
Gravel
5
0
-
12
1
1
-
1
-
1
-
-
;
-
-
-
-
0
-
0
Percent
Fines
5
52
3
12
85
3
-
94
-
90
-
-
96
-
73
-
-
94
-
88
Clay
Fraction
(%<2 urn)
5
16
1
12
44
4
-
43
-
36
-
-
;
-
30
-
53

-
36
wopt
(%)
13.0

10.5


18.5

11.8
18.0

20.5

20.0


16.0
21.5
11.5
16.1
Yd, max
(kN/m3)
18.7

20.1


17.2

18.5
17.0

16.3

16.5


18.4
16.3
19.8
18.0
Compactive
Effort
SP

MP


SP

MP
SP

SP

SP


MP
SP
MP
SP
o
CJ1

-------
   Table C-1.  Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
26


27


28




29
30

31


32

33

34

Location
and Date
Confidential


Confidential


Confidential


ERG Facility
Milan, Ml
1993
Confidential

Confidential


Confidential

Confidential

Confidential

Source of Data
Trast
1993

Trast
1993

Trast
1993

Bergstrom et al.
1995

Personal Files

Personal Files


Personal Files

Personal Files

Personal Files

LL (%)
-
33
-
-
31
-
-
35
-
-
27
-
-
32
7
40
1
-
45
-
29
-
44
PI (%)
-
19
-
-
18
-
-
19
-
-
10
-
-
19
7
24
1
-
27
-
15
-
16
Percent
Gravel
-
7
-
-
8
-
-
3
-
-
2
-
-
;
7
7
4
-
0
-
1
-
0
Percent
Fines
-
85
-
-
74
-
-
89
-
-
76
-
-
;
7
58
3
-
99
-
87
-
96
Clay
Fraction
(%<2 urn)
-
37
-
-
26
-
-
41
-
-
28
-
-
;
7
23
1
-
42
-
40
-
_
wopt
(%)
12.2
17.5
18.5
12.5
16.5
18.5
11.5
16.6
18.5
9.0
13.0
14.4
14.0

12.4


11.0

13.3


17.3
Yd, max
(kN/m3)
19.3
17.7
17.1
19.4
17.8
17.2
19.4
17.5
17.0
20.5
19.1
18.6
18.6

19.3


19.9

18.9


17.1
Compactive
Effort
MP
SP
RP
MP
SP
RP
MP
SP
RP
MP
SP
RP
MP

SP


MP

MP


SP
o
en

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
35

36


37




38
39


40


41



42


Location
and Date
Confidential

Confidential


Indianapolis, IN
1994

ISGS Prototype
Urbana, IL
1986
ISGS Field-Scale
Urbana, IL
April 1988
Confidential


BP Chemicals
SDRI 1
Port Lavaca, TX
Nov. 1988
Celanese
Bishop, TX
July 1986
Source of Data
Personal Files

Personal Files


Personal Files


ISGS
Report

ISGS
Report

Personal Files


McBride-Ratcliff
Report


Personal Files


LL (%)
-
39
9
36
2.5
3
36
3
-
21
-
-
21
-
15
101
5
-
47
_

3
69
3.6
PI (%)
-
19
9
17
1.6
3
17
2
-
7
-
-
7
-
15
71
5
-
30
_

3
45
3.0
Percent
Gravel
-
0
9
2
1.6
3
10
5
-
9
-
-
9
-
-
0
-
-
.
_

2
0
0
Percent
Fines
-
97
9
74
2.6
3
48
3
-
60
-
-
60
-
-
98
-
-
66
_

3
79
3.0
Clay
Fraction
(%<2 urn)
-
-
9
30
3.4
3
16
1
-
26
- (4 urn)






-
.
_

2
49
4.2
wopt
(%)

22.2

13.2


12.4

10.3


10.3


31.6


19.5



23.4


Yd, max
(kN/m3)

17.7

18.3


19.0

20.4


20.4


13.4


16.3



15.1


Compactive
Effort

SP

SP


SP

SP


SP


SP


SP



SP


o

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
43


44




45
46


47


48


49


Location
and Date
GCWDA Test Fill A
Texas City, TX
Nov. 1988
GCWDA Test Fill B
Texas City, TX
Nov. 1988
Texas Eastman
Longview, TX
1987
Puckett Plant
Ft. Stockton, TX
April 1988
Shell
Deer Park, TX
Dec. 1988
Confidential


Confidential


Source of Data
Personal Files


Personal Files


H.B. Zachry Co.
Report

Personal Files


Personal Files


Personal Files


Personal Files


LL (%)
119
62
4.1
119
62
4.1
8
44
4.0
31
35
1.8
41
39
4.7
60
41
2.7
60
42
1.7
PI (%)
119
42
4.3
119
42
4.3
8
28
3.2
31
16
2.3
41
24.3
5.2
60
23
-
60
22
-
Percent
Gravel
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Percent
Fines
119
86
6.2
119
86
6.2
8
70
2.4
31
98
1.1
41
69.5
-
-
86
-
-
86
-
Clay
Fraction
(%<2 urn)
-
-
-
-
-
-
-
-
-
2
22
8.4
-
-
-
-
-
-
-
-
-
wopt
(%)
22.4


22.4


19.5


23.3


14.6


20.0
18.0
13.3
20.0
18.0
13.3
Yd, max
(kN/m3)
15.4


15.4


16.4


15.4


17.7


16.2
16.7
18.7
16.2
16.7
18.7
Compactive
Effort
SP


SP


SP


SP


SP


RP
SP
MP
RP
SP
MP
o
oo

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).

Site
No.
50


51


52




53
54



55



56



57




Location
and Date
Confidential


Confidential


Emelle, AL
Oct. 1984

Confidential


Quarantine Rd
Landfill
Baltimore, MD
Jan. 1994
Savannah
River Plant Panel
A1
March 1988
Savannah
River Plant Panel
A2
March 1988
Savannah
River Plant Panel
B1
March 1988


Source of Data
Personal Files


Personal Files


Colder Assoc.
Report

Personal Files


Personal Files



Mueser-Rutledge
Report


Mueser-Rutledge
Report


Mueser-Rutledge
Report




LL (%)
88
43
3.4
62
40
1.9
2
37
1.4
8
54
2
-
_
.

-
66
.

-
66
_

-
69
_



PI (%)
88
24
-
62
22
-
2
18
2.8
8
31
3
-
_
.

-
35
.

-
35
_

-
38
_


Percent
Gravel
-
-
-
-
-
-
3
10
8.9
8
0
0
-
_
.

-
0
.

-
0
_

-
0
_


Percent
Fines
-
86
-
-
86
-
3
73
15.1
-
-
-
-
_
.

-
93
.

-
93
_

-
98
_

Clay
Fraction
(%<2 urn)
-
-
-
-
-
-
3
38
9.0
8
40
3
-
_
.

-
-
.

-
-
_

-
-
_


wopt
(%)
20.0
18.0
13.3
20.0
18.0
13.3
19.9


19.9


-
_
.

27.4



27.4



26.8




Yd, max
(kN/m3)
16.2
16.7
18.7
16.2
16.7
18.7
16.5


16.4


-
_
.

14.5



14.5



14.6




Compactive
Effort
RP
SP
MP
RP
SP
MP
SP


SP


-
_
.

SP



SP



SP



o
CD

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).

Site
No.
58



59





60

61



62



63



64




Location
and Date
Savannah River
Plant
Panel B2
March 1988
Savannah
River Plant Panel
B3
March 1988
Savannah
River Plant Panel
C1
March 1988
Savannah
River Plant Panel
C2
March 1988
Savannah
River Plant Panel
D1
March 1988
Savannah
River Plant Panel
D2
March 1988
BP Chemicals
Port Lavaca, TX
SDRI2
Dec. 1988


Source of Data
Mueser-Rutledge
Report


Mueser-Rutledge
Report


Mueser-Rutledge
Report


Mueser-Rutledge
Report


Mueser-Rutledge
Report


Mueser-Rutledge
Report


McBride-Ratcliff
Report




LL (%)
-
69
_

-
69
_

-
68
-

-
68
_

-
51
_

-
51
.

-
47
_



PI (%)
-
38
_

-
38
_

-
35
-

-
35
_

-
20
_

-
20
.

-
30
_


Percent
Gravel
-
0
_

-
0
_

-
0
-

-
0
_

-
0
_

-
0
.

-
-
_


Percent
Fines
-
98
_

-
98
_

-
95
-

-
95
_

-
73
_

-
73
.

-
66
_

Clay
Fraction
(%<2 urn)
-
-
_

-
-
_

-
-
-

-
-
_

-
-
_

-
-
.

-
-
_


wopt
(%)
26.8



26.8



26.6



26.6



20.2



20.2



19.5




Yd, max
(kN/m3)
14.6



14.6



14.6



14.6



15.9



15.9



16.3




Compactive
Effort
SP



SP



SP



SP



SP



SP



SP



o
o

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
65


66



67
68


69


70

71

72


Location
and Date
BP Chemicals
Port Lavaca, TX
SDRI3
Dec. 1988
Confidential

Confidential


Confidential


Confidential


Confidential

Confidential

SDDS Longtree LF
Igloo, SD
Feb. 1990
Source of Data
McBride-Ratcliff
Report

Personal Files

Personal Files


Personal Files


Personal Files


Personal Files

Personal Files

S. Dakota Disposal
Systems Report

LL (%)
-
47

-
50
-
49
-
4
35
1
4
22
1
-
42
-
29
3
36
2.0
PI (%)
-
31

-
29
-
27
-
4
17
1
4
9
1
-
26
-
19
3
20
1.7
Percent
Gravel
-
-

-
;
-
-
-
4
2
2
4
6
3
-
0
-
4
-
0
-
Percent
Fines
-
66

-
75
-
62
-
4
67
10
4
50
2
-
88
-
83
-
85
-
Clay
Fraction
(%<2 urn)
-
-

-
;
-
-
-
4
22
4
4
16
1
-
45
-
34
-
35
-
wopt
(%)
13.5


19.0

19.3


14.8
11.5

10.0
8.5

14.9

12.2

18.0


Yd, max
(kN/m3)
19.2


16.1

16.1


17.7
19.0

19.9
21.4

18.7

19.6

16.5


Compactive
Effort
MP


SP

SP


SP
MP

SP
MP

MP

MP

SP


9

-------
   Table C-1.  Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
73


74


75



76
77


78


79


80


Location
and Date
Sea Drift, TX
Sept. 1988

Sea Drift, TX
Sept. 1988

McClellandtown, PA
Sept. 1990
Confidential


Arnoni LF Pad 1
Pittsburgh, PA
Feb. 1994
Arnoni LF Pad 2
Pittsburgh PA
Feb. 1994
DuPont Pad 1
Victoria, TX
Jan. 1989
DuPont Pad 2
Victoria, TX
Jan 1989
Source of Data
McClelland
Engineers Report

McClelland
Engineers Report

Cumberland Geot.,
Consultants Report
Personal Files


Personal Files


Personal Files


Engineering
Sciences Report

Engineering
Science Report

LL (%)
4
76
6.0
4
56
6.0
-
;
4
37
3.0
45
32
0.8
45
32
1.6
12
62
5.3
17
52
1.3
PI (%)
4
53
6.0
4
40
5.0
-
;
4
17
2.0
45
13
0.9
45
16
1.3
12
41
4.9
17
35
1.2
Percent
Gravel
-
-
-
-
-
-
-
;
-
1
-
-
-
-
-
-
-
-
-
-
-
-
-
Percent
Fines
-
-
-
-
64
-
-
;
-
92
-
-
-
-
-
-
-
9
82
3.2
15
84
6.0
Clay
Fraction
(%<2 urn)
-
-
-
-
-
-
-
;
-
-
-
-
=19
-
-
=25
-
-
-
-
-
-
-
wopt
(%)
21.0


18.0


21.0

19.2


9.9


11.5


25.0
17.8

19.6
14.4

Yd, max
(kN/m3)
15.5


16.9


15.6

16.6


19.7


19.6


14.9
16.5

15.9
18.0

Compactive
Effort
-


-


SP

SP


SP


SP


SP
MP

SP
MP

9
ro

-------
   Table C-1.  Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
81


82
83
84

85

86


87


Location
and Date
GE
Waterford, NY
July 1989
Findlay Township,
PA
Aug. 1988
Findlay Township,
PA
Aug. 1988
Montezuma Hills,
CA
Pad A (Dark)
Feb. 1991
Montezuma Hills,
CA
Pad B (Light)
Feb. 1991
Fernald, OH
Pad 1 (Ln. 1)
Nov. 1996
Fernald, OH
Pad 1 (Ln. 2)
Nov. 1996
Source of Data
Clough Harbor &
Assoc. Report

Paul Rizzo
& Assoc. Report
Paul Rizzo
& Assoc. Report
IT Corp.
Report

IT Corp.
Report

GeoSyntec Report


GeoSyntec Report


LL (%)
-
47
-
-
4
39
3.6
-

-

-
43
-
-
43
-
PI (%)
-
22
-
-
4
16
1.9
-

-

-
24
-
-
24
-
Percent
Gravel
-
-
-
-
4
10
3.3
-

-

-
-
-
-
-
-
Percent
Fines
-
-
-
4
84
5.4
4
81
2.7
-

-

-
84
-
-
84
-
Clay
Fraction
(%<2 urn)
-
-
-
4
54
6.7
4
48
8.6
-

-

-
37
-
-
37
-
wopt
25



18.2




17.7


17.7


Yd, max
(kN/m3)
15.3



17.6




17.1


17.1


Compactive
Effort
SP



SP




SP


SP


9
CO

-------
   Table C-1. Material Properties for Natural Clay Liner Materials in Database (Continued).
Site
No.
88


89


Location
and Date
Fernald, OH
Pad2(Ln. 1)
Nov. 1996
Fernald, OH
Pad 2 (Ln. 2)
Nov. 1996
Source of Data
GeoSyntec Report


GeoSyntec Report


LL (%)
-
25
-
-
25
-
PI (%)
-
14
-
-
14
-
Percent
Gravel
-
-
-
-
-
-
Percent
Fines
-
70
-
-
70
-
Clay
Fraction
(%<2 urn)
-
29
-
-
29
-
wopt
(%)
11.6


11.6


Yd, max
(kN/m3)
19.1


19.1


Compactive
Effort
SP


SP


9

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database.
Site
No.
1

2
3

4

5



6
7

8

9

10

11

12

13

Compaction Criteria
w>OWC
RC > 90% MP
None
w>OWC
RC > 95% SP
w > OWC + 2, <+8
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP

w > OWC + 2
RC > 90% MP
w > OWC -2 to +4
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w > OWC + 2, <+5
RC > 90% MP
Compactor
CAT 825

Bomag 210PD
Rex 370

CAT 81 5

RexTrashmaster

CAT 825


CAT 825

Dynapac CT25

-

-

-

-

CAT 825

Compactor
Mass (kg)
32,400

-
30,000

19,800

36,000

32,400


32,400

12,600

-

-

-

-

32,400

Passes per
Lift
6

6
4

-

6

5


4

4

-

-

-

-

4

Lift Thickness
(mm)
150

150
150

150

150

150


150

150

150

150

150

150

150

Number of
Lifts
6

5
8

4

10

6


8

6

10

10

10

10

8

Pad Size
(m x m or m2)
36x15

32x14
30x30

15 x30

Liner

29x12


15x24

24x18

Liner

Liner

Liner

Liner

15x24

9
en

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Compaction Criteria
w > OWC + 2, <+5
RC > 90% MP
w>OWC
R0100%SP
w > OWC + 2, <+5
RC > 90% MP
w > OWC + 2, <+6
RC > 90% SP
Sj>78.5
RC > 90% MP
w>OWC
RC > 90% MP
Sj>82.0
w>OWC
RC > 95% SP
Sj>85.0
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 90% MP
Compactor
-
-
CAT 835
CAT 81 5
CAT D7G bulldozer
CAT 825
CAT 825
CAT 825
-
-
CAT815A
CAT815A
Dynapac CA25
Dynapac CA25
Compactor
Mass (kg)
-
-
39,000
19,800
25,000
32,400
32,400
32,400
-
-
18,900
18,900
18,900
18,900
Passes per
Lift
-
-
-
6
4
5
8
6
-
-
8-12
8- 12
4-6
4-6
Lift Thickness
(mm)
170
200
170
150
150
150
150
150
-
150
150
150
150
150
Number of
Lifts
6
7
5
4
5
10
6
6
-
10
5
5
6
6
Pad Size
(m x m or m2)
9x14
12x26
9x9
15x30
30x12
Liner
45x20
58x26
-
Liner
31 x15
31 x15
27x17
27x17
9
en

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
28
29
30
31
32
33
34
35
36
37
38
39
40
Compaction Criteria
w>OWC
RC > 90% MP
w > OWC-2, +5
RC > 90% MP
RC > 90% MP
w>OWC
RC > 95% SP
w > OWC, <+6
RC > 90% MP
w>OWC
RC > 90% MP
w>OWC
RC > 95% SP
w>OWC
RC > 95% SP
w > OWC, <+6
RC > 95% SP
w>OWC
RC > 95% SP
w>OWC
RC > 90% MP
w: 11 to 12%
RC > 90% SP
w > OWC, +5
RC > 92% SP
Compactor
Dynapac CA25
CAT 825
RexTrashmaster
CAT S563
CAT 825
CAT815B
IR SPF-56 & CAT
815
CAT 824B
CAT 81 5
FWD741
Hyster C852A
CAT815B
sheepsfoot
Compactor
Mass (kg)
18,900
32,400
27,000
-
32,400
19,800
19,800
32,400
17,100
-
-
19,800
59 kg/lin. cm
Passes per
Lift
4-6
6
-
-
-

4
(2 each)
6
8
4
12
12
-
Lift Thickness
(mm)
150
170
150
300
150
150
150
150
60
60
150
130
150
Number of
Lifts
6
9
6
2
8
8
3
6
6
5
6
6
6
Pad Size
(m x m or m2)
27x17
32x16
-
13X26
15x40
15x30
-
-
15x30
-
3x9
14.6x7.3
8x26
9

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Compaction Criteria
RC > 95% SP
-
w > OWC+1
RC > 95% SP
w > OWC+1
RC > 90% SP
w > OWC, <+2
RC > 95% SP
w> OWC+1, <+3
RC > 95% SP
w > OWC+1 , <+5
RC > 90% SP
w > OWC, <+3
RC > 95% SP
w > OWC, <+3
RC > 95% SP
w > OWC, <+3
RC > 95% SP
w > OWC, <+3
RC > 95% SP
w > OWC, <+3
RC > 95% SP
-
-
-
Compactor
CAT 81 5
wedgefoot
IR SPF-48
IR SPF-48
CAT 81 5 &Bomag
BW213PD
Dynapac CA25
CAT815B
CAT815B
CAT815B
CAT815B
CAT815B
-
CAT 825C
-
CAT815B
Compactor
Mass (kg)
19,800
-
7,200
7,200
19,800
14,000
10,900
19,800
19,800
19,800
19,800
19,800
-
-
-
19,800
Passes per
Lifs
40

16
8
4
2
8
6- 10
5-8
3
3
4-5
-
4
-
12
Lift Thickness
(mm)
150
150
150
150
150
200 - 250
85
150
150
150
150
-
-
150
160
Number of
Lifts
4
4
5
5
4
5
10
4
4
4
4
-
4
4
4
Pad Size
(m x m or m2)
93
30x15
37x9
37x9
288
15x30
46x24
46x15
46x15
46x15
46x15
150
-
-
483
9
oo

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
Compaction Criteria
-
-
-
-
-
-
-
-
RC > 95% SP
RC>91% MP
w>OWC+1,<+5
RC > 95% SP
w>OWC+1,<+5
RC > 95% SP
w>OWC+1,<+5
RC > 95% SP
w>OWC+1,<+5
RC > 95% SP
w > OWC-2, <+4
RC > 90% MP
w > OWC-2, <+4
RC > 90% MP
w > OWC, <+6
RC > 95% SP
Compactor
Rex 3-50A & CAT
815B
CAT815B
CAT815B
CAT815B
CAT815B
CAT815B
CAT815B
CAT815B
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT815B
CAT815B
-
-
CAT 825C
Compactor
Mass (kg)
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
19,800
-
-
32,400
Passes per
Lift
6 (Lifts 1-3)
21 (Lift 4)
6
12
12
12
12
12
12
40
80
2
2
6
6
7 to 10
7 to 10
8
Lift Thickness
(mm)
130
140
150
170
170
190
150
230
150
150
100
100
150
150
150
150
150
Number of
Lifts
4
4
4
4
4
4
4
4
4
4
10
11
4
4
6
6
4
Pad Size
(m x m or m2)
483
483
483
483
483
483
483
483
93
186
12x26
12x26
15x36
15x36
24x18
24x18
18x36
9
CD

-------
   Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
73

74

75

76



77
78

79

80

81

82

83

85

Compaction Criteria
w > OWC,  95% SP
w > OWC,  95% SP
w > OWC+3, <+6
RC > 95% SP
w > OWC+3, <+5
RC > 95% SP
w > OWC+0.4
RC > 98% SP

w > OWC+1 .5
RC > 94% SP
w>OWC
RC > 95% SP
w>OWC
RC > 95% SP
w > OWC+4

RC > 96% SP

RC > 96% SP

RC > 90% MP

Compactor
CAT815B

CAT815B

CAT815B

CAT815B

IRSD-100D


IRSD-100D

CAT815B

CAT815B

Dresser VOS
PD84A
CAT 825 &
Vib. Smooth Drum
CAT 825 &
Vib. Smooth Drum
CAT815B

Compactor
Mass (kg)
19,800

19,800

19,800

19,800

10,200


10,200

19,800

19,800

16,200

32,400
-
32,400
-
19,800

Passes per
Lift
22

22

-

6

10


4

8

8

4

6
2
6
2
-

Lift Thickness
(mm)
1 @200
6@100
1 @200
6@100
1 @200
3@150
150

100


100

150

150

150

150

150

1 @300
3@200
Number of
Lifts
7

7

4

4

9


9

8

8

4

4

4

20x24

Pad Size
(m x m or m2)
12x23

12x23

15x30

18x30

15x30


15x30

465

465

465

223

223

-

s
o

-------
  Table C-2. Construction Information for Natural Clay Liner Materials in Database (continued).
Site
No.
85
86
87
88
89
Compaction Criteria
RC > 90% MP
W > OWC, <+4; RC > 95% SP
W > OWC, <+4; RC > 95% SP
W > OWC, <+4; RC > 95% SP
W > OWC, <+4; RC > 95% SP
Compactor
CAT815B
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
Compactor
Mass (kg)
19,800
19,800
19,800
19,800
19,800
Passes per
Lift
-
4
7
4
6
Lift Thickness
(mm)
1 @300
3@200
150
150
150
150
Number of
Lifts
20x24
6
6
6
6
Pad Size
(m x m or m2)
-
13x15
13x15
13x15
13x15
o
ro

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database.
Site
No.
1

2

3
4
5

6

7

8

9


w (%)
34
10.3
0.8
57
26.6
2.2
13.8
4
21.3
0.5
21
17.3
2.2
32
13.8
0.9
33
17.2
1.5
17
15.3
1.2
85
19.6
1.9

yd (kN/m3)
34
19.8
0.036
57
14.4
4.0
19.4
4
16.0
0.17
21
17.3
0.51
37
19.0
0.31
33
17.7
0.47
17
17.7
0.9
85
17.0
0.31

Po
44

28

98
80
95

32

88

8

90


ASj
-2.0

-4.0

+17.7
+3.0
-3.0

-8.2

+1.0

-12.6

+5.8


Distress
None

None

None
None
None

None

None

None

None


Purpose
Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'8 cm/s
Verify k< 1 x 10'7 cm/s ;
show KF = KL using standard
construction methods
Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s

Monitor liner performance and
Verify k< 1 x 10'7 cm/s


Remarks





Compacted slightly wet of modified
Proctor optimum and wet of line of
optimums


Met CQA Spec, but dry of line of
optimums







£
ro

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
10

11

12



13
14


15


16


17
18



w (%)
93
17.8
1.5
91
25.4
3.2
289
26.0
2.3
34
20.7
0.6
18
17.0
0.9
48
30.8
0.5
11
19.8
1.2
16
23.3
1.2
20
16.6
0.9

yd (kN/m3)
93
16.9
0.34
9100
16.0
0.66
289
16.1
0.50
34
16.7
0.24
18
16.8
0.16
48
14.1
0.27
11
16.1
0.22
16
15.7
0.44
20
17.35
0.25

Po
50

75

78


100


78

48
98
-

91

100

85


ASj
+3.5

-7.4

+4.6


+9.0


+4.0

48
+8.0
0.03

+6.0

+3.0

+4.0


Distress
None

None

None

None


None


None


None


None


None

Purpose
Monitor liner performance and
verify k< 1 x 10~7 cm/s

Monitor liner performance and
verify k< 1 x 10~7 cm/s

Monitor liner performance and
verify k< 1 x 10~7 cm/s

Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k < 1 x 10~7 cm/s; show
KF = KL using standard
construction methods
Verify k< 1 x 10~7 cm/s



Remarks






















£
CO

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
19

20

21

22

23
24

25

26
27

w (%)
584
13.6
0.72
37
17.6
0.52
18
19.5
0.3
-
60
22.0
2.8
19
23.6
1.1
18
18.9
0.31
53
15.5
36
13.5

yd (kN/m3)
584
19.0
0.19
37
16.9
0.33
18
16.9
0.15
-
60
16.4
0.57
19
15.8
0.32
18
16.9
0.42
53
17.6
36
18.0

Po
81

8

80

-
89

81

71

17
6

ASj
+3.8

-6.2

+4.3

-3.0
+7.4

+0.4

-0.5

-8.8
-10.4

Distress
None

None

None

None
None

None

None

None
None

Purpose
Monitor liner performance and
verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s

Veryfy Suitability of Si Ity
Material for k < 1 x 10~7 cm/s

Verify k< 1 x 10~7 cm/s
Monitor liner performance and
perify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s
Verify k< 1 x 10~7 cm/s

Remarks


Compacted dry of line of optimums
using acceptable zone approach



Spec, required that So > 90%






In spec, but dry of line of optimums
In spec, but dry of line of optimums
s

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
28

29


30



31
32

33

34

35

36


w (%)
54
16.2
92
13.9
0.71
-
16.2
-
13.1
-
-
13.9
-
13.4
-
17.8
-
20.7
-
15.5

yd (kN/m3)
54
17.7
92
18.8
0.28
-
18.6
-
19.1
-
-
19.2
-
18.7
-
17.1
-
16.8
-
17.6

Po

57

84


65

75


92

80

45

78

77

ASj

-0.3

+1.0


-2.3

-1.5


+7.5

+0.7

+2.5

+2.8

+3.4

Distress
None

None


None

None


None

None

None

None

None


Purpose
Verify K< 1 x 10~7 cm/s

Verify K< 1 x 1Q-7cm/s


Verify K< 1 x 1Q-7cm/s

Verify K< 1 x 1Q-7cm/s


Verify K< 1 x 1Q-7cm/s

Verify K< 1 x 10'7 cm/s

Verify K< 1 x 10~7 cm/s

Verify K< 1 x 1Q-7cm/s

Verify K< 1 x 1Q-7cm/s


Remarks
In spec, but straddles line of optimums

















Constructed with mine spoil

£
en

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
37

38


39




40
41


42


43


44


45



w (%)
-
14.1
24
11.5
3.6
57
11.6
1.1
40
35.5
0.5
13
21.9
0.88
26
25.0
1.5
49
23.4
1.12
49
24.2
1.5
31
19.8
1.03

yd (kN/m3)
-
18.2
24
20.4
0.77
57
17.9
0.82
40
12.8
0.17
13
16.0
0.22
26
15.1
0.36
49
15.4
0.28
49
15.0
0.40
31
103.8
2.61

Po

45
Raw
Data
NA

10


100


92


81


63


47


71


ASj

-1.2

+ 10.2


-19


+0.03


+5.6


+5.5


+3.9


+ 1.0


+0.4


Distress
None

None


None


Desiccation
(Hot HOPE)

None


None


None


None


None



Purpose
Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s



Remarks
Disturbance in tube by gravel
















East Side = A, West = B;
used light roller




Nothing unusual


£
en

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
46


47




48

49



50



51



52


53



w (%)
32
27.3
3.4
51
16.5
1.25
160
17.8
0.40

152
18.9

1.05
216
18.6

1.16
152
17.8

1.10
9
21.2
1.08
32
21.6
2.2

yd (kN/m3)
32
15.4
0.22
51
17.7
0.50
160
17.0
1.1

152
16.7

0.25
216
16.9

0.37
152
17.0

0.40
9
16.1
0.32
32
15.5
0.36

Po

100


100


75



86



84



73



67

32
71


ASj

+8.0


+9.9


1.1



0.3



1.3



-0.7



+ 1.5

32
+0.02
0.04

Distress
None


None


potentially
desiccation or
freeze-thaw
damage

potentially
desiccation or
freeze-thaw
damage
potentially
desiccation or
freeze-thaw
damage
potentially
desiccation or
freeze-thaw
damage
None


None



Purpose
Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s



Verify k< 1 x 10~7 cm/s



Verify k< 1 x 10'7 cm/s



Verify k< 1 x 10'7 cm/s



Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s



Remarks
Cell 1EhadSDRI,
Gs = 2.67 measured




Pads at Sites 48-51 were constructed
with same material by 4 different
contractors. Objective in each case to
obtain KF < 10~7 cm/s, with low bid/low
K contractor winning job.


















o
ro

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
54
55

56



57
58


59


60


61



w (%)
-
27.0
0.7
30.6
0.7
-
29.6
1.1
-
30.7
1.3
-
29.4
1.3
-
26.8
0.8
-
29.8
0.4

yd (kN/m3)
-
15.0
0.22
14.2
0.17
-
14.0
0.30
-
14.3
0.22
-
14.4
0.31
-
15.1
0.20
-
14.4
0.14

Po
-
100*

100*


100*


100*


100*


100*


100*


ASj
-
+5.9

+6.6


+0.5


+8.6


+6.3


+8.4


+7.6


Distress
None
None

None

None


None


None


None


None



Purpose
Verify k< 1 x 10~7 cm/s
Verify suitability of soil for
k< 1 x 10~7 cm/s;
w « wopt
Verify suitability of soil for
k< 1 x 1Q-7cm/s;
w = wopt+3%
Verify suitability of soil for
k < 1 x 1 0'7 cm/s

Verify suitability of soil for
k < 1 x 1 0'7 cm/s

Verify suitability of soil for
k < 1 x 1 0'7 cm/s

Verify suitability of soil for
k< 1 x 10'7 cm/s;
vv~wopt
Verify suitability of soil for
k< 1 x 10'7 cm/s;
w = wopt+3%

Remarks
SDRI Test on Liner
k>1 x 10-7 because soil wasn't wet
enough

Wetting soil up to opt. +3% lowered k
(compared to site 55)
















£
oo

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
62
63
64


65
66

67

68

69

70

w (%)
24.6
0.6
22.7
0.5
8
21.6
0.32
23
17.2
1.07
39
21.7
1.3
59
21.41.2

13
17.6
1.0
8
11.5
0.6
20.6

yd (kN/m3)
15.4
0.17
15.4
0.16
8
16.0
0.22
23
17.4
0.31
59
17.2
0.52
59
17.2
0.52
13
18.0
0.38
8
19.4
1.6
16.1

Po
100*
100*
88

0

95

98

100

75

60

ASj
+10.9
+3.2
+4.3

-6.1

+8.6

+8.1

+6.4

+8.0

-

Distress
None
None
None

None

None

None

None

None

None

Purpose
Verify suitability of soil for
k < 1 x 1 0'7 cm/s
Verify suitability of soil for
k < 1 x 1 0'7 cm/s
Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10" cm/s

Verify k< 1 x 10" cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10" cm/s

Remarks















£
CD

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
71
72
73


74
75

76

77

78


w (%)
14.3
6
23.7
1.0
36
25.2
1.2
30
19.6
1.6
7
25.4
2.0
76
21.8
0.4
111
11.0
0.6
109
12.4
1.5

yd (kN/m3)
18.0
6
15.5
0.2
36
14.8
0.86
30
16.1
0.02
7
15.2
0.7
76
15.9
0.3
111
19.2
0.6
109
18.8
0.5

Po
64
100
97

47

100

86

37

2


ASj
-
11.3
5

-3.7

11.1

1.1

1.1

-7.2


Distress
None
freeze-thaw,
but upper lift
re-worked
before SDRI
None

None

None

None

None

None


Purpose
Verify k< 1 x 10 cm/s
Verify k< 1 x 10 cm/s
Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s

Verify k< 1 x 10 cm/s


Remarks


On site clay

Off site clay









9
CO
o

-------
   Table C-3. Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
79


80


81




82
83


84
85
86


87



w (%)
39
28.2
3.1
37
23.1
2.2
8
28.0
2.5
14
17.8
2.0
16
19.3
1.5
-
-
18
20.5
2.3
19
20.4
2.2

yd (kN/m3)
37
16.2
0.49
39
14.9
0.46
8
14.2
0.20
14
17.1
0.44
16
17.3
0.53
-
-
18
16.6
0.68
19
16.6
0.61

Po
-


-


-


-


-


-
-

100


95


ASj
12.6


11.7


2.7


2.4


12


-
-

5.5


4.1


Distress
None


None


None


None


None


None
None

None


None


Purpose
Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s
Verify k< 1 x 10 cm/s
Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s



Remarks















Dark clay
Light clay






9
CO

-------
   Table C-3.  Quality Control/Quality Assurance Data for Natural Clay Liners in Database (continued).
Site
No.
88


89



w (%)
24
13.2
0.8
29
13.2
1.1

yd (kN/m3)
24
19.2
0.35
29
19.1
0.45

Po

100


100


ASj

11.7


10.8


Distress

None


None


Purpose
Verify k< 1 x 10 cm/s


Verify k< 1 x 10 cm/s



Remarks






o
CO
ro

-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Database.
Site
No.
1


2

3



4
5


6


7


8


Thin-Wall Sampling Tube
k (cm/s)
4
3.2x10-8
0.32
2
3.6x10-9
109
8.0x10-9

4
5.0x10-9
0.34
3
8.7x10-9
0.21
4
2.4x10-8
0.46
8
8.4x10-8
0.35
5
9.0x10-9
0.58
Method, a', i
D5084
69
10
-
-
Flexible-Wall
165
20
9100
-
-
D5084
69
10
Flexible-Wall
-
-
D5084
-
-
D5084
-
-
SDRI
k (cm/s)
2.8x10-7



1.5x10-7
-

1.1 X10-7


9x10-9


2.7x10-7


5.8x10-8


1.2 X10-7


Size (m2)
1.44



1.82
-


2.33

1.49


2.33


2.33


2.33


Lysimeter
k (cm/s)

-


-
9x10-9

-


-


-


-


-


Size (m2)

-


-
15x15

-


-


-


-


-


TSB
k (cm/s)

-


-
-

-


-


-


5
4.3x10-8
0.12
-


30 cm Block
k (cm/s)

2.6x10-7


-
-

-


4x10-8


-


-


-



DF/L

1


1
-

0.3


0.6


0.7


-


-


Vo
(kPa)

60


70
-

-


80


70


-


-


9
CO
CO

-------
   Table C-4. Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
9
10
11
12
13
14
15
16
Thin-Wall Sampling Tube
k (cm/s)
1.0x10-8
8.0x10-9
2.0x10-9
3.0x10-9
8
1.3x10-8
0.18
4
4.8x10-8
0.29
10
4.4x10-9
0.48
7
3.7x10-8
0.48
Method, a', i
-
-
-
-
D5084
69
-
9100
D5084
SDRI
k (cm/s)




1.3x10-8
2.0x10-8
3.3x10-9
3.0x10-8
Size (m2)




2.33
2.33
2.33
2.33
Lysimeter
k (cm/s)
7x10-9
3x10-8
3x10-9
2x10-9




Size (m2)








TSB
k (cm/s)




5
1.4x10-8
0.16

4
1.6x10-8
0.21

30 cm Block
k (cm/s)








DF/L




0.1
0.2
0.7
0.1
Vo
(kPa)








9
CO

-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
17


18


19



20
21


22

23

24


Thin-Wall Sampling Tube
k
7
3x10-9
0.19
5
1.5x10-8
0.12
8
1.9x10-8
0.46
9
3.0x10-9
0.63
2
3.1 X10-7
-
-
2.4x10-8
6
1.5x10-8
0.45
2
9x10-9
-
Method, a', i
9100
34
-
D5084
34
-
-
-
D5084
35
-
D5084
69
10
-
;
-
-
D5084
69
10
SDRI
k (cm/s)



9.8x10-9


-

8x10-7


2.5x10-7


2x10-8

-

1.5x10-8


Size (m2)



2.33


-

2.33


2.33


2.33

-

2.33


Lysimeter
k (cm/s)
6
6x10-9
0.25
-


4.4x10-
8

-


-


-

1.4x10-
8

-


Size (m2)
0.37


-


8x8

-


-


-

-

-


TSB
k (cm/s)
6
5x10-9
0.23
8
9.2x10-9
0.26
-

-


-


-

-

-


30 cm Block
k (cm/s)
-


4
1.4x10-8
0.34
-

-


2
2.2x10-7
-
-

-

1.1 X10-8


DF/L
1.0


-


-

1


1


-

-

1


Vo
(kPa)
-


-


-

30


-


-

-

45


9
CO
en

-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
25


26


27




28
29


30


31


32

Thin-Wall Sampling Tube
k (cm/s)
2
2.3x10-9
-
2
2.9x10-9
-
2
3.0x10-8
-
2
1.9x10-8
-
2
2.2x10-8
-
-
3.0x10-8
-
7
1 .6 x 1 0-8
0.26
2
3.0x10-8
Method, a', i
D5084
69
10
D5084
69
10
D5084
69
10
D5084
69
10
D5084
69
10
Flexible-Wall
-
22
D5084
-
-
Flexible-Wall
_
SDRI
k (cm/s)
8x10-9


2.0x10-7


1.8x10-7


9x10-8


3
1.7x10-8
-
1.1 X10-7


6.0x10-8


3.9x10-8

Size (m2)
2.33


2.33


2.33


2.33


1.85


2.33


2.33


2.33

Lysimeter
k (cm/s)
-


-


-


-


-


-


-


-

Size (m2)
-


-


-


-


-


-


-


-

TSB
k (cm/s)
-


-


-


-


-


-


6
4.7x10-8
0.034
-

30 cm Block
k (cm/s
6x10-9


1.8x10-7


1.5x10-7


1.7x10-7


2
1.7x10-8
-
-


~


-


DF/L
1


1


1


1


1


>0.7


1


1

Vo
(kPa)
35


-


-


-


-


-


32


0

9
CO
en

-------
   Table C-4. Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
33

34

35


36
37

38
39


40


41
Thin-Wall Sampling Tube
k (cm/s)
2
1.3x10-8
6
1.5x10-8
0.62
2
3.0x10-8
6
9.1 X10-9
0.58
-
4.9x10-8
-
6
2.6x10-8
0.14
7
3.5x10-9
0.35
5.5x10-9
Method, a', i
-
;
D5084
69
10
-
-
D5084
21
20
Flexible-Wall
-
-
D5084
14
10
D5084
34
-
-
SDRI
k (cm/s)
3.9x10-8

4x10-7

3.7x10-8

3.0x10-8

1.3x10-8

<3. 6x1 0-8
2.6x10-9

4.3x10-9
2.2x10-8


1.0x10-7
Size (m2)
2.33

2.33

2.33

2.33

2.33

0.16
0.08 m2

1 .76 m2



2.33
Lysimeter
k (cm/s)
-

-

-

-

-


No
Flow

-


-
Size (m2)
-

-

-

-

-


-


-


-
TSB
K (cm/s)
-

-

-

-

-


-


7
1.6x10-8
0.33
-
30 cm Block
k (cm/s)
-

3
3.5x10-7
0.23
-

-

-


-


-


4.1x10-9

DF/L
1

1

>0.7

>0.7

0.5

0.1
0.5


0.8


0.24
Vo
(kPa)
0

-

-

25

~


-


-


-
9
CO

-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner  Materials in Database (continued).
Site
No.
42
43


44


45




46
47
48


49


50


51


Thin-Wall Sampling Tube
k (cm/s)
-
3
2.4x10-9
0.12
3
2.4x10-9
0.13
12
5.8x10-9
0.63
9
1.5x10-8
0.12
-
3
1.1 X10-8
0.21
4
5.1 X10-8
0.67
3
7.4x10-8
0.31
3
4.1 X10-8
0.15
Method, a', i
-
Const. Head
-
-
Const. Head
-
-
-
-
-
-
-
-
-
D5084
69
10
D5084
69
10
D5084
69
10
D5084
69
10
SDRI
k (cm/s)
8x10-8
7x10-8


2x10-7



3.7x10-8


2x10-8

5x10-8
4x10-8


5.0x10-8


2.6x10-7


3.0x10-7


Size (m2)
2.33
2.33


2.33



2.33


2.33

2.33
2.33


2.33


2.33


2.33


Lysimeter
k (cm/s)
-
-


-



-

-


-
-


-


-


-


Size (m2)
-
-


-



-

-


-
-


-


-


-


TSB
k (cm/s)
-
-


-


-


-


-
5
2.1x10-8
0.57
5
3.2x10-7
1.07
5
7.5x10-8
1.20
5
1.1 X10-7
1.08
30 cm Block
k (cm/s)
-
-


-


-


-


-
4.8x10-8


7.7x10-8


3.1x10-6


5.3x10-7



DF/L
-
-


-


0.5


-


0.5
1


1


1


1


Vo
(kPa)
-
-


-


-


-


20
32


35


34


22


9
CO
oo

-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner  Materials in Database (continued).
Site
No.
52


53


54
55


56
57

58
59
60
Thin-Wall Sampling Tube
k (cm/s)
-
-
-
4
1.7x10-8
0.21
-
-
8.1 X10-8
2.8x10-8
-
-
3.4x10-8
2.5x10-8
2.7x10-8
3.4x10-8
Method, a', i
-
-
-
D5084
-
-
-
-
-
:
-
-
-
-
-
-
SDRI
k (cm/s)
2
1.1 X10-7
0.10
2.2x10-8


7x10-8
1.3x10-7

2.4x10-8

5.6x10-8

5.0x10-8
9.4x10-8
1.2x10-7
Size (m2)
7.20


2.33


2.33
2.33

2.33

2.33

2.33
2.33
2.33
Lysimeter
k (cm/s)
-


-


-
-

-

-




Size (m2)
-


-


-
-

-

-




TSB
k (cm/s)
-


5
1.2x10-8
0.35
-
-

-

-




30 cm Block
k (cm/s)
-


-


-
-

-

-





DF/L
-


0.2


1
0.67

0.63

0.71

0.71
0.54
0.63
Vo
(kPa)
2


-


70
-

-

-




9
CO
CD

-------
Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
61

62
63

64
65
66


67


68


69


Thin-Wall Sampling Tube
k (cm/s)
-
4.3x10-8
1 .6 x 1 0-7
1.7x10-7
5.5x10-9
-
-
3
3.7 x10-8
0.31
3
3.0 x10-8
0.30
4
7.8 x10-9
0.14
4
2.1 x10-8
0.33
Method, a', i
-
;
-
-
;
-
-
D5084
-
-
D5084
-
-
D5084
22
34
D5084
22
34
SDRI
k (cm/s)
3.7x10-8

3.1 x10-7
3.9x10-7
2.3x10-7

1.8x10-7

1.2x10-8


8.3 x10-8

2
2.3 x10-8
0.017
2
1.3x10-8
0.002
Size (m2)
2.33

2.33
2.33
2.33

2.33

2.33


2.33


2.33


2.33

Lysimeter
k (cm/s)
-



-

-
-


-


-


-


Size (m2)
-



-

-
-


-


-


-


TSB
k (cm/s)
-



-

-
5
1.1 x10-8
0.26
5
8.5 x10-8
0.21
5
2.6 x10-8
0.11
5
5.6x10-8
0.12
30 cm Block
K (cm/s)
-



4.1 X10-9

-
-


-


-


-



DF/L
0.63

0.75
0.54
0.25

0.80

>0.5


>0.5


>0.7


>0.7

Vo
(kPa)
-



-

-

26


34


60


46


-------
Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Database (continued).
Site
No.
70
71
72
73
74
75
76
77
78
79
80
81
Thin-Wall Sampling Tube
k (cm/s)
2x10-8
2x10-8
2
1 .4 x 1 0-8
0.06

-
-
4
4.7x10-8
1.1
-
-
5
3.3x10-9
0.22
3
1.8x10-9
0.15
2
4.2x10-8
0.27
Method, a', i
Flexible-Wall
Flexible-Wall
52

-
-
52
-
-
-
-
-
SDRI
k (cm/s)
4x10-8
8.3x10-8
2.0x10-8
8x10-8
1 X10-9
5x10-8
3x10-8
2x10-8
2x10-8
2
4.5x10-8
2
4.0x10-8
1.5x10-7
Size (m2)
2.33
2.33
2.33
2.33
2.33
2.33
2.33
2.33
2.33
2.33
2.33
2.33
Lysimeter
k (cm/s)




-
-

-
-



Size (m2)




-
-

-
-



TSB
k (cm/s)




-
-

-
-



30 cm Block
k (cm/s)




-
-

-
-



DF/L


0.5
0.6
0.5
1
1
0.5
0.5

0.4
0.8
Vo
(kPa)




-
-

-
-




-------
   Table C-4.  Hydraulic Conductivity for Natural Clay Liner Materials in Data Base (continued).
Site
No.
82


83


84
85


86
87

88
89
Thin-Wall Sampling Tube
k (cm/s)
4
1.5x10-8
0.29
4
1.7x10-8
0.05
-
-
2
2.2x10-8
-
2
2.6x10-8
2
3.9x10-8
2
3.1 X10-8
Method, a', i
Flexible-Wall
34
-
Flexible-Wall
34
-
-
-
Flexible-wall
14
-
Flexible-wall
14
Flexible-wall
14
Flexible-wall
14
SDRI
k (cm/s)
3x10-8


4.5x10-8


1.3x10-7
2.8x10-8
1.5x10-8


1.4x10-8

2.3x10-8
2.1 X10-8
Size (m2)
2.33


2.33


2.33
2.33
2.25


2.25

2.25
2.25
Lysimeter
k (cm/s)
-


-


-
-
-


-



Size (m2)
-


-


-
-
-


-



TSB
k (cm/s)
-


-


-
-
-


-



30 cm Block
k (cm/s)
-


-


-
-
-


-




DF/L
0.4


0.4


0.8
0.8
-


-



Vo
(kPa)
-


-


-
-
-


-



o

4*.
IV)

-------
   Table C-5.  Material Properties for Soil-Bentonite Liners in Database.
Site
No.
1

2
3

4
5

6

7


8

Location
and Date
Oxford, NJ
1991
Southern
Nebraska
Southern
Nebraska
Southern
Nebraska

Southern
Nebraska
Kettleman
City, CA
1987
Kettleman
City, CA
1987
Borfer, TX
1988

Source of Data
Colder Assoc

D.L. Osadnick
D.L. Osadnick
D.L. Osadnick

D.L. Osadnick

Colder Assoc.

Colder Assoc.


McBride-Ratcliff


LL (%)
-
;
31
51
31
51
31
51
-
31
51
-
-
-
.
-
-
56

PI (%)
-
;
31
36
31
36
31
36
-
31
36
29
-
-
29
-
-
31
Percent
Gravel
-
;
32
13
32
13
32
13
-
32
13
-
-
-
.
-
-
-
Percent
Fines
-
;
32
32
32
32
32
32
-
32
32
81
-
-
81
-
-
55
Percent
Bentonite)
-
3.75
9.0
9.0
9.0
-
-
9.0



.


7.8
wopt
(%)

-
15.0
15.0
15.0


15.0
23.8


23.8


18.6
Yd, max
(kN/m3)

-
17.1
17.1
17.1


17.1
15.4


15.4


16.9
Compactive
Effort


SP
SP
SP


SP
SP


SP


SP
o

-------
   Table C-5.  Material Properties for Soil-Bentonite Liners in Database (continued).
Site
No.
9

10


11




12
Location
and Date
Borfer, TX
1988
San Mateo
County, Ca
1993
Lead, South
Dakota
1994
Mobile, AZ
1990

Source of Data
McBride-Ratcliff

BFI


Colder Assoc.


Colder Assoc.


LL (%)
-
65
-
51
-
-
.
-
-
60
-
PI (%)
-
39
-
36
-
-
.
-
-
38
-
Percent
Gravel
-
;
-
2
-
-
.
-
-
-
-
Percent
Fines
-
63
-
21
-
-
.
-
-
39
-
Percent
Bentonite)

10.5

10.0

-
14.7
-

4.0

wopt
(%)

20.1

9.0


17.7


13.5

Yd, max
(kN/m3)

16.5

19.9


16.8


18.5

Compactive
Effort

SP

MP


SP


SP

o

-------
   Table C-6. Construction Information for Soil Bentonite Liners in Database.
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
Compaction Criteria

w > OWC, <+4
RC > 95% SP
w > OWC, <+4
RC > 95% SP
w > OWC, <+4
RC > 95% SP
w > OWC, <+4
RC > 95% SP
w>OWC
RC > 90% MP
w > OWC +3
RC > 90% SP
w>OWC
RC > 95% SP
w > OWC +2
RC > 92% SP
w > OWC+2, +5
RC > 905 MP
w > OWC, <+3
RC > 98% SP
w > OWC +2
RC > 95% SP
Compactor
Ingersol
Rand S100
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT 81 5
CAT 825
-
CAT 81 5 (1-4
lift), CAT
CP433B (5th
lift), Sakai SV
(6th lift)
Compactor
Mass (kg)
-
20,000
20,000
20,000
20,000
20,000
20,000
12,600
20,000
32,400
-

Passes per
Lift
10
6
6
4
4
2
2
6
6
4
-
4
Lift Thickness
(mm)
150
150
150
150
150
150
150
150
150
150
150-230
150
Number of
Lifts
4
6
6
6
6
7
7
6
6
6
3
6
Pad Size
(m x m or m2)
9x9
31 x11
31 x11
31 x11
31 x11
43x15
43x15
13x28
13x28
15x15
11 x11
36x18
o

-k
en

-------
   Table C-7. Quality Control/Quality Assurance Data for Soil-Bentonite Liners in Database.
Site
No.
1


2

3



4
5

6

7

8

9


w (%)
28
12.3
1.4
2
14.7
2
16.0
2
143
-
2
15.2
-
28.4
-
28.4
38
20.2
38
21.4

yd (kN/m3)
28
16.0
0.3
2
17.2
2
16.7
2
16.8
-
2
16.5
-
14.8
-
14.8
38
15.9
38
15.9

Po

-


~

~

-


~

54

54

~

"

ASj

-


~

~

-


~

~

~

~

"

Distress
None


None

None


None

None

None

None

None

None


Purpose
Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'8 cm/s

Verify k< 1 x 10'7 cm/s


Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s

Verify k< 1 x 10'7 cm/s


Remarks




















o

-------
Table C-7. Quality Control/Qualit
Site
No.
10


11

12


w (%)
34
12.5
0.77
8
18.5
32
15.5

yd (kN/m3)
34
19.1
0.22
8
16.8
32
17.9

Po

100


75

"

ASj

-


~

"
\j Assurance Data for Soil-Bentonite Liners in Database (continued)

Distress
None


None

None


Purpose
Verify k< 1 x 10~7 cm/s


Verify k< 1 x 10'7 cm/s

Verify k< 1 x 10~7 cm/s


Remarks







o

-------
   Table C-8.  Hydraulic Conductivity for Soil-Bentonite Liners in Database.

Site
1

2
3

4
5

6
7

8

Thin-Wall Sampling Tube
k (cm/s)
4
5.5x10-8
2
3.0x10-8
2
1.9x10-8
2
6.0x10-8
-
2
7.5x10-8
7
6.9x10-9
7
6.9x10-9
0.35
-
_
Method, a', i
D5084
-
D5084
34
D5084
34
D5084
34
-
D5084
34
-
-
-
-
_
SDRI
k (cm/s)

?
3.0x10-8
1.0x10-8
3.0x10-8


2.0x10-8
1.6x10-8
6.2x10-8


2.2x10-8
Size (m2)

-
2.33
2.33
2.33


2.33
2.33
2.33


2.33
Lysimeter
k (cm/s)













Size (m2)













TSB
k (cm/s)













30 cm Block
k (cm/s)














DF/L













Vo
(kPa)













o

4*.
oo

-------
   Table C-8.  Hydraulic Conductivity for Soil-Bentonite Liners in Database (continued).

Site
9

10
11

12
Thin-Wall Sampling Tube
k (cm/s)
-
-
10
2.6x10-8
-
6
3.2x10-8
-
Method, a', i
-
-
-
-
;
-
SDRI
k (cm/s)

1.0x10-7
3.0x10-8
2.0x10-9
2.0x10-8

Size (m2)

2.33
2.33
2.33
2.33

Lysimeter
k (cm/s)






Size (m2)






TSB
k (cm/s)






30 cm Block
k (cm/s)







DF/L






Vo
(kPa)






o

4*.
CD

-------
                 Appendix D
Cincinnati  Geosynthetic Clay Liner
                   Test Site
                        by

                David E. Daniel, Ph.D., P.E.
                   University of Illinois
                   Urbana, IL 61081
                    performed under

             EPA Cooperative Agreement Number
                    CR-821448-01-0
                    Project Officer

                  Mr. David A. Carson
         United States Environmental Protection Agency
             Office of Research and Development
         National Risk Management Research Laboratory
                  Cincinnati, OH 45268

-------
                                 Appendix D
               Cincinnati Geosynthetic Clay Liner Test Site

D-1  Introduction
This appendix contains additional information that augments Chapter 3 in the main body
of the report regarding the research program related to field test plots constructed at a
site in Cincinnati, Ohio.  The test plots were constructed to evaluate the internal shear
resistance of geosynthetic clay  liners  (GCLs) that were constructed on 2H:1V and
3H:1V slopes in prototype landfill cover systems.

D-2  Test Plots
The  main objective in  constructing the field test plots was to  investigate the internal
(mid-plane) shear  strength of GCLs in carefully  controlled, field-scale tests.  Other
objectives were to verify that GCLs in landfill cover systems will remain stable on 3H:1V
slopes with a factor of safety of at least 1.5, to monitor the displacement and creep of
GCLs in the  field for as long as possible,  to  develop  information  on erosion  control
materials, and to better understand the field performance of GCLs as a component in
liner and cover systems.

Fourteen test plots have been constructed at the ELDA Landfill  in Cincinnati, Ohio.
Nine of the plots were constructed on 2H:1V slopes and five were constructed on 3H:1V
slopes.   Each  plot  is  about 9 m wide by 20  or  29 m long and is  covered  by
approximately 0.9 m  of cover soil.  Instrumentation was placed in each test plot (with a
few  exceptions)  in  order  to  monitor the  moisture  content of  the  subsoil  and
displacements of the  GCL.   An  additional  plot  consisting only of cover  soil  was
constructed on the 2H:1V slope.  This plot did not contain geosynthetic materials and
was  used as  a control plot to study the effect of erosion on the cover soil on a plot that
did not contain any synthetic erosion control material.

Slope angles of 2H:1V and 3H:1V were selected to test the shear strength limits of the
GCLs.  The rationale for selecting these slope inclinations was as follows. Many landfill
final  covers  have slopes of approximately 3H:1V.  If GCLs are to be widely used in
landfill covers, they will have to be stable at a slope angle of 3H:1V.  Thus, the 3H:1V
slope was selected to  be representative of  a typical landfill cover.  However, it is not
sufficient to demonstrate that GCLs are stable on 3H:1V slopes — it must be shown that
they are stable with an  adequate  factor  of  safety.    Many  regulators  and  design
engineers require that permanent  slopes have a  minimum factor of safety  for static
loading of 1.5.
For an infinite slope  in a cohesionless  material, with no seepage, the factor  of safety
(FS) is:

      FS = tan(4)) / tan(p)                                                   (D.1)

                                      D-1

-------
where (j> is the friction angle and p is the slope angle.  If a GCL remains stable on the
2H:1V slope,  the friction angle of the  GCL (assuming  zero cohesion) must be at least
26.6°, and for this friction angle, the factor of safety on a 3H:1V slope must be at least
1.5.  Thus, the logic was to try to demonstrate a minimum factor of safety of 1.5 on
3H:1V slopes, and in order to do this,  it was necessary to test the GCLs on 2H:1V
slopes.    It was recognized that constructing a 2H:1V slope was  pushing the test to
(and possibly beyond)  the limits of stability, not necessarily of the mid-plane of the
GCLs but certainly at various interfaces within the system.

D-2.1 Expectations at the Beginning of the Project
During the conception and design of the  field test plots, there were several expectations
concerning the performance of the GCLs. First, it was assumed that if the GCLs were
placed with the bentonite in contact with the subgrade soils that the bentonite would
hydrate by absorbing water from the adjacent soils. However, it was also assumed  that
if a  geomembrane  (GM) separated the bentonite component  of  the  GCL from  the
underlying subsoil,  and a GM was placed over the bentonite to encase the bentonite
between two  GMs, that the bentonite  would be isolated from adjacent soils (except at
edges) and would not hydrate.

A key expectation was that none of the GCLs would fail internally on any of the field  test
plots. This expectation was based on the results of mid-plane laboratory shearing  test
on fully-hydrated GCLs.    Interface  shear slides were viewed as possible,  but  the
greatest  concern  was  with   the  GCL/subsoil  interface.    It was   predicted   that
displacements of the GCLs would  be downslope with the largest displacements on the
2H:1V slopes.  Creep of the GCLs was considered possible.   Differential (shear)
displacements were expected to be nominal.

D-2.2 Layout of the Test Plots
Fourteen test plots containing  a GCL as a component  were  constructed.  The layout of
the plots is shown in Figure D-1. Each plot was assigned a letter. Five plots (plots A-E)
were constructed on a 3H:1V slope, and nine plots (plots F to L, N, and P) were built on
a 2H:1V slope.  An additional  plot, plot M, which  consisted of only cover soil and no
geosynthetics, was  an  erosion control  plot that  was installed  on a 2H:1V slope to
document the degree of erosion that would occur if no synthetic erosion control material
was  placed over the cover soil. In all other plots, a synthetic erosion control material
covered the surface of the test plot.   Plots on the 2H:1V slope  were about 20 m long
and 9 m wide; plots on the 3H:1V slope were about 29 m long and 9  m wide.

D-2.3 Plot Compositions
Four different types of GCLs were placed at the site: Gundseal, Bentomat ST, Claymax
500SP, and Bentofix. Two styles of Bentofix were employed:  Bentofix NW contained

                                      D-2

-------
                                   Geomembrane Placed over GCL
                    Crest of Slope

     3H:1V Plots:

                     Toe of Slope —»•


A
t
B

B

c

D

E
entomat f Bentofix f
ST I NS I
                                  Gundseal      Claymax      Gundseal
                                  (Bentonite       500SP       (Bentonite
                                     Up)                       Down)
                                                             Note:

                                                             All Test plots Were Nominally
                                                             9 m Wide and either 20 m Long
                                                             (21-1:1 V Slopes) or 29m Long
                                                             (31-1:1 V Slopes)
                               Geomembrane Placed
                                    over GCL
                                          No Geomembrane
                                                 Geomembrane
                                                  Placed over
                                                     GCL
D
CO
Crest of Slope
     2H:1V Plots:
             Toe of Slope —»•
G,
P

H

1

J

K

L

M

N
                          Gundseal  A   Claymax   A   Bentomat
(Bentonite
   Up)
                                        500SP
                                                       ST
                                              Bentofix
                                                NW
                                                    Bentofix
                                                      NW
Bentofix
  NS
                                                           No GCL
                                                           (Erosion
                             Bentomat ST (Plot G);
                             Gundseal (Plot P) with
                              Bentonite Up
                                              Claymax
                                               500SP
            Figure D-1.  Layout of field test plots.

-------
nonwoven geotextiles (GTs) on both surfaces.  Bentofix NS contained a woven GT on
the side that faced downward and a nonwoven GT on the side that faced upward.

Two  general  designs  were employed.   The  principal design involved  a subgrade
overlain by a  GCL, textured GM, geotextile/geonet/geotextile drainage composite, and
0.9 m of cover soil, as shown in Figure D-2.  This cross section is typical of many final
cover systems for landfills  being designed today.  The  GTs were heat-bonded to the
geonet (GN).  A nonwoven, needlepunched GT was used between the textured GM and
GN in an effort to develop  a high coefficient of friction between the  GM and drainage
layer.
                         Erosion
                         Control
                        Cover Soil
                        GT/GN/GT
                            GM
                           GCL

                         Subsoil •
Figure D-2.    Typical test  plot cross section employing a composite textured
              HOPE GM/GCL liner system.
The second design involves a GCL overlain by 0.3 m of drainage soil, a GT, and 0.6 m
of cover soil, as shown in Figure D-3.  This design is  also typical of current GCL designs
for final cover systems in which a GM  is not used.

The GM-supported GCL was employed  in two configurations, i.e., with the bentonite
encased between two GMs as shown in Figure  D-4A, or with the  bentonite in contact
with the subgrade, as shown in  Figure D-4B.  In the former case, the  bentonite was
designed to stay  dry.  In  the latter  case, it was expected that the bentonite would
hydrate by absorbing moisture from the subgrade.
                                     D-4

-------
                          Erosion
                          Control
                        Cover Soil
                        Geotextile
                     Drainage Sand
                             GCL

                          Subsoil
Figure D-3.    Alternative test plot cross section employing a GCL with no GM.
Geotextile-encased, needlepunched GCLs consisted of materials that either had woven
and nonwoven GTs on the surfaces, or two nonwoven GTs.  For the GCLs containing a
woven GT on one  surface, the woven  GT faced upward in some cases (Figure D-5A)
and downward in other cases (Figure D-5B)

Plot M is an erosion control section and consisted only of 0.9  m of cover soil.  There
were no geosynthetic materials or instrumentation at the erosion control  section.  The
plot was constructed to document the erosion that would occur without any geosynthetic
erosion control material on the surface.

General cross sections are shown in  Figure D-6 for plots constructed on the 2H:1V
slope and in Figure D-7 for plots constructed on the 3H:1V slope. A cross section in the
perpendicular direction is  shown in Figure D-8.  Each plot width was equal to two GCL
panels minus a 150-mm-wide overlap.  The spaces  between plots on the 2H:1V slope
ranged between 0 and 1.5 m,  and were typically 1.5  m on the 3H:1V slope. There were
graded drainage swales  only  on  the 3H:1V slopes.  Table D-1 lists the  slope angles,
plot, type of GCL, and a description of the plot cross-section from top to bottom. Table
D-2 lists the composition, dimensions, etc., of each plot.

D-2.4 Anchor Trenches
Anchor trenches were constructed at the crest of each test plot.   On the 3H:1V and
2H:1V  slopes all of the geosynthetic materials (GCL, GM, and GN, if present) were
brought into the anchor trench.  A GM cap strip was placed over the GCL in the anchor
trench with  the purpose of preventing moisture from  entering the GCL from the crest of
the plot. A typical anchor trench detail is shown in Figure D-9.

                                      D-5

-------
Soil in the anchor trench was nominally compacted. The anchor trench was used only
for  the  purpose of holding the geosynthetics  in position  during construction.  As
discussed later, the geosynthetics above the mid-plane of the GCLs were cut next to the
anchor trenches so that the shear force from the cover would be transmitted to the
internal  structure of the GCL (and  not simply carried by tension in the geosynthetics
overlying the mid-plane and anchored in the anchor trench).

D-2.5 Toe Detail
At the toe of the slope the GM and  GN were extended beyond the GCL in the plots on
the  3H:1V and 2H:1V slopes.  The  extension is shown in Figures D-6 and  D-7 for test
plots employing a GN for the drainage  layer.  Both the GM and the GN  were  extended
(daylighted) approximately 1.5 m past the end of the cover soil. For test plots in which a
sand drainage layer was used, a GN was extended beyond the sand drainage material
as shown in Figure D-10. The toe was designed to provide no buttressing effect for the
cover soil.

D-2.6 Instrumentation
The objectives of the instrumentation for the field test plots were to monitor the wetting
of the subsoil and the bentonite in the GCLs, and to monitor displacements of the GCLs.
Moisture sensors were  installed to verify that the  bentonite was hydrated, or in the case
of plots A, F,  and P, to  verify that the bentonite was dry. Extensometers were installed
to document  the internal shear and creep of the GCLs in each plot.  As there was  a
limited budget,  the  instrumentation was selected  based on simplicity, low cost, and
redundancy.

      D-2.6.1 Moisture Sensors.  Moisture sensors were installed in each test plot in
order to assess the  moisture conditions impacting the bentonite within the GCLs.  Two
types of sensors were used in  the project: a gypsum block sensor  and a fiberglass
mesh sensor (Figure D-16). The  gypsum block sensors were placed in  the subsoil
beneath the GCLs; the fiberglass sensors were placed within the  bentonite of the GCLs.
Both sensors operate on a resistance basis.  The fiberglass sensors contain a porous
fiberglass mesh embedded in two  wire screens .  The  resistance to flow of electric
current between the two screens is  dependent on the moisture present in the fiberglass
mesh. The resistance  is measured and converted to moisture content  by comparison
with a calibration chart.  The calibration is a function of soil type and the constituents of
the  soil moisture.   The gypsum block sensors have  two concentric  spirals of wire
between which resistance of gypsum is determined.  The electrical resistance of the
gypsum  is a function  of  the moisture content of the gypsum.   The resistance is
measured using a digital  meter manufactured  specifically  to measure resistance for
these sensors.

                                      D-6

-------
 (A) Plots with Bentonite Component Facing Upward
                   (B) Plots with Bentonite Component Facing Downward
      Cover Soil
   Geocomposite
   Drainage Layer

   Textured HOPE
   Geomembrane

GCL (Bentonite Up) {
        Subsoil
-Bentonite
                                    • Textured HOPE
                                    Geomembrane
                          Cover Soil
                       Geocomposite
                       Drainage Layer
                 GCL (Bentonite Down)
                                                                 Subsoil
Textured HOPE
Geomembrane
 Bentonite
   Figure D-4. Placement of Gundseal with bentonite facing upward or downward.

-------
(A)  Woven Geotextile Interfacing with Geomembrane
                    (B) Nonwoven Geotextile Interfacing with Geomembrane
       Cover Soil
    Geocomposite
    Drainage Layer

    Textured HOPE
    Geomembrane
GCL (BentonitSOp.) {
         Subsoil
Woven
Geotextile
  Bentonite
                                     ^ Nonwoven
                                       Geotextile
                            Cover Soil
Geocomposite
Drainage Layer


        GCL
                               Subsoil
Nonwoven
Geotextile
                                                                                                     • Bentonite
                                                              Woven
                                                              Geotextile
   Figure D-5. Orientation of GCL with either woven or nonwoven GT facing upward.

-------
Drainage Layer
(GT/GN/GT or Granular)
     Figure D-6. General cross section of plots on a 2H:1V slope.
      Geocomposite
         Drainage
         Material
     Figure D-7. General cross section of plots on a 31-1:1 V slope.
                                         D-9

-------
               0.9m
                               GCL
                                   GM
                                       Drainage Layer:
                                         GT/GN/GT
                                         or Granular
Erosion
Control
 Mat
                          Total VUdth of Test Plot: 10 to 13 m
Figure D-8. Cross section along width of test plots.
              Geomembrane
                Cap Strip
           Geocomposite-
           Drainage Layer
                                    All Geosynthetics above the
                                    Mid-Plane of the GCL Were Cut,
                                    Including the Upper Geotextile or
                                    Geomembrane Component of the
                                    GCL (If Present)
                      Anchor
                      Trench
                      Backfill
                                      Geomembrane
Figure D-9. Typical anchor trench detail.
                                       D-10

-------
Table D-1. Components of the GCL Field Test Plots.
Plot
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
GCL
Gundseal
Bentomat ST
Claymax 500SP
Bentofix NS
Gundseal
Gundseal
Bentomat ST
Claymax 500SP
Bentofix NW
Bentomat ST
Claymax 500SP
Bentofix NW
Erosion
Control
Bentofix NS
Gundseal
Target
Slope
(deg.)
18.4
18.4
18.4
18.4
18.4
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
26.6
Actual
Slope
(deg.)
16.9
17.8
17.6
17.5
17.7
23.6
23.5
24.7
24.8
24.8
25.5
24.9
23.5
22.9
24.7
Cross-section
(from top to bottom)
Soil/GN/GM/GCL (Bent, up)
Soil/GN/GM/GCL(Wup)
Soil/GN/GM/GCL (W-W)
Soil/GN/GM/GCL (NW up)
Soil/GN/GCL (Bent, down)
Soil/GN/GM/GCL (Bent, up)
Soil/GN/GM/GCL (W up)
Soil/GN/GM/GCL (W-W)
Soil/GN/GM/GCL (NW-NW)
Soil/GT/Sand/GCL (W up)
Soil/GT/Sand/GCL (W-W)
Soil/GT/Sand/GCL (NW-NW)
Soil
Soil/GN/GM/GCL (NW up)
Soil/GN/GM/GCL (Bent, up)
where:
      Soil = cover soil
      GN = geonet
      GM = textured GM
      GT = geotextile
      GCL = geosynthetic clay liner
      Bent, up = bentonite side of Gundseal facing upward (GM against subgrade)
      Bent, down =  bentonite side of Gundseal against subgrade
      W up = woven GT of GCL up, nonwoven side of GCL against subgrade
      NW up = nonwoven GT of GCL up, woven side of GCL against subgrade
      NW-NW = both sides of GCL nonwoven
      Bentofix I is Bentofix NW, with a nonwoven GT on both sides
      Bentofix II is Bentofix NS, with a woven GT facing upward.
                                   D-11

-------
Table D-2. Summary of Test Plots.
Plot
A
B
C
D
E
F
G
H
1
J
K
L
M
N
P
GCL
Type
Gundseal
Bentomat
Claymax
Bentofix NS
Gundseal
Gundseal
Bentomat
Claymax
Bentofix NW
Bentomat
Claymax
Bentofix NW
Erosion
Control
Bentofix NS
Gundseal
GM
(Y/N)
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
Y
Drain
Type
GN
GN
GN
GN
GN
GN
GN
GN
GN
Sand
Sand
Sand
Sand
Sand
GN
Slope
3H:1V
3H:1V
3H:1V
3H:1V
3H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
2H:1V
Slope
Length
(m)
28.9
28.9
28.9
28.9
28.9
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
20.5
Crest
Elev.
(m)
179.2
179.2
179.2
179.2
179.2
157.9
157.9
157.9
157.9
157.9
157.9
157.9
157.9
157.9
157.9
Toe
Elev.
(m)
170.0
170.0
170.0
170.0
170.0
148.7
148.7
148.7
148.7
148.7
148.7
148.7
148.7
148.7
148.3
Test
Plot
Width
(m)
10.5
9.0
8.1
9.1
10.5
10.5
9.0
8.1
9.1
9.0
8.1
9.1
7.6
9.1
10.5
Notes:
     1.
     2.
     3.
     4.
Bentofix NW contained a nonwoven GT on both sides.
Bentofix NS was installed with the nonwoven GT facing upward.
Bentomat ST was installed with the woven GT facing upward.
Gundseal was installed with the bentonite facing upward in plots A, F, and P,
and with the  bentonite facing downward at plot E.
                                     D-12

-------
        GCL
Figure D-10.    Detail of drainage at toe for sections with GN drainage layer (not
              to scale).
           Gypsum Block
Fiberglass Moisture Sensor
                       40mm     40mm
                                         I!
              25 mm
   25mm
Figure D-11. Schematic diagrams of moisture sensors.
                                 D-13

-------
The  sensors were placed on the centerline of one of the  two GCL panels at  three
locations - top  , middle,  and bottom - of each plot as shown  in Figure D-12.   The
sensors were installed 5.2 m, 10.7 m, and 16.8 m from the crest on the 2H:1V slope and
6.1 m, 15.2 m, and 24.4 m from the crest on the 3H:1V slope. At each location two, and
in some cases three, moisture sensors were placed in the subsoil, at the subsoil-GCL
interface, and in a few instances, above the GCL.  The purpose of the sensors was to
monitor the moisture content of the  bentonite and soil adjacent to the  bentonite.
Because most plots contained a GM above the GCL, placing sensors in the cover soil
would  not provide information  on  moisture conditions within  or  near the GCLs.
Therefore, moisture sensors were generally placed  adjacent to or beneath the GCLs.  A
cross section of the typical moisture sensor installation  in all  plots except for plots A, F,
and  P, is  shown in Figure D-13.   Figure D-14 shows how the moisture sensors were
installed in plots A and F.

The  moisture sensors in  Plot P  were  installed differently than the other plots.  Only
fiberglass moisture sensors were installed in  Plot  P.   Sixteen moisture sensors were
placed in a 4 x 4 grid on the upper side of the bentonite of the GCL but underneath the
overlying GM.

The  gypsum blocks  and  digital  meter were obtained from Soil Moisture  Equipment
Corporation  of  Santa  Barbara,  CA.   The  fiberglass  sensors were  obtained  from
Techsas, Inc. of Houston,  TX.

1B
•
2m
3B

4m
•
5m
CREST
I •
I
I •
I
MIDDLE
I
I u
I
TOE •
                                           $ Cluster of
                                             Moisture Sensors
                                           | Extensometer
                     Left GCL Panel   Rig ht GCL Panel
Figure D-12. Locations of moisture sensors and extensiometers.

                                      D-14

-------
                                   Cover Soil
           GT/GN/GT -
                GM -

                GCL-
                Gypsum
                Block
                                            Subsoil
-Fiberglass Sensor

-Fiberglass Sensor
Figure D-13. Location of moisture sensors in all plots except A and F.
        GT/GN/GT
              GM

             GCL-
               Gypsum_
               Block
                                                Subsoil
  Fiberglass Sensor
Figure D-14. Location of moisture sensors in plots A, F, and P.
As mentioned above, the electrical resistance of a moisture sensor is measured and
converted to  moisture content by comparison with a calibration chart.  The moisture
sensor readout device used on this project reads from 0 to 100, with 0 corresponding to
no soil moisture and 100 corresponding to a very wet soil.  However, the calibration is a
                                      D-15

-------
function of soil type. There are generally four different soil types at the site.  Three soils
are distributed generally as shown in Figure D-15 for the 2H:1V test plots.  Soil A is a
gray fat clay, soil B is a clayey silt, and soil C is a silty clay (field classifications).  The
subsoil on the 3H:1V slope is primarily a clayey silt (soil D).

Calibration  tests were  performed for both the  gypsum block and fiberglass moisture
sensors for soils A, B,  C, and D.  A 1000 ml beaker was filled with soil, and a circular
piece of Gundseal was placed above the soil with the bentonite portion of the GCL in
contact with the soil.  A small layer of sand was placed over the GCL and a pressure of
18 kPa was applied to  the specimen. A gypsum  block was inserted within the subsoil,
and  a fiberglass moisture sensor was  placed at the interface of the GCL and the
subsoil.  The subsoil was incrementally  wetted, and after the moisture gauge reading
had  equilibrated,  the resistance  reading was recorded and a sample of the soil was
obtained for measurement of water content.  A typical calibration curve for the gypsum
block in the subsoil and the fiberglass moisture gauge at the soil/GCL interface is shown
for Soil A in Figure D-16.
                  Soil A: Gray Fat Clay
                  Soil B: Clayey Silt
                  Soil C: Silty Clay
           Crest
            Toe
B





B









B





B









B





B
x








B




X
X
/C









B

X
X
X

C









B"
X




C
X
'







C




'
."A









C

>
X
X


A









px
s




A

            PLOT:  FGHIJKLMN
Figure D-15. Soil types at 2H:1V test plots.

The calibration of the fiberglass moisture sensor with bentonite was performed as
follows.  A fiberglass sensor was sandwiched between two prewetted pieces of
Gundseal so that the  sensor was surrounded by bentonite. Sand was placed below and
above the GCLs, and a pressure of 18 kPa was applied.  After the moisture gauge
reading  had stabilized, the moisture gauge reading was recorded.
                                      D-16

-------
          O)
          CO
          (U
          b
          (U
          CO
          (U
100

 80-

 60+

 40

 20
OA
  •
                               '
  0
   0
—i	1	h-
 10      15      20

   Water Content (%
                                                          25
          30
Figure D-16.  Calibration of gypsum block moisture sensors (typical calibration).
The  calibration  curve  for the fiberglass  moisture sensor with bentonite is shown  in
Figure D-17. The scatter is due to the use of 15 different sensors in the development of
the calibration curve  (each  moisture sensor  should  ideally  have  its own individual
calibration curve).   This  calibration  curve  can  be used to qualitatively  distinguish
whether the  bentonite  is relatively dry or saturated.  Beyond that, however, statistical
scatter limits resolution. For example, a moisture gauge reading of 20 indicates that the
water content of the bentonite could range  between 40 and  150%, and for a  gauge
reading of 80 the water content of the bentonite could range between 190  and  290%.
However, a gauge reading of close to 0 clearly  indicates that the bentonite is dry, and a
reading close to 100 clearly indicates that  it is wet.

      D-2.6.2  Displacement Gauges.  Displacement gauges, or  extensometers, were
installed in each plot to measure displacements and to  assess shear strains  in the GCL
at multiple locations. Twenty displacement gauges were installed  in each plot (10 pairs
on each panel).  Five gauges in each panel were attached to the upper side of the GCL
Figure D-18.  With gauges on the upper  and lower side of the GCL, the difference in
total  displacement between the  upper  and  lower gauges provides  a measure  of
shearing displacement. Figure D-19 shows the  attachment of the hooks to the upper
                                     D-17

-------
              120
100    150    ZOO
  Water Content (%)
250
                                                            300
     Figure D-17. Calibration of fiberglass moisture sensors with bentonite.
and lower GTs of the GCLs.  Each extensometer consisted of a braided steel wire (for
flexibility) running from its point of attachment to above the crest of the slope.  The wire
was contained within a 6-mm OD (outside diameter) plastic tubing, and was connected
to a fishhook at the end of the wire (Figure D-19). The fishhook was attached by epoxy
to the surface of  the GT component of the GCL.   Gauges on the upper and  lower
surfaces  were used to measure  differential displacement, as  shown in Figure  D-20.
Each wire extended from the  fishhook to a monitoring station, or displacement table, at
the crest of the slope. A displacement table is shown  in Figure D-21.

D-2.7 Construction

Construction  of the plots  began  on November 15, 1994, and was  completed on
November 23, 1994. The construction sequence was as follows:
                                     D-18

-------
   1.   Subgrade preparation.
   2.   Installation of moisture sensors in  the  subgrade and  at  the  surface of the
       subgrade.
   3.   Placement of GCL.
   4.   Installation of the extensometers and displacement cables.
   5.   Installation of moisture gauges within the GCL (plots A, F, P).
   6.   Placement of GM (not applicable to plots J, K, L, and M).
   7.   Placement of GN composite or granular drainage layer (plots J, K, L, M).
   8.   Placement of GT (plots J, K, L, M only).
   9.   Placement of cover soil.
   10.  Construction of displacement tables.
      Crest
                                                          Toe
Figure D-18.  Locations of displacement sensors.
                                     D-19

-------

                               -mm OD Plastic Tubing
                                      3-mm Diameter Steel Cable
Figure D-19. Attachment of displacement monitoring hook to GCL.
                                 D-20

-------
           Geomembrane

         Bentonite
       Bentonite
        Extensometer
           Cable
                               O. 0)
                               ° E
      Upper Gauge

2 L = Differential Deformation
     Lower Gauge
                                                 Time
Figure D-20.  Location of displacement gauges to measure differential movement.
                                    D-21

-------
                                                   Displacement
                                                     Indicator
Figure D-21.  Displacement table at crest of slope.
D-2.8 Cutting of the Geosynthetics

With other geosynthetic materials besides the GCL leading into the anchor trench, part
of the down-slope component of force created by the cover soil is carried by tension in
these geosynthetic materials.  To concentrate all of the shear stress within the mid-
plane of the GCL, the geosynthetic materials above the mid-plane  of the GCL were
severed.  The geosynthetics above the mid-plane of the GCLs in plots A through D
(3H:1V slope) were cut on April 13, 1995, and the geosynthetics above the mid-plane of
the GCLs on the 2H:1V slopes and plot E (3H:1V slope) were cut on May 2, 1995.

In plots with GT-encased GCLs,  the  GN composite, GM, and the upper GT of the GCL
were cut at the crest of the slope down to the mid-plane of the GCL as shown in Figure
D-22. The geosynthetic materials in plots constructed with a granular drainage layer
were cut  down to the mid-plane of the GCL as shown  in Figure D-23.  The granular
drainage  material did not extend into the anchor trench, so the GT was cut as well as
the upper GT in the GCL.
                                     D-22

-------
                            Upper Geotextile
                               inGCLCut
         GT/GN/GT
Figure D-22.
Cross-section at crest of slope showing cutting of geosynthetics
down to mid-plane of GCL on test plots with a GM.
                        Upper Geotextile
                          in GCL Cut
       GCL-
                                                            Granular
                                                            Drainage
Figure D-23.
Cross-section at crest of slope showing cutting of geosynthetics
down to mid-plane of GCL on test plots without a GM.
                                   D-23

-------
The cutting of anchor trench materials in plots with Gundseal is shown in Figures D-24
and D-25.  In the case with the bentonite side of the GCL facing up (Figure D-24), the
GN and GM were cut leaving the entire GCL intact.  In the case with the bentonite side
of the GCL facing downward (Figure D-25), the GN and the GM of the GCL were cut.
                               Only GT/GN/GT
                                 & GM Cut
                                                            Topsoil
        Gundseal GCL
        (Bentonite Up)
Figure D-24. Cutting of slope with Gundseal, bentonite side facing upward.
                                GT/GN/GT & GM
                              Component of GCL
                                    Cut
             GT/GN/GT
          Gundseal GCL
         (Bentonite Down)
Figure D-25. Cutting of slope with Gundseal, bentonite side facing downward.
D-2.9 Supplemental Analyses of Subsoil Characteristics
In the summer of 1997, displacements developed in several test plots that appeared to
be  consistent with  the soil  patterns  depicted  in  Figure D-20.   The various soil
                                    D-24

-------
boundaries shown  in  Figure  D-15 were  determined in the field  at  the  time  of
construction, based on visual observations.  Cracks developed in several test  plots
parallel to the lines shown in Figure D-20.

In an attempt  to refine Figure D-15, additional samples  were obtained  between the
2H:1V test plots and analyzed for liquid and  plastic limits following American Society for
Testin