EFFECT OF FREEZE-THAW ON THE HYDRAULIC
CONDUCTIVITY OF BARRIER MATERIALS: LABORATORY
AND FIELD EVALUATION
by
Jason F. Kraus and Craig H. Benson
Environmental Geotechnics Program
Department of Civil and Environmental Engineering
University of Wisconsin-Madison
Madison, Wisconsin
Cooperative Agreement
CR-821024-01-0
Environmental Geotechnics Report 94-5
Project Officer
Robert Landreth
U.S. Environmental Protection Agency
National Risk Management Research Laboratory
Cincinnati, Ohio 45268
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
CONTACT
Robert Landreth is the EPA contact for this report. He is presently with the newly organized
National Risk Management Research Laboratory's new Land Remediation and Pollution
Control Division in Cincinnati, OH (formerly the Risk Reduction Engineering Laboratory).
The National Risk Management Research Laboratory is headquartered in Cincinnati, OH, and
is now responsible for research conducted by the Land Remediation and Pollution Control
Division in Cincinnati.
-------
DISCLAIMER
, , , . „, , |
The information in the document has been funded wholly or in part by the
United States Environmental Protection Agency under assistance agreement
number CR-821024-01-0. It has been subject to the Agency's peer and
administrative review and has been approved for publication as a U.S. EPA
document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
-------
FOREWORD
i
The U.S. Environmental Protection Agency is charged by Congress With protecting
i
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
i
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. i
_ . . .... j . , .
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks from threats
to human health and the environment. The focus of the Laboratory's research program is
on methods for the prevention and control of pollution to air, land, water and subsurface
resources; protection of water quality in public water systems ; remediation of
contaminated sites and ground water; and prevention and control of indoor air pollution.
The goal of this research effort is to catalyze development and implementation of
innovative, cost-effective environmental technologies; develop scientific arid engineering
information needed by EPA to support regulatory and policy decisions; and provide
technical support and information transfer to ensure effective implementation of
environmental regulations and strategies. i
i • !
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
111
-------
ABSTRACT ;
Tests were performed in the laboratory to assess the impact jof freeze-thaw
on the hydraulic conductivity of two compacted clays and three alternative barrier
materials: a sand-bentonite mixture, three geosynthetic clay liners (GCLs), and
three paper mill sludges. Results of laboratory tests on the clays, sand-bentonite
mixture, and GCLs were compared to data obtained from the COLDICE
(Construction of Liners Deployed in Cold Environments) project coriducted by the
U. S. Army Cold Regions Research and Engineering Laboratory i(CRREL) and
CH2M Hill, Inc. The COLDICE project is a large-scale field study designed to
evaluate the effect of freeze-thaw on the hydraulic conductivity of barrier materials.
Results of laboratory tests performed on the paper mill sludges were compared to
results of hydraulic conductivity tests performed in a small-scale field study
conducted at the University of Wisconsin-Madison. The small-scale field study
consisted of compacting paper mill sludge in large PVC pipes and measuring their
hydraulic conductivity before and after exposure to freeze-thaw. Tests were also
performed on the paper mill sludges to determine how effective stress and
permeation time affect hydraulic conductivity.
Results of this project show that: \
(1) The hydraulic conductivity of compacted clay increases as a result of
exposure to freeze-thaw. Cracks that form due to desiccation (induced by
freezing) and formation of ice lenses are responsible for the increase in
hydraulic conductivity. Furthermore, greater increases Jin hydraulic
conductivity occur in the field relative to those that are observed in freeze-thaw
tests conducted in the laboratory. Larger cracks and a more blocky structure
occur in the field. The exact cause of this difference in structure iis not known.
It possibly can be attributed to differences in soil structure prior to freezing.
i
|
(2) The hydraulic conductivity of the bentonitic barrier materials (sand-bentonite
mixture, geosynthetic clay liners (GCLs)) tested in this study was Insensitive to
freeze-thaw. This was observed in the field (COLDICE data) and in the
laboratory. Although segregated ice was observed in sections of frozen
specimens, no cracks were observed in the sand-bentonite mixture or GCLs
after thawing. Apparently, cracks in the soft bentonite close during thawing
and thus no increase in hydraulic conductivity occurs. This behavior is in
IV
-------
direct contrast to the behavior of compacted clays, which are relatively stiff and
retain the cracks incurred during freezing after thawing has occurred.
i
(3) Hydraulic conductivities less than 1 x 10-9 m/s were obtained when
compacting the three paper mill sludges wet of optimum water content. Two of
the sludges behaved like clay when subjected to freeze-thaw. :Their hydraulic
conductivity increased one to two orders of magnitude. In contrast, the other
sludge was resistant to freeze-thaw if it was permeated after each thaw.
However, if this sludge was frozen and thawed without intermittent
permeation, the hydraulic conductivity increased approximately one order of
magnitude.
, . _ !
(4) The small-scale field tests with the sludge were inconclusive. |When the field
specimens were permeated in the pipes, a reduction in hydraulic conductivity
was observed after one winter of freeze-thaw. However, when 'the specimens
were removed from the pipes as slices and permeated in flexible-wall
permeameters, increases in hydraulic conductivity of approximately one order
of magnitude were observed. This discrepancy in hydraulic conductivity may
have been the result of disturbance incurred when the specimens were sliced
from the pipes. Nevertheless, the effect that freeze-thaw has on paper mill
sludges in the field is not clear. Large-scale field tests are recommended to
address this issue.
-------
VI
-------
TABLE OF CONTENTS
DISCLAIMER
FOREWORD
.......... :- ............. ................ • ........................... • ......................... . ............. .....ill
ABSTRACT ..... [[[ ..... . ............................... ;L jv
LIST OF FIGURES... [[[ . ! vi!
•••••»...*......,...,.,.......,,... ....,..»...,,....,,.,, .All
LISTOFTABLES ............... .
1. INTRODUCTION
2. BACKGROUND
2
2.1 IMPACTS OF FREEZE-THAW ON COMPACTED CLAY ! ?
2.2 BENTONITIC BARRIER MATERIALS ................................. »!!"I"!!!""!!""!!!."".5
2.2.1 Sand-Bentonite Mixtures ..................................... 5
2.2.2 Geosynthetic Clay Liners (GCLs) ................. !""."'." ™."3!!!!!™/Z'"'""6
2.3 PAPER MILL SLUDGE ...................................... ..... ....................... . 8
3, MATERIALS,..: ................................ . .................................... ; ..... _ 13
3.1 CLAYS .................................
•r:
13
3.1.1 Index Properties...., 13
3.1.2 Compaction ."......^ 13
3.2 BENTONITIC BARRIER MATERIALS \ 16
3.2.1 Sand-Bentonite Mixture ; _ -|g
3,2.1.1 Index Properties 15
3.2.1.2 Compaction. ^"."'."'.'"""'"'19
3.2.2 Geosynthetic Clay Liners (GCLs) i 19
3.3 PAPER MILL SLUDGES i... . 23
i
3.3.1 Index Tests ! 23
3.3.2 Compaction ".'."""ll"'.T.". 24
4. METHODS i 27
-------
4.1 CLAYS.
4.1.1 Field Methods.
4.1.1.1 COLDICE Test Pads..
4.1.1.2 In Situ Box Infiltrometers.,
.27
.27
.27
.29
4.1.1.3 Laboratory Assessment of Field-Scale
Hydraulic Conductivity 29
.- . !
4.1.2 Laboratory Methods ; ' 35
4.1.2.1 Hydraulic Conductivity-Water Content
Relationships !. 35
4,1.2.2 Standard Freeze-Thaw Tests I......:..............."!... 36
4.1.2.3 One-Dimensional Freeze-Thaw Tests .......!............."."38
4.2 SAND-BENTONITE MIXTURE ......; 40
i
4.2.1 Field Methods [ _ 40
4.2.1.1 COLDICE Test Pads. ! 40
4.2.1.2 In Situ Box Infiltrometers.. , i..'".'.'.'.'".'.'.'.'.'.'.'.'."40
4.2.1.3 Laboratory Assessment of Field-Scale
Hydraulic Conductivity ; ..40
i """"
4.2.2 Laboratory Methods i 43
4.2.2.1 Hydraulic Conductivity-Water Content '<
Relationship j 43
4.2.2.2 Standard Freeze-Thaw Tests ."......."..."!.!.."!."43
4.3 GEOSYNTHETIC CLAY LINERS (GCLs).... j 44
4.3.1 Field Methods _ 44
4.3.1.1 COLDICE Test Ponds and Pans-Field •
Measurement of Hydraulic Conductivity. 44
4.3.1.2 Laboratory Assessment of Field-Scale
Hydraulic Conductivity 45
4.3.2 Laboratory Methods 45
4.3.2.7 Laboratory Measurement of Hydraulic ;
Conductivity ( 45
4.3.2.2 Standard Freeze-Thaw Tests ........".."i................l."48
4.4 PAPER MILL SLUDGES ... ; 49
viii
-------
4.4.1 Field Methods ; 49
4.4.1.1 Compaction of Pipe Specimens j 49
4.4.1.2 Instrumentation and Burial of Pipe Specimens........'"""...49
4.4.1.3 Permeation of Pipe Specimens .' .."..........52
4.4.1.4 Measurement of Hydraulic Conductivity in "
Flexible-Wall Permeameters 54
4.4.2 Laboratory Methods.. , 54
4.4.2.1 Hydraulic Conductivity-Water Content \
Relationships ;... 54
4.4.2.2 Standard Freeze-Thaw Tests .'.'.'.'.'.'.'.'.'.'".'.'.'.'.'.'.'I'.' 55
4.4.2.3 Effective stress Tests !!!"Z!"!!""!!!!!;!!! 57
4.4.2.4 Long-Term Hydraulic Conductivity reste.....l......................57
5. RESULTS: CLAYS [ 58
5.1 FIELD TESTS ; 58
5.1.1 Freeze-Thaw Monitoring 58
5.1.2 In Situ Box Infiltrometers -".""""""""""; 58
5.1.3 Laboratory Assessment of Field-Scale !
Hydraulic Conductivity '. ] QQ
5.1.3.1 Block Specimens .„! . 50
5.1.3.2 Specimens Collected in Thin-Wall Tubes"'''l"f"''^""^'"'. 64
5.1.3.3 Frozen Core Specimens ...........!!..... 68
5.1.3.4 Structure of Compacted Clay Subjected to l
Freeze-Thaw. ; .......... 68
5.2 LABORATORY TESTS [ ......72
! ""
5.2.1 Hydraulic Conductivity-Water Content Relationships;... 72
5.2.2 Standard Freeze-Thaw Tests 72
5.2.3 One-Dimensional Freeze-Thaw Tests ............."....!!."."!"1"!.1.""76
5.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS ...76
6. RESULTS: SAND-BENTONITE MIXTURE [. 82
6.1 FIELD TESTS ; .J 82
6.1.1 Freeze-Thaw Monitoring 82
6.1.2 In Situ Box Infiltrometers „ ......"!l.....f... 82
6.1.3 Laboratory Assessment of Field-Scale "'
Hydraulic Conductivity 83
ix i
-------
6.2 LABORATORY TESTS ;- 85
6.2,1 Hydraulic Conductivity-Water Content Relationships '• 85
6.2.2 Standard Freeze-Thaw Tests ........" 85
i
6.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS 87
i
7. RESULTS: GEOSYNTHETIC CLAY LINERS (GCLs) [ 92
7.1 FIELD TESTS'. '; Q2
7.1.1 GCL Test Ponds and Pans .„.' 92
7-1.2 Laboratory Assessment of Field-Scale :
Hydraulic Conductivity : 95
7.2 LABORATORY TESTS \J_ 95
7.2.1 Laboratory-Scale Hydraulic Conductivity Tests 95
7.2.2 Structure of Frozen Specimens ' -JOP
7.2,3 Bentonite in Effluent »~!!Z"!!!!"!:!!!!!!!!!".'""!!.'io2
7.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS 104
8. RESULTS: PAPER MILL SLUDGES ! 105
8.1 FIELD TESTS • 105
8.1.1 Compaction of Pipe Specimens •.....105
8.1.2 Freeze-Thaw Monitoring of Pipe Specimens 105
8.1.3 Hydraulic Conductivity of Pipe Specimens ..ZZ!!!""^//.!!".3".1'.106
8.1.3.1 Hydraulic Conductivity Measured in Pipes 106
8.1.3.2 Hydraulic Conductivity Measured in Flexible-Wail
Permeameters. 106
8.2 LABORATORY TESTS -, -, 1
8.2.1 Hydraulic Conductivity-Water Content Relationships.
8.2.2 Standard Freeze-Thaw Tests ...............111
8.2.2.1 Low-K Molding Water Contents 111
8.2.2.2 As-Received Molding Water Contents ....."..........118
8.2.3 Effective Stress Tests
8.2.4 Long-Term Hydraulic Conductivity
8.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS 122
-------
9. SUMMARY AND CONCLUSIONS , [ ..124
9.1 COMPACTED CLAYS 104
9.2 SAND-BENTONITE MIXTURE ' 105
9.3 GEOSYNTHETIC CLAY LINERS (GCLs) ."." ] "lofi
9.4 PAPER MILL SLUDGES ™ZIZZ"!"""Il26
REFERENCES j 128
APPENDIX A i 132
APPENDIX B.... : :. 13g
XI
-------
LIST OF FIGURES
Rgure2.1. Hydraulic conductivity vs. number of freeze-thaw cycles for three
Wisconsin clays (after Othman and Benson 1993).
i
F:igure 2.2. Hydraulic conductivity vs. depth for tests performed • on specimens
sliced from "field" specimen (after Benson and Othman !1993).
Figure 2.3. Hydraulic conductivity vs. number of freeze-thaw cycles for sand-
bentonite mixtures (after Wong and Haug 1991). |
Figure 2.4. Hydraulic conductivity vs. number of freeze-thaw cycles for
Bentomat® for test set 1 (a) and test set 2 (b) (after Robert L Nelson &
Associates 1991). i
Figure 2.5. Hydraulic conductivity vs. number of freeze-thaw cycles at various
effective stresses (after Zimmie et al. 1994).
Figure 3.1. Plasticity chart showing locations of Parkview and Valley Trail clays.
Figure 3.2. Results of particle size analyses for Parkview (a) and Valley Trail (b)
clays. ; v '
Figure 3.3. Compaction curves for Parkview clay. \
Figure 3.4. Compaction curves for Valley Trail clay. i
Figure 3.5. Results of particle size analysis for sand component of sand-bentoriite
mixture. \
Figure 3.6. Compaction curves for sand-bentonite.
Figure 3.7. Schematic diagrams of Bentofix® (a), Bentomat® (b), and Clavmax®
(c)GCLs. i
Figure 3.8. Compaction curve for Sludge A. !
Figure 3.9. Compaction curve for Sludge B.
Figure 3.10. Compaction curve for Sludge C.
Figure 4.1. Compaction curves and field compaction data for Parkview (a) and
Valley Trail (b) test pads. . i
Figure 4.2. COLDICE project field test layout (from Erickson et al. 1994).
!
Figure 4.3. Schematic of in situ box infiltrometer (from Erickson et all 1994).
xii
-------
Figure 4.4. Flexible-wall permeameter. '•
F:igure4.5. Large flexible-wall permeameter used for measuring hydraulic
conductivity of block specimens. \
Figure 4.6. Trimming of block specimen (a) and separating block specimen from
underlying soil using a wire saw (b). i
Figure 4.7. CRREL frozen soil core barrel sampler. '
Figure 4.8. Temperature vs. time for Valley Trail clay during three-dimensional
freezing. i
Figure 4.9. One-dimensional freezing set-up. j
Figure 4.10. Freezer and data-acquisition system used for one-dimensional
freezing of compacted clay specimens. !
(
Figure 4.11. Temperature vs. time for one-dimensional freezing of Parkview clav in
the laboratory. ; •
Figure 4.12. Compaction curve and field compaction data for siand-bentonite
mixture.
Figure 4.13. Schematic of GCL test pan used at the COLDICE field site (from
Erickson etal. 1994).
Figure 4.14. Photograph of trimmed Bentomat® specimen for laboratory testing.
Figure 4.15. Temperature vs. time for Bentomat® GCL frozen three-dimensionally.
Figure 4.16. Small-scale field pipe specimens. I
Figure 4.17. Compaction of paper mill sludge pipe specimen.
Figure 4.18. Data acquisition set-up for buried pipe specimens. i
;
Figure 4.19. Burial of pipe specimens in the ground outside the Environmental
Geotechnics Laboratory at the University of Wisconsin-Madison.
Figure 4.20. Photograph of sliced sludge specimen for hydraulic! conductivity
testing in flexible-wall permeameter. ;
Figure 4.21. Test set-up for hydraulic conductivity testing of compacted paper mill
sludge specimens. ;
XIII
-------
Figure 5.1. Frost depth vs. time in the Parkview test pad (from Chamberlain 1994
personal communication). i
Figure 5.2. Frost depth vs. time in the Valley Trail test pad (from'chamberlain
1994, personal communication). !
Figure 5.3. Photograph of soil inside box infiltrometer-Parkview test pad.
Figure 5.4. Hydraulic conductivity vs. depth for specimens removed from the
Parkview (a) and Valley Trail (b) test pads. !
Figure 5.5. Interior of block specimens removed before winter from the Parkview
(depth = 0-0.3 m) (a) and Valley Trail (depth = 0.6-0.9 m) i(b) test pads.
Figure 5.6. Interior of block specimens removed after winter from a depth of 0-0 3
m from the Parkview (a) and Valley Trail (b) test pads.
Figure 5.7. Interior of block specimens removed after winter from a depth of 0 6-
0.9m from the Parkview (a) and Valley Trail (b) test padsj
Figure 5.8. Specimens removed from the Parkview and Valley Trail test Dads
using a thin-wall tube (a) and their internal structure (b). :
' ' - '-
Figure 5.9. Specimen removed from Parkview test pad while frozen using CRREL
core barrel (from Erickson 1994, personal communication).
Figure 5.10. Test pit excavated in Valley Trail test pad showing: blocky soil
structure caused by freeze-thaw. . i y
Figure 5.11, Hydraulic conductivity vs. molding water content for Parkview clay.
Figure 5.12. Hydraulic conductivity vs. molding water content for Valley Trail clay.
Figure 5.13. Hydraulic conductivity vs. number of freeze-thaw cycles for Parkview
(a) and Valley Trail (b) clay. i
Figure 5.14. Hydraulic conductivity vs. freezing rate for Parkview clay.
Figure 5.15. Hydraulic conductivity vs. freezing rate for Valley Trail clay!.
Figure 5.16. Interior of Parkview specimen frozen and thawed in the laboratory.
Figure 6.1. Depth of frost penetration vs. time for the COLDICE sand-bentonitp
test pad (after Chamberlain 1994, personal communication).
Figure 6.2. Hydraulic conductivity vs. molding water content for sand-bentonite.
XIV
-------
Figure 6.3. Hydraulic conductivity vs. number of freeze-thaw cycles for sand-
bentonite. 1
Figure 6.4. Photograph of sand-bentonite block specimens before! exposure to
freeze-thaw (a) and after exposure to freeze-thaw (b) showing internal
structure and bentonite-rich zones.
I
Figure 6.5. Photograph of internal structure of sand-bentonite specimen
compacted in the laboratory and subjected to freeze-thaw.
Figure 6.6. Photograph of the internal structure of sand-bentonite specimen
removed from the COLDICE test pad after winter using a thin-wall
tube.
• • i •
Figure 7.1. Temperature vs. time beneath the GCLs in the COLDICE test ponds
for the winters of 1992-93 (a) and 1993-94 (b) (from Chamberlain
1994, personal communication).
i
Figure 7.2. Photograph of the two Bentomat® specimens removed from the GCL
test ponds at the COLDICE field site. !
Figure 7.3. Results of freeze-thaw tests on specimens of Bentofix® frozen arid
thawed in the laboratory.
Figure 7.4. Results of freeze-thaw tests on specimens of Bentomat® frozen arid
thawed in the laboratory. ;
Figure 7.5. Results of freeze-thaw tests on specimens of Claymax® frozen and
thawed in the laboratory.
Figure 7.6. Photograph of section of frozen Claymax® GCL.
Figure 7.7. Photograph of hydrated Bentomat® specimens before and after
freeze-thaw.
Figure 8.1. Temperature vs. time at various depths within field specimen
consisting of paper mill sludge A.
Figure 8.2. Temperature vs. time at various depths within field specimen
consisting of paper mill sludge B. »
Figure 8.3. Temperature vs. time at various depths within field specimen
consisting of paper mill sludge C.
Figure 8.4. Hydraulic conductivity vs. molding water content for paper mill sludge
A.
xv
-------
Figure 8.5. Hydraulic conductivity vs. molding water content for paper mill sludge
- B. ;
i
Figure 8.6. Hydraulic conductivity vs. molding water content for paper mill sludge
C.
Figure 8.7. Hydraulic conductivity vs. number of freeze-thaw cycles for low-K
specimens (a) and as-received specimens (b): paper mill sludge A.
Figure 8.8. Hydraulic conductivity vs. number of freeze-thaw cybles for low-K
specimens (a) and as-received specimens (b): paper mill sludge B.
- - |
Figure 8.9. Hydraulic conductivity vs. number of freeze-thaw cycles for low-K
specimens (a) and as-received specimens (b): paper mjll sludge C.
Figure 8.10. Hydraulic conductivity vs. effective stress for paper mill sludges A, B,
and C. . .:
Figure 8.11. Hydraulic conductivity vs. time for paper mill sludge A. >
Figure 8.12. Hydraulic conductivity vs. time for paper mill sludge B. :
i
Figure 8.13. Hydraulic conductivity vs. time for paper mill sludge C. :
Figure B.1. Air temperature vs. time during winter of 1993-94 at srhall-scale field
test site.
XVI
-------
LIST OF TABLES
Table 3.1.
. Table 5.1.
Table 5.2.
Table 5.3.
Table 5.4.
Table 5.5.
Table 5.6.
Table 5.7.
Table 6.1.
Table 6.2.
Table 6.3.
Table 6.4.
Table 7.1.
Table 7.2.
Table 7.3.
Table 8.1.
Results of GCL mass per unit area and free swell tests.:
Results of hydraulic conductivity tests on block specimens removed
from the Parkview test pad. !
Results of hydraulic conductivity tests on block specimens removed
from the Valley Trail test pad. ;
Results of hydraulic conductivity tests on thin-wall tube specimens
removed from the COLDICE test pads. ;
i
Results of hydraulic conductivity tests on specimens removed from the
COLDICE test pads with CRREL frozen core barrel (from Chamberlain
1994, personal communication).
^ i
Results of freeze-thaw tests on specimens of Parkview and Valley
Trail clay compacted in the laboratory. ;
Summary of hydraulic conductivity tests performed on Parkview field
and laboratory specimens. ;
Summary of hydraulic conductivity tests performed on Valley Trail
field and laboratory specimens. ;
Results of hydraulic conductivity tests on block specimens removed
from the sand-bentonite test pad.
Results of hydraulic conductivity tests on thin-wall tube specimens
removed from the sand-bentonite test pad.
Results of freeze-thaw tests on sand-bentonite specimens compacted
in the laboratory. , i
Summary of freeze-thaw test results for sand-bentonite mixture.
Summary of hydraulic conductivity test results for the GCL test pans
used in the COLDICE project. '.
Results of hydraulic conductivity tests on GCL specimens removed
from the COLDICE test ponds.
Summary of freeze-thaw tests on GCL laboratory specimens.
Results of hydraulic conductivity tests on paper mill sludge conducted
in the pipes.
XVII
-------
Table 8.2.
Table 8.3.
Table A. 1.
Table A.2.
Table A.3.
Table A.4.
Table A.5.
Table A.6.
Results of hydraulic conductivity tests on sludge
flexible-wall permeameters.
specimens tested in
Summary of freeze-thaw test results for paper mill sludges
Results of compaction and hydraulic conductivity tests
Parkview clay.
Results of compaction and hydraulic conductivity tests
Valley Trail clay.
Results of compaction and hydraulic conductivity tests
sand-bentonite.
Results of compaction and hydraulic conductivity tests
paper mill sludge A.
Results of compaction and hydraulic conductivity tests
paper mill sludge B.
Results of compaction and hydraulic conductivity tests
paper mill sludge C.
XVIII
performed on
performed on
performed on
performed on
performed on
performed on
-------
ACKNOWLEDGMENTS
i •
!
Financial support for this study was provided by the United States
Environmental Protection Agency (EPA) under cooperative agreement CR 821024-
01-0 and by the National Council of the Paper Industry for Air and Stream
Improvement (NCASI). Mr. Robert Landreth was the Project Officer for EPA and Mr.
Van Maltby was the contact person at NCASI. Support for sampling andlhydraulic
conductivity testing of the block specimens was provided by Jamies Clem
Corporation and Colloid Environmental Technologies Company (CETCO). This
report has not been reviewed by NCASI, James Clem, or CETCQ and no
endorsement should be implied.
Appreciation is extended to Mr. Allan Erickson of CH2M Hill, Inc. and Mr.
Edwin Chamberlain of the U. S. Army Corps of Engineers Cold Regions Research
and Engineering Laboratory (CRREL). Mr. Erickson and Mr. Chamberlain
permitted sampling of the COLDICE test pads, GCL test ponds and pans, and
provided field and laboratory data. Their assistance and cooperation were
essential to the successful completion of this study.
XIX
-------
SECTION 1
INTRODUCTION
The low hydraulic conductivity of compacted clays has led to their frequent use
in landfill liners and covers, caps over contaminated soil, and other applications where
minimizing fluid flow is desired. However, the hydraulic integrity of compacted clay
can be compromised if it freezes. When the temperature of compacted clay falls below
0°C, water present in the clay freezes and ice lenses form. Concurrently, soil below
the freezing front desiccates as water is pulled to the growing ice lenses: (Benson and
Othman 1993, Chamberlain et al. 1995). As the temperature of the clay later rises
above freezing, the ice lenses thaw, leaving behind a network of cracks that allow
rapid transmission of water (Chamberlain et al. 1990, Benson and Othman 1993).
Because of the detrimental impact that freeze-thaw has on compacted clay,
alternative materials are being considered for use in hydraulic barrier applications in
northern climates. Three alternative materials and two clays were examined in this
study. The alternative materials were a sand-bentonite mixture, geosynthetic clay
liners (GCLs), and paper mill sludges. Previous research indicates thatjthe hydraulic
conductivity of sand-bentonite mixtures decreases after exposure to freezing and
thawing (Wong and Haug 1992). Recent studies have also suggested that GCLs are
resistant to the detrimental effects of freeze-thaw (e.g., Geoservices 1989, GeoSyntec
1991, Shan and Daniel 1991). Because of their high water content and fiprous nature,
paper mill sludges may also be resistant to increases in hydraulic conductivity.
The objective of this study was to evaluate and compare the effect! that freezing
and thawing has on the hydraulic conductivity of clays, sand-bentonite mixtures, GCLs,
and paper industry sludges under laboratory and field conditions, to meet this
objective, a battery of hydraulic conductivity tests were conducted in the laboratory on
specimens prepared under conditions that yield low hydraulic conductivity. The
hydraulic conductivity of each specimen was measured before and after the specimen
had been exposed to a specified number of freeze thaw cycles. Results of the
laboratory tests were compared to data from the COLDICE project, a recent large-
scale field study evaluating the impact of freeze-thaw on compacted clays and GCLs
(Chamberlain et al. 1994), and small-scale field tests using paper mill jsludges that
were conducted as part of this study. i
-------
SECTION 2 i . '
BACKGROUND
2.1 IMPACTS OF FREEZE-THAW ON COMPACTED CLAY
The deleterious effect that freeze-thaw has on the hydraulic; conductivity of
compacted clays is well documented. Chamberlain et'al. (1990) have shown that
changes in soil structure that occur during freeze-thaw can result in significant
increases in the hydraulic conductivity of compacted clays. They compacted five clays
at optimum water content and measured their hydraulic conductivities in
consolidometers. For four of the clays, the hydraulic conductivity increased one to two
orders of magnitude after being exposed to freeze-thaw. Chamberlain et al. (1990)
attributed the increases in hydraulic conductivity to macroscopic horizontal and vertical
cracks that formed during freeze-thaw. j
Othman and Benson (1993a) examined how freeze-thaw affected the hydraulic
cpnductivity of three compacted clays. Specimens were compacted i using standard
and modified Proctor efforts at water contents equal to or exceeding ioptimum water
content. Then, they were subjected to one or three-dimensional freeze-thaw. Factors
investigated were molding water content, compactive effort, dimensionality of freezing,
ultimate temperature below freezing, and temperature gradient. Othman and Benson
(1993a) found that the hydraulic conductivity of all three clays increased about two
orders of magnitude and attributed the increases in hydraulic conductivity to a network
of cracks that formed during freezing. Results of their study are shown in Figure 2.1.
Similar results have been documented by other investigators (eJg., Zimmie and
LaPlante 1990, Kim and Daniel 1992, Bowders and McClelland 1994).; Othman et al.
(1994) provide a review and synthesis of these studies. They report that compacted
clays having an initial hydraulic conductivity less than 1x10'9 m/s generally undergo
an increase in hydraulic conductivity of 1 to 2 orders of magnitude when subjected to
freeze-thaw. Greater increases in hydraulic conductivity occur for clays having lower
initial hydraulic conductivity, specimens frozen more quickly, and tests conducted at
lower effective stress. :
Three studies have been performed in the field to assess the impact of freeze-
thaw on the hydraulic conductivity of compacted clay. Benson and Othman (1993)
performed small-scale field tests. Two specimens of compacted clay were prepared in
PVC pipes (diameter = 0.30 m, length = 0.91 m). One specimen was kept in the
-------
laboratory as a "control" specimen, and was permeated to determine the as-
compacted hydraulic conductivity. The other specimen ("field" specimen) was buried
in the ground in January 1991 and subjected to two months of winter! weather. The
field specimen was instrumented with thermocouples and fiberglass moisture probes
at various depths to monitor temperature and water content while the specimen was in
the ground. |
The field specimen was removed from the ground in Mafch 1991 and
permeated in the laboratory. No significant change in overall hydraulic conductivity
was observed between the control and field specimens. However, Jwhen the field
specimen was sliced horizontally into separate specimens (which were then
permeated in a large-scale flexible-wall permeameter), increases in hydraulic
conductivity of 1.5 to 2 orders of magnitude were observed for specimens sliced from
above the freezing plane. In contrast, a specimen removed from just below the
freezing plane showed a slight increase in hydraulic conductivity, and the specimen
removed from 0.3 m below the freezing plane showed no'increase in hydraulic
conductivity. Figure 2.2 shows the results obtained by Benson and Othman (1993).
Benson and Othman (1993) attributed the increase in hydraulic conductivity above the
freezing plane to horizontal and vertical cracks that formed as a result bf freeze-thaw.
They attributed the increase in hydraulic conductivity just below the freezing plane to
vertical cracks that formed as a result of desiccation, which occurred as a result of
water redistribution during freezing. ;
Benson et al. (1995) conducted a full-scale field test to determine if increases in
hydraulic conductivity occur in the field that are similar to those observed in laboratory
tests. A test pad was constructed and instrumented at a site in southeastern Michigan.
Temperatures within the test pad were monitored and recorded throughout the winter
of 1992-93. I
Hydraulic conductivity of the test pad was measured in the laboratory and in the
field before and after winter. Laboratory hydraulic conductivity tests were performed in
flexible-wall permeameters on large block specimens (diameter = OJ30 m) and on
small specimens (diameter = 71 mm) taken from the test pad using thin-wall sampling
tubes (Shelby tubes). Field measurements of hydraulic conductivity w4re made using
sealed double-ring infiltrometers (SDRIs). !
Results of the laboratory tests indicated that an increase in hydraulic
conductivity occurred as a result of freeze-thaw. Increases in hydraulic conductivity
were seen in specimens extracted from the zone in which freezing occurred (depth <
-------
i
'•5
o
O
Figure 2.1.
0 i 23 4 5 6 :
Number of Freeze-Thaw Cycles ;
Hydraulic conductivity vs. number of freeze-thaw cycles for three
Wisconsin clays (after Othman and Benson 1993a). I
.c
a.
-------
0.5 m), whereas no increase in hydraulic conductivity was found for specimens taken
from below the depth of frost penetration. In contrast, results of the SDRI tests showed
no increase in hydraulic conductivity after winter exposure. Benson et al. (1995) state
that no increase in hydraulic conductivity was observed in the SDRI te;sts because soil
below the depth of freezing controlled the infiltration rate. ;
Erickson et al. (1994) examined the effect of freezing and thawing on two
compacted clay test pads constructed for the COLDICE (Construction Of Liners
Deployed In Cold Environments)/CPAR (Construction Productivity Advancement
Research) project at a landfill near Milwaukee, Wisconsin. The test pads were
instrumented with thermistors and a weather station was used to'record the soil
temperature distribution, number of freeze-thaw cycles, depth of frost penetration and
climatic conditions throughout the winter. Hydraulic conductivities of the test pads
were measured before and after exposure to winter weather. j
Hydraulic conductivities of the test pads were assessed in the laboratory using
specimens collected by three sampling methods: large blocks, thin-wall sampling
tubes (Shelby tubes), and as cores, which were removed from the test pads while the
soil was frozen. The large block specimens (diameter = 0.30 m) were taken before
and after the winter of 1992-93.. Thin-wall sampling tubes (diameter == 71 mm) were
used to remove specimens in June 1993. Frozen cores (diameter =, 71 mm) were
collected in March 1993 and March 1994. i
Increases in hydraulic conductivity of up to 4 orders of magnitude were
observed for the specimens removed from the soil subjected to freeze-thaw. Erickson
et al. (1994) state that the increase in hydraulic conductivity of the compacted clay is a
result of the formation of cracks due to the formation of ice lenses and shrinkage of the
soil caused by redistribution of water. i
2.2 BENTONIT1C BARRIER MATERIALS
2.2.1 Sand-Bentonite Mixtures !
Wong and Haug (1991) examined how freeze-thaw affected! the hydraulic
conductivity of sand-bentonite mixtures consisting of a sodium bentonite from western
Canada and Ottawa sand. Varying amounts of sodium bentonite were rnixed with the
Ottawa sand and the mixtures were compacted according to ASTM D i698 (standard
Proctor). The specimens were extruded and permeated in fixed- and flexible-wall
permeameters. After initial permeation was complete, the specimens were exposed to
an ultimate temperature of -20°C for a minimum of 6 hours in a closed system (no
-------
external water supply). The specimens were then allowed to Ithaw at room
temperature. After thawing, their hydraulic conductivity was measured. This
procedure was repeated until 5 freeze-thaw cycles were completed. ':
Figure 2.3 shows the results of their study. A decrease in hydraulic conductivity
was observed in all specimens as a result of freeze-thaw. Wong arid Haug (1991)
provided two possible explanations for the decrease in hydraulic conductivity: (1)
freeze-thaw promotes hydration of the bentonite, lowering the hydraulic conductivity
towards long-term test values or (2) thaw consolidation compresses the bentonite into
gaps between the sand grains. !
Erickson et al. (1994) examined how freeze-thaw affected a sand-bentonite
mixture in the field. A test pad with a sand-bentonite mixture was constructed for the
COLDICE project, adjacent to the two clay test pads described in Section 2.1. A large
box infiltrometer (1.3 m x 1.3 m) was constructed in the field to assess the in situ
hydraulic conductivity of the test pad after winter. Hydraulic conductivity was also
measured on specimens removed from the test pad as blocks, using thin-wall tubes,
and as frozen cores. The hydraulic conductivity of the sand-bentonite test pad was 4 x
10-10 m/s before winter and 5 x 10"10 m/s after winter. j
Erickson et al. (1994) concluded that sand-bentonite mixtures that are mixed
properly with a large enough percentage of bentonite can be resistant to increases in
hydraulic conductivity caused by freeze-thaw.
2.2.2 Geosynthetic Clay Liners (GCLs) I
Several studies have been performed to evaluate how freeze-thaw affects the
hydraulic conductivity of geosynthetic clay liners (GCLs). Geoservices (1989), Shan
(1990), and Chen-Northern (1988) studied how freeze-thaw affects: the hydraulic
conductivity of the GCL Claymax®. Geoservices (1989) measured an mitial hydraulic
conductivity of 4 x 10-12 m/s for ciaymax®. The hydraulic conductivity tests were
conducted at an effective stress of 196 kPa and hydraulic gradient of Approximately
1000. The hydraulic conductivity of the specimens after 10 cycles of freeze-thaw was
1.5x10-12 m/s. ;
Shan (1990) measured the hydraulic conductivity of Claymax® before and after
5 cycles of freeze-thaw. The samples were permeated using an effective stress of 14
kPa and a hydraulic gradient of 10. The initial hydraulic conductivity was 2.0 x 10'11
m/s. After 5 cycles of freeze-thaw, the hydraulic conductivity was 2.2 x 10-"11 m/s.
-------
^o,
&
>
O
o
10
-7
10"
10
-9
10
-10
T—i—i—r
4,5% bentonite i
6.0% bentonite !
8.3% bentonite '
13% bentonite
25% bentonite '
1 2 34
Number of Freeze-Thaw Cycles
£— :^L_^_ ?
— i — i — i — i — i — i — -i — i i i ••
A
V _
j . 1 . .
O ; -
-
" V— : i
_j i \ ,_,_. 1, ,
Figure 2.3. Hydraulic conductivity vs. number of freeze-thaw cycles for sand-
bentonite mixtures (after Wong and Haug 1991). !
-------
Chen-Northern (1988) subjected specimens of Claymax® to 10 cycles of freeze-
thaw. The specimens were permeated after 0, 3, and 10 cycles of freeze-thaw using
procedures similar to those used by Shan (1990). The initial hydraulic conductivity
was 1.0 x-10'11 m/s. After 3 freeze-thaw cycles the hydraulic conductivity was 2.3 x 10-
11 m/s, and after 10 freeze-thaw cycles, it was 2.2 x 10-11 m/s. •
GeoSyntec (1991) studied how freeze-thaw affected the hydraulic conductivity
of the GCL Bentomat®. Initially and after each freeze-thaw cycle, the specimen was
permeated with de-aired water at an effective stress of 34.5 kPa and hydraulic gradient
of 30. Flexible-wall permeameters were used. The hydraulic conductivity of the
Bentomat® fluctuated between 1 x 10'11 and 6 x 10'11 m/s for various cycles of freeze-
thaw, with no increasing or decreasing trends being observed. !
Freeze-thaw tests were also performed on Bentomat® by Robert L. Nelson and
Associates, Inc. (1993). Two sets of tests were performed. In the first set, several
specimens were permeated after undergoing a designated numberi of freeze-thaw
cycles (up to six cycles). In the second set, a single specimen was used that was
permeated after each freeze-thaw cycle (up to 10 freeze-thaw cycles)! The hydraulic
conductivities ranged between 1.1 x 10-11 m/s and 4.0 x 10'11 m/s for the first set of
tests, and 1.9 x 10-11 m/s and 3.3 x 10'11 m/s for the second set. Results of the tests
performed by Robert L. Nelson & Associates, Inc. (1993) are shown in Figure 2.4.
2.3 PAPER MILL SLUDGE 'i
Because some pulp and paper mill sludges (wastes createp! in the paper
making process) have been shown to possess properties similar to clays, interest has
arisen in using them in construction of landfill liners and covers (NCASI 1989).
Nonetheless, information regarding the geotechnical properties of paper mill sludge is
scant (Genthe 1993).
The National Council of the Paper Industry for Air and Stream Improvement
(NCASI) performed hydraulic conductivity tests on 11 paper mill sludges of various
origins (NCASI 1989). Tests were performed on specimens compacted using
standard Proctor procedures (ASTM D 698), except only 10 blows were applied per
lift. The specimens were compacted at their as-received water content, which ranged
fromi 120 to 409%. Their hydraulic conductivity ranged from 4.2 x 10J6 to 5.8 x 10-10
m/s. The variability in the hydraulic conductivity test results was prob[ably caused by
the variability in molding water content and sludge type. j
8
-------
10-9
;f io-10
TJ
O
O
10
-12
10'9
>>
> 1
-G
X)
O
O
I
•o
>,
x
io-12
(a)
2 34 5
Number of Freeze-Thaw Cycles
(b)
0 2 4 6 8 10
Number of Freeze-Thaw Cycles
Figure 2.4. Hydraulic conductivity vs. number of freeze-thaw cycles for Bentomat® for
test set 1 (a) and test set 2 (b) (after Robert L. Nelson & Associates 1991).
-------
Zimmie et al. (1994) examined the geotechnical properties of a paper mill
sludge from Erving Paper Co. in Massachusetts. Compaction, hydraulic conductivity,
consolidation, and shear strength tests were performed. The effects that freeze-thaw
and effective stress have on the hydraulic conductivity of sludge were also examined.
Zimmie et al. (1994) report that the compaction curve for the sludge was
developed by gradually drying the sludge from its high as-received wafer content and
compacting specimens at various water contents using standard Proctor compaction
effort (ASTM D 698). The maximum dry unit weight and optimum water content were
determined, and the curve was similar to curves for compacted clays.; However, the
maximum dry unit weight (8.1 kN/mS) was much lower than that for most clays,
whereas optimum water content (49%) was much higher. I
The compacted specimens were permeated in flexible-wall pejrmeameters to
determine the hydraulic conductivity-water content relationship. The hydraulic
conductivity was found to be much lower for specimens compacted at [water contents
wet of optimum, relative to the hydraulic conductivity of specimens compacted at water
contents near or dry of optimum. This behavior is similar to that of compacted clays
(Mitchell et al. 1965, Benson and Daniel 1990, Daniel and Benson 1990). However,
the lowest hydraulic conductivities were obtained for specimens compacted 50% wet
of optimum water content, where for clays the lowest hydraulic conductivities are
generally obtained 1-3% wet of optimum. For sludge specimens compacted greater
than 50% wet of optimum, hydraulic conductivities less than 1 x 10'9 m/s were
obtained. , ;
Zimmie et al. (1994) also performed hydraulic conductivity tests
-------
compaction, the specimens were wrapped in plastic to prevent desiccation and frozen
one-dimensionally. After the desired number of freeze-thaw cycles,! the specimens
were placed in flexible-wall permeameters and permeated at effective stresses of 34.5,
69, and 138 kPa using a hydraulic gradient of 21. ;
Results of the freeze-thaw tests showed that the paper mill sludge was affected
by freeze-thaw (Figure 2.5). The hydraulic conductivity of the sludge increased
approximately one order of magnitude after 10 freeze-thaw cycles. Similar increases
in hydraulic conductivity occurred at each effective stress. However; increasing the
effective stress from 34.5 kPa to 138 kPa decreased the hydraulic conductivity of the
sludge approximately one order of magnitude at each freeze-thaw cycle; (Fig. 2.5).-
Maltby and Eppstein (1994) describe a field study in which paper mill sludges
were used in landfill cover test cells. The field study was undertaken to compare the
performance of paper mill sludge and compacted clay when used as hydraulic
barriers. The comparison was based on water balance computations (e.g., runoff and
percolation). One of the two test cells containing paper mill sludge was constructed
with a combined (primary and secondary treatment) sludge and the other was
constructed with a primary sludge. Construction of the test cells wa;s completed in
November 1987. I
Maltby and Eppstein (1994) observed after 5 years that: (1) the test cells
constructed with sludge consolidated more than those constructed with clays, (2) the
cells containing sludge had greater amounts of runoff than the cells constructed with
clay,, (3) percolation was lower for the cells containing paper mill sludge, and (4) the
average field hydraulic conductivities for the cells containing sludge were lower than
those containing compacted clay (9.6 x 10-10 m/s and 4.4 x 10-9 m/s for the cells
containing sludge; 1.4 x 10-8 m/s and 1.5 x to-8 m/s for the cells containing compacted
clay).
11
-------
U)
E
o
O
.o
10-io
10-11
effective stress = 34.5 kPa;
effective stress = 69 kPa i
effective stress = 138 kPa •
5 10 15 20
Number of Freeze-Tf\aw Cycles
25
Figure 2.5. Hydraulic conductivity vs. number of freeze-thaw cycles at various
effective stresses (after Zimmie etal. 1994) ;
12
-------
SECTION 3
MATERIALS
i
. •' . . _ .... _ ... . . . . ! '
3.1 CLAYS
3.1.1 Index Properties !
Two clays were used in the testing program: Parkview clay and Valley Trail clay.
Parkview clay is a low plasticity glacial till, whereas Valley Trail clay jis a moderately
plastic glacio-lacustrine clay. Both clays have been used in the construction of landfill
liner and cover systems and were used to construct test pads for the CpLDICE project
(Sec. 2.1). Bulk samples of both clays used in this project were obtained from the
COLDICE test pads. :
Parkview clay is from a glacial deposit in Germantown, Wisconsin. It has a
liquid limit of 30, plastic limit of 15, and plasticity index of 15 (ASTfv} D 4318). The
location of Parkview clay on the plasticity chart is showh in Fig. 3.1J. Particle size
analyses of the clay (Fig. 3.2) showed that it is composed of 0.5% gravfel, 11.5% sand,
and 88.0% fines (particle sizes < 0.075 mm) based on definitions in the Unified Soil
Classification System (USCS) (ASTM D 421). Clay content (particles finer than 2 |im)
was found to be 36.5% by hydrometer analysis (ASTM D 422). Specific gravity tests
on Parkview clay showed the clay has a specific gravity of 2.81 (ASTM D 854).
Parkview clay is classified as CL.in the USCS (ASTM D 2487). l
Valley Trail clay is from a lacustrine deposit in Berlin, Wisconsin. It has a liquid
limit of 45, plastic limit of 19, and plasticity index of 26. The location of Valley Trail clay
on the plasticity chart is shown in Fig. 3.1. Particle size analyses of the clay (Fig. 3.2)
showed that it is composed of 0.6% gravel, 3.8% sand/and 95.6% fines based on
definitions in the USCS. Clay content was found to be 51.2% by hydrometer analysis.
Specific gravity tests on Valley Trail clay showed the clay has a specific gravity of 2.77.
Valley Trail clay is classified as CL-CH in the USCS. '
3.1.2 Compaction :
Compaction tests were conducted on the two clays to determine the
relationships between dry unit weight, water content, and compactive effort. Two
compactive efforts were used: standard Proctor (ASTM D 698) and modified Prortor
(ASTM D 1557). . i
13
-------
X
0)
-a
"o
'w
SS
a.
• Parkview Clay
• Valley Trail Clay
50
Liquid Limit (LL)
1 100
Figure 3.1. Plasticity chart showing locations of Parkview and Valley! Trail clays.
14
-------
CD
IE
IE-
CD
CD
Q.
100
80
60
40
20
0
PI
in itJ i g^ MJ i i. i , | mi I | | | III 1 1 1 1 I — nil i i i i — |
" E3 o
• ^ ra
- • • o Q
o 3 -
- ° Sra
- O ^
0 0 . -
0 Testl ^o
D Test 2 u
o Test 3
( n\ -
V"/
"I'll 1 1 Inn 1 1 1 . 1 1 1,,, | , , , , . [,,,, , l ( |
10
0.1 0.01 0.001 0.
Particle Size (mm)
0001
CD
C
LJL
i*_i
CD
e
CD
Q-
100
80
60
40
20
0
n
og
o
o
o
o Test 1
n Test 2
o Test 3
"II 1 - 1 - Illll 1 1 1 - 1 - In, ..... |,
(b)
10
1 0.1 0.01 0.001 0.0001
Particle Size (mm)
Figure 3.2. Results of particle size analyses for Parkview (a) and
clays.
15
Valley Trail (b)
-------
Prior to compaction, the soil was air dried and crushed to pass the No. 4 sieve
(particle size < 4.75 mm). The soil was then uniformly mixed with tap water from
Madison, Wisconsin to a pre-determined water content. The moistened clay was
sealed in a plastic bag or bucket and allowed to hydrate for at least 24 hours prior to
compaction. '
Compaction curves are shown in Figs. 3.3 and 3.4. For Parkview clay, optimum
water contents of 14.1 and 9.6% were obtained for standard and mbdified Proctor
compactive efforts, respectively. The maximum dry unit weights corresponding to
these water contents are 18.8 and 21.1 kN/m3. For Valley Trail clay, optimum water
contents of 17.9 and 14.2% were obtained for standard and mqdified Proctor
compactive efforts, respectively. The maximum dry unit weights corresponding to
these water contents are 17.3 and 19.1 kN/m3. ;
3.2 BENTONITIC BARRIER MATERIALS
3.2.1 Sand-Bentonite Mixture
3.2.1.1 Index Properties '.'•!.
One sand-bentonite mixture was used in the testing program. The sand-
bentonite mixture was obtained from one of the COLDICE test pads (Sec. 2.2.1). The
sand-bentonite mixture was prepared in the field using a pugmill prior to construction
of the test pad (Erickson et al. 1994). The sand component is a clean mortar sand,
which was purchased from a concrete supplier near the landfill site. The sand is a
poorly graded, clean, medium to fine sand and is classified as SP in thei USCS. More
than 90% of the sand passed the No. 30 sieve and less than 5% passed the No. 200
sieve (Fig. 3.5). The bentonite component is a granular sodium bentohite (American
Colloid CG-50) with no polymer additives. !
The sand-bentonite mixture has a liquid limit of 42, plastic limit of 29, and
plasticity index of 13. Methylene blue titration tests were performed to measure the
bentonite content of the sand-bentonite. Measurements were made onjgrab samples
of the mixture from a drum stored in the laboratory. The average bentonite content
was 12% by weight. Specific gravity tests showed that the sand-bentonite mixture has
a specific gravity of 2.70. • ;
16
-------
I
o>
i
.tr
c
Q
^
f
§
'c
ID
0
C.C.
21
20
19
18
17
16
I
. • • ' ' i ' ' ' i i i i i i i — i — i — i — i — | — i — i — i — i —
1 O Standard
r^" N n Modified ~
y f-i C" „. , ,.
: % n :
Dn :
o QD -
'. o \
— i — i — i — ' i i • • • i . . i . i i , , , i , , , , ~
3 5 10 15 20 2£
Molding Water Content (%)
Figi
20
19
18
17
16
15
jre 3.3. Compaction curves for Parkview clay:
7"1 ' ' n • • ' i i i i i i 'i i i i i i •— | — i — i — i —
O Standard
^(ttp — \ D Modified J
^^& \. '-
an :
\n ~
°/^\° ^
o/ ^\p
c/° \ '-
0 7 xo :
o o .
—> — > — 1 — 1 — 1 — 1 — 1 — 1 1 . . 1 . . , 1 . , , 1 , , ,
8 12 16 20 24
Molding Water Content (%)
28
Figure 3.4. Compaction curves for Valley Trail clay.
17
-------
o
HZ
I
CD
Q_
100
80
60
40
20 -
O Test 1
D Test 2
O Test 3
Mill - 1 - 1 - 1
^&S_
10
1 0.1
Particle Size (mm)
0.01
Figure 3.5. Results of particle size analysis for sand component of sand-bentonite
mixture.
18
-------
3.2.1.2 Compaction , i
Compaction tests were conducted to determine the relationships between dry
unit weight, water content, and compactive effort. Two compactive efforts were used:
standard Proctor and modified Proctor. !
Prior to compaction, the sand-bentonite was air dried and crushed to pass the
No. 4 sieve. The sand-bentonite was then uniformly mixed with itap water from
Madison, Wisconsin to a pre-determined water content. The moistened mixture was
sealed in a plastic bag or bucket and allowed to temper for at least 24 hours prior to
compaction. |
Compaction curves for the sand-bentonite mixture are shown in Fig. 3.6.
Optimum water contents of 16.5 and 12.3% were obtained for standard and modified
Proctor compactive efforts, respectively. The maximum dry unit weights corresponding
to these water contents are 17.4 and 18.8 kN/m3. ;
3.2.2 Geosynthetic Clay Liners (GCLs)
Geosynthetic clay liners (GCLs) are commerciaiiy manufactured products that
consist of highly swelling bentonitic clay either sandwiched between two geotextiles or
glued to a geomembrane. Three geosynthetic clay liners were used in this- study:
Bentofix®, Bentomat®, and Claymax®. Two of the GCLs (Bentomat® iand Claymax®)
were used as liners for test ponds and test pans in the COLDICE project (Erickson et
al. 1994). :
Two rolls of each GCL were shipped to the University of Wisconsin-Madison by
their manufacturers. The rolls were wrapped in plastic to prevent absorption of water
and stored in the Environmental Geotechnics Laboratory at the; University of
Wisconsin-Madison. i
Bentofix® is manufactured by the National Seal Company (NSC) by needle-
punching loose granular bentonite between two geotextiles (Fig. 3.7). For this project,
a Bentofix® GCL with a 140 g/m2 upper, woven polypropylene geotextile and a 270
g/m2 lower, non-woven polypropylene geotextile was used. The mass per unit area of
the Bentofix® GCL was 6.8 kN/m2 (ASTM D 5261). Free swell tests performed on the
bentonite from Bentofix® using GRI GCL-1 showed that the average free swell was
18.1 mm. Results of the mass per unit area and free swell tests are .shown in Table
3.1.
Bentomat® is manufactured by the Colloid Environmental |Technologies
Company (CETCO) by needle-punching granular Volclay® bentonite between two
19
-------
«"-
E
-c
O)
Q
20
19
18
17
17
16
1 — 1 — — 1 — 1 _ 1
O Standard
D Modified
8 1.2 16 20
Molding Water Content (%)
24
Figure 3.6. Compaction curves for sand-bentonite.
20
-------
Needle-Punched Fibers
Woven Polypropylene Gedtextile
Non-Woven Polypropylene Geotextile
(a) Bentofix®
Needle-Punched Fibers
Woven Polypropylene Geotextile
Non-Woven Polypropylene Geotextile '
(b) Bentomat®
Woven Polypropylene Geotextile,
Polyester, Open Weave Geotextile'
(c) Claymax®
Figure 3.7. Schematic diagrams of Bentofix® (a), Bentomat® (b), and
0>CLS.
Claymax® (c)
21
-------
Table 3.1. Results of GCL mass per unit area and free swell tests
Specimen Number
Bentofix-1
Bentofix-2
Bentofix-3
Bentomat-1
Bentomat-2
Bentomat-3
Claymax-1
Claymax-2
Claymax-3
Mass per Unit Area
(kg/m2)
7.4
6.5
6.4
6.1
6.1
no test
6.5
6.4
6.4
Free Swell
(mm)
12.3
9.5
7.f
16.5
18.0
19.8
5.6!
6.6
7.9!
22
-------
geotextiles (Fig. 3.7) (Daniel and Estornell 1990). For this project, Bentomat® "CS"
with a 140 g/m2 upper, woven polypropylene geotextile and a 140 g/m2 lower, non-
woven polypropylene geotextile was used. The mass per unit area ofthe Bentornat®
was 6.1 kN/m2. Free swell tests performed on bentonite from Bentomat® showed that
the average free swell was 9.6 mm (Table 3.1). ;
Claymax® is manufactured by the James Clem Corporation in Arlington
Heights, Illinois. It is comprised of amended sodium bentonite sandwiched between
two woven, polypropylene geotextiles (Fig. 3.7). The bentonite is adhered to the two
geotextiles using a non-toxic, water soluble adhesive (Daniel and Estornell 1990).
For this project, Claymax® 200R with a 140 g/m2 upper polypropylene
geotextile and 25 g/m2 lower polyester, open-weave geotextile was used. The mass
per unit area of the Claymax® used was 6.4 kN/m2. Free swell tests performed on
bentonite from Claymax® showed that the average free swell was 6.7 mm (Table 3,1).
3.3 PAPER MILL SLUDGES
3.3.1 Index Tests
Three paper industry sludges were used in the testing program, and are
referred to as sludges A, B, and C. Sludge A is a combined sludge; the sludge is
composed of primary sludge from the clarification of raw wastewater^nd biological
sludge from an activated sludge treatment plant. Sludge A is from a non-integrated
mill, which produces specialty grades and coated and uncoated book grades. Sludge
B is a primary paper mill sludge from a non-integrated mill which produces specialty
coated, lightweight coated, coated bag, and pressure sensitive paper, kludge C is a
combined (primary and biological) sludge from a de-inking mill which produces
disposable garments and napkin and tissue paper. Sludges A and B are from paper
mills in Michigan, whereas Sludge C is from a paper mill in Massachusetts. Bulk
samples ofthe sludges were supplied by the National Council ofthe PapW Industry for
Air and Stream Improvement (NCASI) office in Kalamazoo, Michigan. ;
Attempts were made to determine the liquid and plastic limits of the three
sludges. However, difficulties were encountered when performing the Atterberg limits
tests. The fibrous nature of the sludge prevents changes in the behavior of the sludge
with changes in water content. Therefore, measurements of the liquid and plastic
limits were not possible. Similar difficulties with Atterberg limit tests have been
reported by NCASI (1989) and Genthe (1993). ;
23
-------
<
Sludge A contains 58.9% fines as determined by wash sieving the sludge past
the No. 200 sieve. The ash content of Sludge A is 56.0% (ASTM D 2974).
f-ractionation of Sludge A by wet screening was found to be 78.5% (TAPPI T 261).
The weighted average fiber length of Sludge A is 0.24 mm (TAPPI T 233).
Sludge B contains 75.8% fines as determined by wash sieving the sludge past
the No. 200 sieve. The ash content was found to be 53.1%. Fractionatjon of Sludge B
by wet screening was found to be 85.5%. The weighted average \fiber length of
Sludge B is 0.12 mm.
Sludge C contains 79.8% fines as determined by wash sieving p&st the No. 200
sieve. The ash content of Sludge C is 44.4%. Fractionation of Sludge C by wet
screening is 76.5%, whereas the weighted average fiber length of Sludge C is 0.29
mm. :
3.3.2 Compaction j
Compaction tests were conducted on the three sludges to determine the
relationship between dry unit weight and water content. Standard Proctor compactive
effort was used. ;
Prior to compaction, the sludge was allowed to air dry from its "as-received"
water content. At various times, a grab sample was taken, sealed in a plastic bag or
bucket, and allowed to equilibrate at least 24 hours prior to compaction. This method
was used because previous research has shown that re-wetting completely air-dried
sludge results in a loss of plasticity (NCAS11989, Zimmie et al. 1994). !
Compaction curves for the three sludges are shown in Figs. 3.8-3.10. For
Sludge A, the optimum water content is 40% and the maximum dry unit weight is 8.7
kIN/m3. For Sludge B, optimum water content is 97% and the maximum 'dry unit weight
is 6.0 kN/m3. For Sludge C, the optimum water content is 72%, whereas1 the maximum
dry unit weight is 6.4 kN/m3. j
Results of the compaction tests indicate that sludges havfe compaction
properties similar to those of clays. However, the optimum water contents determined
for the sludges are higher in comparison to typical optimum water contents for clays
(-10-30%) and the maximum dry unit weights are lower than those foriclays (-15-19
kN/m3). :
24
-------
i
D)
I
.*;
C
0 50 100 150
Molding Water Content'(%)
Figure 3.8. Compaction curve for Sludge A.
200:
.E
D)
I
• O
3 -
0 50 100 150 200 250 300
Molding Water Content (%)
Figure 3.9. Compaction curve for Sludge B.
25
-------
D)
50 100 150 200
Molding Water Content (%)
250
Figure 3.10. Compaction curve for Sludge C.
26
-------
SECTION 4
METHODS
4.1 CLAYS
4.1.1 Field Methods i
4.1.1.1 COLDICE Test Pads \
The tests performed on clay in this study were conducted for comparison to field
data collected from tests on large-scale test pads constructed/for the COLDICE project.
The COLDICE project was conducted by CH2M Hill, Inc., the U. S. Army Cold Regions
Research and Engineering Laboratory (CRREL), and a suite of industrial partners at a
landfill near Milwaukee, Wisconsin (Erickson et al. 1994) as k Construction
Productivity Advancement Research (CPAR) project. Assistance was also provided by
the University of Wisconsin-Madison. The objective of the COLDICEJ project was to
evaluate how freeze-thaw affects the hydraulic conductivity of compacted clays and
geosynthetic clay liners (GCLs) at field-scale. \
Four test pads in the COLDICE project were constructed with compacted clay.
Two of the pads were constructed with Parkview clay and two were constructed with
Valley Trail clay. The test pads were labeled: PV-2, PV-3, VT-4, and VT-5
(Chamberlain et al. 1995). Each test pad was 9 m by 21 m. Two.test pads had
thicknesses of 0.6 m (PV-2 and VT-5), and two had thicknesses of 0.9 m| (PV-3 and VT-
4). Two thicknesses were used such that the effects of fully- or partially-penetrating
frost could be examined (Erickson etal. 1994). ;
A layer of high-density polyethylene (HOPE) geomembrane was placed above
a prepared subgrade and a geocomposite drain (non-woven geotextile on each side
of a geonet) was placed above the HOPE geomembrane (Erickson et k\. 1994). Clay
was placed in 0.15 m loose lifts and the lifts were compacted with a Caterpillar 825C
tamping foot compactor at water contents 2-5% wet of standard Proctor optimum.
Relative compaction exceeding 95% of standard Proctor maximum dry unit weight was
obtained for each test pad (Benson et al. 1994). All of the measurements of molding
water content and dry unit weight made during construction fell wet of the line of
optimums (Benson et al. 1994) (Fig. 4.1). Construction of the test pads was completed
in October 1992 (Erickson et al. 1994). ;
The 0.6 m-thick pads (PV-2 and VT-5) were covered with a 0.15 mm thick
polyethylene sheet and 0.1 m of sand to minimize desiccation. Test pad PV-3 was
27 ;
-------
IE
iz
.4_i
.D>
1
H-^
'c
Z>
Q
1=
2
^
•#-»
0)
' ' i i i — i — i — i — | — i — i — i — i— i
O Standard -
D Modified -
9-,^~^Dv + Field :
I — I \^
/ \ -
n rj n
L * \ .:
"- 8^- :
1 / O^jt :
: / ^(j- ~
: o/o an :
— O . Q-}
: o . ^°v -
— i — i — i — i — i — i — i iii... , , , , , i . . .
05 10 15 20 25
Molding Water Content (%)
20
19
18
•\ -7
I"1 ' ' i ' ' • t • ' ' r~> ' ' — T"1 — ' — ' — i — i — i — i — i
O Standard I
O\ D Modified -
^^X + Field
0 X
n — -u \ , :
1 °\n -
\D
\
\
_ n Q_
c
Z)
Q
8 12 16 20 24 28
Molding Water Content (%)
Figure 4.1. Cojrnpaadon^curw.jand field compaction data for Parkview (a) and
28
-------
covered with a needle-punched non-woven geotextile and 0.3 m of gravel. Test pad
VT-4 was covered with 0.3 m of well-graded sand.
An array of thermistors was placed within the test pads and! an automated
weather station was installed at the COLDICE field site (Erickson et ai. 1994).
Measurements of frost depth and climatic conditions were recorded every five minutes
using a Campbell Scientific datalogger. Hourly averages were bomputed and
transferred to a computer at the CRREL laboratory in Hanover, New Hampshire via a
cellular phone and computer modem (Erickson et al. 1994). Figure 4.2 is diagram of
the COLDICE project field test layout. i
! .
4.1.1.2 In Situ Box Infiltrometers
Erickson et al. (1994) installed box i nf i It ro meters to measure ;the field-scale
hydraulic conductivity of the test pads. The infiltrometers were designed to measure
hydraulic conductivity before and after exposure to freeze-thaw without disturbing the
structure of the soil. Infiltrometers were placed in test pads PV-2 and PV-3. No
• infiltrometers were installed in either of the Valley Trial test pads. i
Figure 4.3 is a schematic of the box infiltrometers. Immediately following
construction of the test pads, trenches were excavated to leave a 1.3 m tpy 1.3 m
block of undisturbed soil. An HOPE box with an open top and bottJm was placed
around the block of soil and seamed to the underlying HOPE geomeriibrane. Pipes
and filters were connected to the geocomposite drain to carry any infiltration to a sump
for measurement. The annular space between the block of clay and the HOPE was
filled with a bentonite grout to prevent sidewall leakage. :
A non-woven geotextile was placed on top of the block of soil land was then
covered with a layer of washed gravel. An HOPE lid was seamed to the walls of the
box which contained a riser pipe for adding water to the system and for measuring
inflow. The infiltrometers were covered with 0.6 m of sand to prevent uplift of the lids.
The sand layer was reduced to a thickness of 0.1 m (level with the surrounding
overburden) during winter months to allow penetration of the freezing plane into the
compacted clay. '
4.1.1.3 Laboratory Assessment of Field-Scale Hydraulic Conductivity !
Hydraulic conductivity tests were performed in the laboratory on specimens
removed from the test pads to assess field-scale hydraulic conductivity, j Specimens
were removed from the field by three methods: as large blocks (diameter = 0.4 m)
29
-------
MOTTOSCAU
OAnlNc* ttttnlhlck ttflmlMck I OAnlNck OimlNc*
k
&3~
Sand
/
Storw
aim
Section A-A'
NOtKOCMt
Figure 4.2. COLDICE project field test layout (from Erickson et ai. 1994).
°K>*OO» /SSf2£!?rr»tw
Figure 4.3. Schematic of in situ box infiltrometer (from Erickson et al. 1994).
30
-------
using thin-wall sampling tubes (Shelby tubes, diameter = 71 mm), and as frozen cores
(diameter = 71 mm). The specimens removed as blocks and with thin-wall tubes were
collected by personnel from the University of Wisconsin-Madison. Personnel from
CRREL and CH2M Hill, Inc. obtained the frozen cores.
Large block specimens were removed before freezing in December 1992 and
after thawing in June 1993. Thin-wall sampling tubes were used in |June 1993. A
frozen core barrel sampler developed at CRREL was used to remove frozen
specimens in March 1993. The frozen core barrel sampler is designed to remove a
soil specimen while the soil is frozen. Removing specimens while frozen prevents
disturbance and preserves the soil and ice structure (Benson et al. 1994). Flexible-
wall permeameters were used to measure the hydraulic conductivity of all specimens
removed from the COLDICE test pads (Figs. 4.4 and 4.5).
The large block specimens were removed using a procedure developed at the
University of Wisconsin-Madison. Othman et al. (1994) describe the procedure in
detail. The following is a brief summary of the sampling procedure: :
1. A location for sampling is chosen and marked and the surface is
cleaned and leveled. For this project, cleaning consisted of removing
0.1 to 0.3 m of gravel or sand (and underlying geosynthetics)
overlying the clay.
2. Trenches are excavated around the marked area using a shovel. The
specimen is trimmed from the remaining block of soil. '
3. A PVC ring (inside diameter = 0.40 m, height =,0.30 m) is placed on
the surface of the block of soil. The ring has a vertical cut that is used
to expand the ring during removal of the specimen in the laboratory.
Excess soil around the ring is gently trimmed away and the ring is
slowly pushed by hand over the block of soil in a manner similar to
trimming a specimen for consolidation testing (Fig. 4.6). Trim!ming of
excess soil is continued until the ring is pushed to full depth (-0.30
m). !
i
4. The specimen is separated from the underlying soil by pressing a flat
spade into the underlying soil around the perimeter of the ring or by
31
-------
.Vent Port
Top Plate
Acrylic Cylinder «». =
Porous Disks =
Base Pedestal
Base Drain
Top Drain
c- 5 _„ Flexible Tube
,"O" Ring Seal
.Latex Membrane
Ring Seal
Hom
Base Drain
Top Drain
Cell Drain
Figure 4.4. Flexible-wall perm'eafrfeter.
Figure 4.5. Large flexible-wall permeameter used for measuring hydraulic
conductivity of block specimens. ;
32
-------
Figure 4.6.
Trimming of block specimen (a) and separating block specimen from
underlying soil using a wire saw (b). ;
33
-------
using a wire saw (Fig 4.6): Once the specimen is separated'from the
underlying soil, excess soil is trimmed away and it is placed on a
pallet.
5. The specimen is sealed with two layers of plastic wrap to' prevent
desiccation during transport to Madison, where it is placed inja humid
room until testing. :
Block specimens were taken along one continuous profile at three depths (0-0.3
m, 0.3-0.6 m, and 0.6-0.9 m) from the 0.9 m-thick test pads (PV-3 and VT-4). Three
specimens were removed from the test pad before winter (December 11992) and four
specimens were removed after one winter (June 1993). A second specimen was
obtained in June 1993 from a depth of 0-0.3 m to obtain an additional measurement of
the surficial hydraulic conductivity. ;
Before being placed in permeameters, the PVC ring was removed from the
specimens and excess or disturbed soil was trimmed away until the specimens had a
diameter of 0.3 m and a height of 0.15 to 0.20 m. The block specimens were placed in
large flexible-wall permeameters and permeated at an average effective stress of 10.5
kPa, hydraulic gradients of 4 to 5, and a backpressure of 413 kPa. Tap water from
Madison, Wisconsin was used as the permeant. The specimens were permeated until
the hydraulic conductivity measurement was steady (no upward or downward trend
over time) and the ratio of outflow to inflow was between 0.75 and 1.25. ;
Specimens were also removed from the COLDICE test pads ijising thin-wall
sampling tubes (diameter = 71 mm). Specimens were removed from the test pads in
June 1993, after on.e winter of exposure. The sampling tubes were seajed in the field
and shipped to the University of Wisconsin-Madison. In the laboratory, ihe specimens
were extruded from the sampling tubes, sealed in plastic wrap, and stored in a high
humidity room prior to hydraulic conductivity testing. ]
The specimens collected in thin-wall tubes were permeated in flexible-wall
permeameters at an average effective stress of 7 kPa and hydraulic gradients ranging
from 2 to 5. A backpressure of 280 kPa was used. Tap water from Madison,
Wisconsin was used as the permeant. Specimens were permeated until hydraulic
conductivity was steady and the ratio of outflow to inflow was between 0.75 and 1.25
The core barrel sampler developed at CRREL was used to remove specimens
from the COLDICE test pads in March 1993 while the soil was frozen (Erickson et al.
.' ' i
34 !
-------
1994, Benson et a!. .1994). The core barrel sampler removes a specimen having a
diameter of 71 mm and length up to 0.9 m (Fig 4.7). A power auger was used to slowly
advance the core barrel so as not to melt the frozen soil. The barrel was then retracted
and the frozen specimen was removed from the barrel. The specimens were then
wrapped in plastic and packaged in an insulated box with ice packs for overnight
shipping to CRREL. All of the specimens obtained with the core barrel were removed
by personnel from CH2M Hill, Inc. or CRREL as part of the COLDIGE project.
Storage and placement of the core barrel specimens in flexible-wall
perrneameters was conducted in a cold room at CRREL by CRREL staff (Chamberlain
et al. 1995). The frozen core specimens were allowed to thaw and consolidate at an
average effective stress of 7 kPa. Hydraulic conductivity tests were then performed at
the same average effective stress using hydraulic gradients ranging from 2 to 5. A
backpressure of 380 kPa was used and tap water from Hanover, New Hampshire was
used as the permeant (Chamberlain et al. 1995). '
4.1.2 Laboratory Methods
4.1.2.1 Hydraulic Conductivity-Water Content Relationships \
Specimens of the Parkview and Valley Trail clays that were |compacted to
determine compaction curves corresponding to standard and modified Proctor efforts
(Sec. 3.1.2) were also used to determine the corresponding hydraulic conductivity-
water content relationships. After compaction, the specimens were extruded from .the
compaction molds, sealed in plastic wrap to prevent desiccation, and stored until
permeation. .!
The specimens were placed in flexible-wall perrneameters for Saturation and
measurement of hydraulic conductivity. The falling head-rising tailwater method was
used (ASTM D 5084-Method D). An average effective stress of 7 kPa, hydraulic
gradient of 12, and backpressure of 128 kPa were used. Tap water from Madison,
Wisconsin was used as the permeant. ! .
Each test was terminated when the hydraulic conductivity was steady (change <
± 25% and no increasing or decreasing trend) and the ratio of outflow1 to inflow was
between 0.75 and 1.25 for four consecutive readings. The hydraulic conductivity was
reported as the arithmetic mean of the last four measurements. j
35
-------
4.1.2.2 Standard Freeze-Thaw Tests ;
Three specimens each of Parkview clay and Valley Trail clay were compacted
at standard Proctor effort at water contents similar to the water content used in
construction of the COLDIGE test pads. These water contents also yielded the lowest
hydraulic conductivities determined in Sec 4.1.2.1. Each specimen was extruded from
its compaction mold, sealed in plastic wrap to prevent desiccation, and stored until
permeation. \ '
The clay specimens were placed in flexible-wall permeameters to determine
their initial hydraulic conductivity (before exposure to freeze-thaw). Test conditions
identical to those described in Sec. 4.1.2.1 were employed. After the initial hydraulic
conductivity tests were complete, the specimens were carefully removed from the
permeameters. They were then sealed.in plastic wrap and duct tkpe to prevent
desiccation and placed in a laboratory freezer for cooling. Extreme Icare was used
when handling the specimens to avoid disturbing the soil structure.
The free-standing procedure described by Othman et al. (1994) was used to
freeze the specimens in a closed system (no external supply of water). They were
cooled to an ultimate temperature of -20°C at a freezing rate of approximately 52
mm/hr. The free-standing procedure results in three-dimensional freezing as opposed
to one-dimensional freezing which occurs in the field. Othman and Benson (1993b)
and Zimmie and LaPlante (1990) have shown, however, that the dimensionality of
freezing does not affect the change in hydraulic conductivity, even though it does
result in somewhat different ice and crack structures. ;
t
The specimens were left in the freezer for at least 24 hours, at which point they
were removed and allowed to thaw at room temperature (25°C). After 24 hours of
thawing, the specimens were placed back in the freezer. Data collected from control
specimens instrumented with thermocouples showed that the 24 hour period was
more than adequate to ensure that complete freezing or thawing occurred (Fig. 4,8).
This procedure was repeated until the desired number of freeze-th^w cycles was
attained. I
The hydraulic conductivity of each specimen was measured Jn flexible-wall
permeameters after 1, 3, and 5 freeze-thaw cycles. To minimize disturbance, frozen
specimens were removed from the freezer and placed in flexible-wall permeameters to
thaw. The cell pressure and hydraulic gradient were applied immediately. The falling
head-constant tailwater method for measuring hydraulic conductivity was used (ASTM
36
-------
• -;-;, -r-Sr
^ ' - * -iS
Figure 4.7. CRREL frozen soil core barrel sampler.
£
£
3
5
0)
Q.
1
\J\J
i
20
10
0
-10
-20
-30
I 1 1 1 [ r— T 1 1 1 1 1 r— i j 1 1 1 1 1 1 1 1 1
1 O TopTC
3 D Middle TC
ft o Bottom TC
iS . Specimen — ^' /&
-8 _ /
C3 ' ^^^
~ v g H Middle ~^
". W S Bottom ^.
8
-
2.
,
_
-
-
"
-
7 . Thermocouple ~
: e Locations :
S3 -
1 1 1 « 1 1 1 1 1 t 1 1 1 1 1 1 1 f 1 1 1 1 1 1
1
1
f
i
i
i
!
0 5 10 15 20 25 !
Time (hours) i
Figure 4.8. Temperature vs. time for Valley Trail clay during three-dimensional
freezing. -i
37
-------
[
05084-Method C). An average effective stress of 10 kPa and hydraulic gradient of 12
was used. Termination criteria identical to those described in Sec. 4.1.2.1 were used.
4.1.2.3 One-Dimensional Freeze-Thaw Tests ]
Compacted specimens of Parkview and Valley Trail clay were subjected to one-
dimensional freeze-thaw to replicate freezing processes that occur in the field. The
hydraulic conductivity of these specimens was then measured to; compare with
hydraulic conductivities of specimens removed from the COLDICE test pads. The
compacted specimens were subjected to the same number of freeze-thaw cycles
observed in the COLDICE test pads. !
i
Three specimens each of Parkview and Valley Trail clay were; compacted at
standard Proctor effort at water contents similar to the water consents used for
construction of the COLDICE test pads. The specimens were removed from their
compaction molds and sealed in plastic wrap. >
The specimens were instrumented with thermocouples at the top, middle,
bottom, and along their side. They were then wrapped in 0.1 m thick ft-11 fiberglass
building insulation and placed on a sheet of expanded polystyrene. A circular heating
element (diameter = 0.15 m) was placed beneath the polystyrene sheet.; The top of the
specimen was left exposed to the ambient temperature of the freezer (-20°C). Figure
4.9 is a diagram of the one-dimensional freezing set-up; a similar method to one-
dimensionally freeze compacted specimens was used by Othmah and Benson
(1993b). Figure 4.10 contains photographs of the equipment used to induce and
monitor one-dimensional freezing. <
Temperatures within the specimens were recorded while freezing (Fig. 4.11).
The specimens were removed from the freezer when no further decrease in
temperature was occurring. The specimens were allowed to thaw at room temperature
(25°C) for at least 24 hours.
Three specimens of each clay were subjected to three different; freezing rates
(one freezing rate per specimen). The freezing rates were varied by changing the
thickness of the polystyrene between the specimen and heating ejement. After
freezing and thawing, the hydraulic conductivity of each specimen was measured in a
flexible-wall permeameter. An average effective stress of 10 kPaiand hydraulic
gradient of 12 was used. No backpressure was applied. Tests were terminated when
the hydraulic conductivity was steady and the ratio of outflow to inflow was between
38
-------
Fiberglass Insulation
Compacted
Clay
Specimen
\
/.•Type-T
^•^Thermocouple
fLocations
//Expanded Polystyrene Sheets-^
'/ (variable thickness) J
/
0.15 m-Diameter
Circular Heating Element
Figure 4.9. One-dimensional freezing set-up.
Figure 4.10. Freezer and data acquisition system used for one-dimensional freezing
of compacted clay specimens.
39
-------
0.75 and 1.25 for four consecutive readings. The hydraulic conductivity was reported
as the arithmetic mean of the last four hydraulic conductivity measurements.
4.2 SAND-BENTONITE MIXTURE ;
4.2.1 Field Methods I
4.2.1.1 COLDICE Test Pads \
A test pad consisting of a sand-bentonite mixture was also constructed in the
COLDICE project, adjacent to the four clay test pads described in Sec. 4.1.1.1. The
sand-bentonite test pad, labeled SB-1, was 9 m by 21 m and had a thickness of 0.6 m
(Erickson et al. 1994) (Fig. 4.2). It was constructed with similar procedures as the test
pads constructed with clay, except that it was compacted using a smooth vibrating
wheel compactor. The test pad was covered with a 0.15 mm-thick polyethylene sheet
and 0.1 m of sand to minimize desiccation. Construction of the sand-bentonite test
pad was completed in October 1992. All of the measurements of'molding water
content and dry unit weight made during construction of the test pad felj wet of the line
of optimums (Fig. 4.12). Like the clay test pads, the sand-bentonite test pad was
instrumented with thermistors to record temperature at different depths. •
i
4.2.1.2 In Situ Box Infiltrometers \
A box infiltrometer was also installed in the sand-bentonite test pad to measure
hydraulic conductivity in the field (Fig 4.3). The infiltrometer was the same
as those described in Sec. 4.1.1.2, and was installed using similar methods (Erickson
et al. 1994). j
i
4.2.1.3 Laboratory Assessment of Field-Scale Hydraulic Conductivity \
Hydraulic conductivity tests were performed in the laboratory'on specimens
removed from the .sand-bentonite test pad to assess the field-scale hydraulic
conductivity. Specimens were removed from the field by three methods: as large
blocks, using thin-wall sampling tubes (Shelby tubes), and as frozen cores. Sampling
of the sand-bentonite test pad was performed concurrently with sampling from the clay
test pads and similar sampling methods were employed. The sampling methods and
schedule are described in Sec. 4.1.1.3. Flexible-wall .permeameters! were used to
measure the hydraulic conductivity of specimens removed from the sand-bentonite test
pad. ;
40
-------
O
o
£
1
CD
CL
E
T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 r-
Specimen
. Middle
'.—• Side
- Bottom
_Thermocouple
Locations
X:-
•J.9P
— — -Side
_j i i •
10
20 30
Time (hours)
,40
50
Figure 4.11. Temperature vs. time for one-dimensional freezing of Parkview clay in
the laboratory. , :
.g>
I
• &•
Q
20
19
18
16
O Standard
D Modified
+ Field
b
A
\
8 12 16 20
Molding Water Content (%)
24
Figure 4.12. Compaction curve and field compaction data for sand-ben' tonite mixture.
41
-------
Three block specimens were removed from the sand-bentohite test pad in
December 1992 (before freezing) and in June 1993 (after freezing);. Because the
sand-bentonite pad was only 0.6 m thick, only two specimens were removed along
one continuous vertical profile (0-0.3 m and 0.3-0.6 m). The third; specimen was
removed from a location adjacent to the first two specimens at an intermediate depth of
0.15-0,45 m. . • |
Before being placed in permeameters, the PVC ring was removed from the
specimens and excess or disturbed soil was trimmed away until the specimen had a
diameter of 0.3 m and a height of 0.15 to 0.20 m. The block specimens were
permeated at an average effective stress of 10.5 kPa, hydraulic gradients of 4 to 5, and
using a backpressure of 413 kPa. One specimen (before winter, 0.15-0.45 m) was
tested at a higher hydraulic gradient (8.0) to reduce the time required to reach
equilibrium. The hydraulic gradient was increased by reducing the effluent pore water
pressure to 406 kPa. Increasing the hydraulic gradient also increased the average
effective stress to 14 kPa. Tap water from Madison, Wisconsin was used, as the
permeant. The specimens were permeated until the hydraulic conductivity was steady
(no upward or downward trend over time) and the ratio of outflow1 to inflow was
between 0.75 and 1.25. The latter criterion was not met by the specimen collected
before winter from a depth of 0.15-0.45 m, which was permeated for 55 days without
reaching equal inflow and outflow. ,
Specimens were also removed from the sand-bentonite test pad with thin-wall
sampling tubes (Shelby tubes) in June 1993, after one winter of exposure. The
sampling tubes were sealed in the field and shipped to the University of Wisconsin-
Madison. The specimens were extruded from the sampling tubes, sealed in plastic
wrap, and stored in a high humidity room to await hydraulic conductivity'testing.
The specimens removed in thin-wall tubes were permeated in flexible-wall
permeameters at an average effective stress of 28 kPa using a hydraulic gradient of
25. No backpressure was used. Tap water from Madison, Wisconsin was used as the
permeant. The specimens were permeated until hydraulic conductivity was steady
and the ratio of outflow to inflow was between 0.75 and 1.25 ;
Specimens were also removed from the sand-bentonite test pad in March 1993
while the soil was frozen. Personnel from CH2M Hill, Inc. and CRREL collected the
specimens using the core barrel described in Sec. 4.1.1.3 (Erickson et al. 1994).
Storage and hydraulic conductivity testing were performed at CRREL. Hydraulic
conductivity of the frozen cores was measured in flexible-wall permeameters. The
' ' 42" ,
-------
frozen core specimens were allowed to thaw and consolidate under an average
effective stress of 7 kPa. Hydraulic conductivity tests were then performed at that
average effective stress using hydraulic gradients ranging from 2 to 5. A backpressure
of .380 kPa was used. Tap water from Hanover, New Hampshire was used as the
permeant.
i
4.2.2 Laboratory Methods i
4.2.2.1 Hydraulic Conductivity-Water Content Relationship ;
Specimens of sand-bentonite that were compacted to determine compaction
curves corresponding to standard and modified Proctor efforts (Sec 3.2.1.2) were also
used to determine the hydraulic conductivity-water content relationships. The
specimens were extruded from the compaction molds, sealed in pla'stic wrap, and
stored until permeation. | '
The specimens were placed in flexible-wall permeameters for saturation and to
measure their hydraulic conductivity, which was measured while the specimens were
saturating. The falling head-rising tailwater method was used. An average effective
stress of 21 kPa and hydraulic gradient of 30 were used. A backpressure of 345 kPa
was applied. Tap water from Madison, Wisconsin was used as the permeant fluid.
Each test was terminated when the hydraulic conductivity was steady (change <
±25% and no increasing or decreasing trend) and the ratio of outflow to inflow was
between 0.75 and 1.25 for four consecutive readings or when the'test time had
reached 30 days. For specimens which had reached steady conditions, the hydraulic
conductivity was reported as the arithmetic mean of the last four measurements. For
specimens which did not reach steady conditions in 30 days, the hydraulic conductivity
was reported as the last measurement. For all of the specimens which had not
reached steady conditions in 30 days, the hydraulic conductivity was steady and the
ratio of outflow to inflow was approaching 1.0.
4.2.2.2 Standard Freeze-Thaw Tests \
Three specimens of the sand-bentonite mixture were compacted at a water
content similar to the water content used for construction of the sand-bentonite test pad
(Sec. 4.2.1.1). Each specimen was extruded from the compaction mold and sealed in
plastic wrap to prevent desiccation and stored until permeation. The 4and-bentonite
specimens were placed in flexible-wall permeameters to determine their initial
43
-------
hydraulic conductivity. Conditions and termination criteria identical to those described
in Sec. 4.2.2.1 were employed. ;
After the initial hydraulic conductivity tests were complete, the specimens were
carefully removed from the permeameters. Extreme care was used to avoid disturbing
the structure of the specimens. They were then sealed in plastic wrap to prevent
desiccation and placed in a laboratory freezer for cooling. The free-standing
procedure was used to freeze the sand-bentonite specimens (Sec.:4.1.2.2). The
specimens were left in the freezer for at least 24 hours, at which point they were
removed and allowed to thaw at room temperature (25°C). After 24 hours of thawing,
the specimens were placed back in the freezer. This procedure was repeated until the
desired number of freeze-thaw cycles was attained. i
The hydraulic conductivity of each specimen was measured after 1, 3, and 5
freeze-thaw cycles. To minimize disturbance, the frozen specimens were removed
from the freezer and directly placed in flexible-wall permeameters forkhawing. The
cell pressure and hydraulic gradient were applied immediately. An average effective
stress of 21 kPa and hydraulic gradient of 30 were applied. A backpressure of 345
kPa was used. j
i
I
4.3 GEOSYNTHETIC CLAY LINERS (GCLs)
4.3.1 Field Methods
4.3.1.1 COLDICE Test Ponds and Pans-Field Measurement of Hydraulic Conductivity
Large-scale GCL test ponds were constructed in October 1992 and large-scale
test pans were constructed in September 1993 by personnel from CH2lv1 Hill, Inc. as
part of the COLDICE project (Fig. 4.2). The ponds and pans were used;to assess the
impact that freeze-thaw had on the hydraulic conductivity of GCLs (Erickson et al.
1994). The test pans were constructed because of difficulties encountered the
previous year (1992) with the seepage collection system in the GC|L test ponds
(Erickson et al. 1994). Consequently, the only successful hydraulic conductivity tests
were performed using the GCL test pans. Tests were performed on! three GCLs:
Bentomat®, Claymax®, and Gundseal®:
Nine rectangular test pans were constructed in three groups (one group for
each GCL). Each group of test pans contained one large test pan (surface area = 1.8
m2) and two smaller test pans (surface area = 0.7 m2). The large test pans and one of
the two small test pans contained GCLs with seams that were installed ih accordance
44
-------
with the manufacturers' specifications. The third pan in each group contained a
seamless GCL specimen (Erickson et al. 1994).
The large test pans were constructed by welding pieces of HOPE
geomembrane plate stock together. The smaller test pans were manufactured storage
bins constructed of HOPE. Each test pan contained a seepage collection system with
drains. A GCL specimen was placed over a thin layer of pea gravel to act as a support
and drain for the GCL The GCL was covered with a 0.25 m-thick layer of pea gravel.
The test pans were then surrounded by pea gravel to the same level of the gravel
within the pans (Fig, 4.13) (Erickson et al. 1994).
Water was initially added to a depth of 30 mm in the test pahs to allow the
bentonite to hydrate. After one week, the water level was increased to 0.25 m. The
bentonite was allowed to hydrate in this condition for one month. After one month,
seepage data were recorded. The water level was kept constant during! the duration of
the tests, and the water was not drained prior to winter. The hydraulic gradient ranged
from 5 to 15 and averaged 10. Hydraulic conductivity of the GCLs was determined by
measuring outflow. Measurements of before-winter hydraulic conductivity were made
through December 1993. Measurements after winter began in April 1994 (Erickson et
al 1994). • | . .
4.3.1.2 Laboratory Assessment of Field-Scale Hydraulic Conductivity \
During decommissioning of the COLDICE GCL test ponds and pans in June
1994, four specimens of Bentomat® and Claymax® (two specimens'of each GCL)
were removed from the test ponds by personnel from the University! of Wisconsin-
Madison. The GCLs in the test ponds had been exposed to two winters and two
freeze-thaw cycles. Specimens, approximately 0.8 m by 0.8 m, were cut using a razor
knife, placed on stiff sheets of 10 mm-thick HDPE plate stock, and sealed with plastic
wrap for shipping. Extreme care was taken to prevent disturbance of the hydrated
bentonite when cutting, sealing, and transporting the specimens. The specimens were
stored prior to permeation in a high humidity room at the University 'of Wisconsin-
Madison. ;
Large flexible-wall permeameters were used to measure ihe hydraulic
conductivity of the GCL specimens. The protocol described in Geosynthetic Research
Institute (GRI) test method GCL-2 was employed, which is an adaptation of ASTM D
5084 specifically for use with GCLs. Circular specimens were cut from the 0.8 m by
0.8 m sections with a razor knife to a diameter of 0.45 m. A bentonite paste was
i
45 ;
-------
applied to the edge of the specimens to prevent sidewatl leakage between the
specimen and the membrane. The falling head-constant tailwater method was
employed for permeation. An effective stress of 28 kPa and hydraulic; gradient of 75
were used. No backpressure was applied. The hydraulic conductivity tests were run
until the hydraulic conductivity was steady and outflow equaled inflow for four
consecutive measurements. i .
The two Bentomat® specimens had very high flow rates. Dye was injected into
the influent and the permeameter was then disassembled to attempt to; locate the high
seepage zones. It was found that sidewall leakage between the specimen and the
membrane was the cause of the high flow rates for both Bentomat® specimens. The
Bentomat® specimens were removed from the permeameter and trimmed to a
diameter of 0.3 m and placed in another flexible wall permeameter. Bentonite paste
was applied around the edges of the specimen to alleviate the sidewall leakage
problem. Similar testing conditions and termination criteria were used when re-testing
the Bentomat® specimens. Sidewall leakage continued to be a problem for one of the
smaller Bentomat® specimens, and additional measures were not successful in
correcting the problem. '
4.3.2 Laboratory Methods ! • .
Laboratory tests were conducted on Bentofix®, Bentomat®, and Claymax® (Fig.
3.7). Tests were not conducted on Gundseal®, because freeze-thaw is unlikely to
affect the hydraulic performance of the geomembrane component of the Gundseal®
GCL. Also, the GCL Bentofix® was only used in the laboratory freeze-thaw tests and
was not evaluated in the field. |
4.3.2:7 Laboratory Measurement of Hydraulic Conductivity \
Specimens for hydraulic conductivity testing were selected by unrolling the
GCLs on the laboratory floor and locating a point away from the edge of the roll where
the specimens were likely to have uniform bentonite content. A piece of plywood was
placed underneath the GCL and a razor knife was used to carefully cut out circular
specimens (diameter = 0.15 m). Efforts were made to keep loose bentonite in the
GCLs from spilling out of the edges. This procedure was used to make three
specimens each of Bentofix® and Bentomat® and four specimens of Claymax® (Fig.
4.14). The specimens were weighed and their diameter and thickness were measured
46
-------
1cm Pea Gravel
\
fVC Covor Strip
4cm Minus Stone ^-Vent Pipe
\
4cm Minus"
Stone
Bentonlte Edge
Seal (Paste)
Bentonlte Paste
;Seepage Drain
Side Leakage Drain
20 ml PVC Membrane
PVC Seepage
Collection
Control Varve
Figure 4.13. Schematic of GCL test pan used at the COLDICE field site (from Erickson
etal. 1994).
Figure 4.14. Photograph of trimmed Bentomat® specimen for laboratory testing
47
-------
using calipers. A bentonite paste was then applied to the edges of each specimen to
prevent short circuiting during permeation. :
The specimens were placed in flexible-wall permeameters for saturation and to
define their initial hydraulic conductivity. The hydraulic conductivity was measured
while the specimens were saturating. The protocol described in GRI test method GCL-
2 was employed. An average effective stress of 7 kPa and hydraulic gradient of 75
were applied, and no backpressure was used. Tap water from Madison, Wisconsin
was used as the permeant. !
The tests were terminated when the hydraulic conductivity was steady (change
smaller than ± 25% and no increasing or decreasing trend) and the ratio of outflow to
inflow was between 0.75 and 1.25 for four consecutive readings. > The hydraulic
conductivity was reported as the arithmetic mean of the last four measurements.
4.3.2.2 Standard Freeze-Thaw Tests
After the initial hydraulic conductivity tests were complete, the GCL specimens
were carefully removed from the permeameters by sliding them onto sheets of
.expanded polystyrene. Extreme care was used to avoid disturbing the structure of the
hydrated bentonite. Calipers were used to measure their dimensions. They were then
sealed in plastic freezer bags to prevent desiccation and placed in a. laboratory freezer
for cooling. An identical procedure was used after each freeze-thaw-pej-meate cycle.
The free-standing procedure (Othman et al. 1994) was used' to freeze the
specimens in a closed system. They were cooled to an ultimate temperature of -20°C
at a freezing rate of approximately 12 mm/hr. This procedure results in three-
dimensional freezing as opposed to one-dimensional freezing which' occurs in the
field. Nevertheless, because the GCL specimens are thin and wide relative to
specimens of compacted clay, freezing was .approximately one-dim'ensional (Fig
4.15). ;
The specimens were left in the freezer for at least 24 hours, at which point they
were removed and allowed to thaw at room temperature (25°C). After 24 hours of
thawing, the specimens were placed back in the freezer. Data collected from control
specimens instrumented with thermocouples (Fig. 4.15) showed that !24 hours was
more than adequate to ensure that complete freezing or thawing occurred. This
procedure was repeated until the desired number of freeze-thaw cycles was attained.
The hydraulic conductivity of each specimen was measured after 1, 3, 5, and 20
freeze-thaw cycles. To minimize disturbance, frozen specimens were removed from
48 :
-------
the freezer and directly placed into flexible-wall permeameters in the frpzen state. The
cell pressure and hydraulic gradient were applied immediately. Hydraulic conductivity
was measured during and subsequent to thawing. Test parameters and termination
criteria identical to those described in Sec. 4.3.2.1 were used. j
4.4 PAPER MILL SLUDGES
4.4.1 Field Methods
4.4.1.1 Compaction of Pipe Specimens
Large specimens of compacted paper mill sludge were constructed by
compacting paper mill sludge in a poly vinyl chloride (PVC) pipe (diameter = 0.35 m,
length— 0..6 m) (Fig. 4.16). Two specimens were constructed for each sludge. One
specimen of each sludge was used to determine the initial hydraulic conductivity
(herein referred to as the "control" specimen). The other specimen was instrumented
with thermocouples and buried in the ground outside the Environmental Geotechnics
Laboratory at the University of Wisconsin-Madison (herein referred tb as the "field"
specimen). The field specimens were.buried in the ground in early December 1993
and allowed to freeze one-dimensionally. The specimens were removed from the
ground in March 1994 and were permeated to determine the hydraulic conductivity
after exposure to freeze-thaw. Benson and Othman (1993) conducted similar small-
scale field tests on a pipe specimen of compacted clay. ;
A PVC plate was used at the bottom of the pipe to confine the specimen. A
georiet and non-woven geotextile were placed at the bottom of the specimen prior to
compaction to provide for filtration and drainage during permeation, the specimens
were compacted directly in the pipe using a compaction energy equal to standard
Proctor energy and molding water contents which yield the lowest hydraulic
conductivities for the three sludges (see Sec. 4.4.2.1). The specimens were
compacted in six 0.1-m-thick lifts. Compaction was performed by dropping a 11.8 kg
cylindrical weight with a diameter of 100 mm 120 times on each lift from, a height of 0.3
m (Fig. 4.17). A flexible PVC geomembrane was placed on top of the compacted
paper mill sludge in the field specimens and sealed to the inside of the PVC pipe to
prevent desiccation. ;
4.4.1.2 Instrumentation and Burial of Pipe Specimens '•
Thermocouples were placed in the field specimens at the center of the pipe
between each lift so the temperature within the specimen could be pleasured with
49 i
-------
0
o
2
1
8.
O)
H
i
1 ^
I O
10
5
0
-5
-10
-15
.on
L , i . , i . . I . - i • • • i • • • i • • • i • • • i • • A-
1 Top. Middle '
KJ \ / Qj
V6 . VaLrk,;... ..;••:.-,;: .. .j'.^' , i^jr. :
i-O .x /
ZB Bottom Side A :
^ o :
' SB 8 ° 1
- S 8 a ' g ^^ ^oo -:
A A ' ~.
@-
:
O Top :
D Middle ;
O Bottom as
A Side Q A :
' , , , I , i , 1 i i i 1 i . i 1 i r I V t . J _i — 1 — i — i — 1 I I i 1 —
0 2 4 6 8 10 12 14 - 16
Time (hours) •
Figure 4.15. Temperature vs. time for Bentomat® GCL frozen three-dimensionally.
Figure 4.16. Small-scale field pipe specimens.
50
-------
Figure 4.17. Compaction of paper mill sludge pipe specimen.
51
-------
depth. Teflon® coated Type-T thermocouples were used. Five thermocouples were
placed in each pipe specimen. Temperatures within the pipe specimens were
recorded by a Campbell Scientific CR10 Datalogger every hour:between rnid-
December 1993 and late March 1994. Air temperature and relative hurfiidity were also
recorded every hour by the datalogger. A graph of air temperature vs.i time is located
in Appendix B. Figure 4.18 is a photograph of the data acquisition set-up.
The field specimens were buried in the ground in early December 1993, before
any ground freezing had occurred. A large trench was excavated and the three field
specimens were placed next to each other in the trench (Fig 4.19). The trench was
then backfilled level with the top of the compacted sludge in the pipe specimens. The
top of the sludge, overlain by the PVC geomembrane was exposed to the ambient air
temperature. ;
i
4.4.1.3 Permeation of Pipe Specimens !
The pipe specimens were designed to be used as la;rge rigid-wall
permeameters. The,control specimens were permeated while the fibld specimens
were buried in the ground, whereas the field specimens were permeated after
exposure to one winter. Brass fittings were placed in the top and bottom plates of the
pipe specimens to allow for inflow and outflow of water. The constant headwater-
constant tailwater method was used to measure the hydraulic conductivity of the pipe
specimens in accordance with methods described in the ASTM draft standard for rigid-
wall hydraulic conductivity tests. Mariotte bottles were used to apply a constant head.
Tests were performed at a hydraulic gradient of 3. •
The tests were terminated when the hydraulic conductivity was steady and the
ratio of outflow to inflow was between 0.75 and 1.25 for four consecutive readings.
The hydraulic conductivity of the specimens is reported as the arithmetic mean of the
last four hydraulic conductivity measurements.
High initial values of the outflow to inflow ratio were observed for all pipe
specimens during testing. It was determined that excessive gas build-jup was taking
place within the specimens. This gas build up was preventing inflow and increasing
outflow as a result of increased pore pressures in the specimens. To correct this
problem, venting pipes were installed in the top plates. After the venting pipes were
installed, accurate readings of outflow and inflow were possible. i
52
-------
Figure 4.18. Data acquisition set-up for buried pipe specimens.
Figure 4.19. Burial of pipe specimens in the ground outside the Environmental
Geotechnics Laboratory at the University of Wisconsin-Madison.
53
-------
i ft j| (ft ,
4.4.1.4 Measurement of Hydraulic Conductivity in Flexible-Wall Permeameters
Following permeation, the pipe specimens were sliced into horizontal sections
that were subsequently permeated to investigate how depth influenced hydraulic
conductivity. Benson and Othman (1993) showed that the hydraulic conductivity of the
compacted clay varied with depth for their pipe specimen. \
A circular saw was used to slice the pipe into horizontal sections. The sections
were then separated from each other by slicing through the compacted paper mill
sludge with a carpenter's saw. Vertical cuts were made in the pipe around the sludge
sections and the remaining pieces of PVC were carefully removed to prevent
disturbance (Fig 4.20). The sections of compacted paper mill sludge were trimmed to
a diameter of 0.3 m and heights ranging from 0.1 to 0.2 m. The specimens were then
placed in large flexible-wall permeameters (Fig. 4.5). An average effective stress of 18
kPa was applied to the specimens. Hydraulic gradients ranged from; 8 to 16. No
backpressure was used. Termination criteria described in Section^ 4.4.1.3 were
followed.
4.4.2: Laboratory Methods
4.4.2.1 Hydraulic Conductivity-Water Content Relationships j
Specimens that were compacted to determine the compaction qurves for the
three paper mill sludges (Sec. 3.3.2) were used to determine the hydraulic
conductivity-water content relationships. The specimens were placed in rigid-wall
compaction mold permeameters for saturation and to measure their hydraulic
conductivity. The rigid-wall permeameters contained a double ring in the base plate to
provide a check for sidewall leakage. The constant head-constant tailwater method
was used in accordance with methods described in the ASTM draft standard for rigid
wall hydraulic conductivity tests. A lead weight was placed on the! specimen to
simulate an overburden stress of 7 kPa during hydraulic conductivity testing.
Hydraulic gradients ranging from 6 to 12 were applied, and tap water from Madison,
Wisconsin was used as the permeant. ;
Each test was terminated when the hydraulic conductivity was steady (change <
±25% and no increasing or decreasing trend) and the ratio of outflow jto inflow was
between 0.75 and 1.25 for four consecutive readings. The hydraulic conductivity was
reported as the arithmetic mean of the last four measurements. '
Difficulties were encountered when permeating the paper mill sludge in the
rigid-wail permeameters. Gases produced as a result of biological activity within the
,.,.., 54 !
-------
paper mill sludge blocked inflow and increased outflow as a result of increased pore
pressures within the specimens. A venting tube, similar to those used when
permeating the pipe specimens (Sec. 4.4.1.3), was placed in the top plate of the
perrneameter (Fig 4.21). Use of the venting tube alleviated the problern and reduced
testing time. . ;
i
i
4.4.2.2 Standard Freeze-Thaw Tests . j
Four specimens of sludge A and three specimens of sludges B and C were
compacted at standard Proctor effort at water contents yielding the lowest hydraulic
conductivities determined in Sec 4.4.2.1 (herein referred to as "low-K" water contents)
and at water contents at which they were received at the University! of Wisconsin-
Madison (herein referred to as "as-received" water contents). Each specimen was
compacted, sealed in its compaction mold using plastic wrap to prevent desiccation,
and stored until permeation. \
The sludge specimens were placed in rigid-wall compaction mold
permeameters to determine their initial hydraulic conductivity (before exposure to
freeze-thaw). Test conditions and termination criteria similar to those described in
Sec. 4.4.2.1 were employed for the initial hydraulic conductivity tests. After the initial
hydraulic conductivity tests were complete, the excess water was drained from the
permeameters and the entire perrneameter was placed in the laboratory freezer.
The free-standing procedure (Othman et al. 1994) was used! to freeze the
specimens in a closed system. They were cooled to an ultimate temperature of -20°C.
The specimens were left in the freezer for at least 24 hours, at which point they were
removed and allowed to thaw at room temperature (25°C). After 24 hours of thawing,
the specimens were placed back in the freezer. This procedure was repeated until the
desired number of freeze-thaw cycles was attained. The hydraulic conductivity of each
specimen was measured after-a various number of freeze-thaw cycles (up to 5 cycles).
Sidewall leakage was a problem for the specimens exposed to freeze-thaw.
Volume change of the specimen due to freezing and subsequent thawing resulted in a
gap between the specimen and the wall of the compaction mold. To;eliminate this
problem, the specimens were extruded from their compaction molds and placed in
flexible-wall permeameters for hydraulic conductivity testing. The i falling head-
constant tailwater method for measuring hydraulic conductivity Was used for
specimens which had been subjected to freeze-thaw. An average effeptive stress of
1.0 kPa and hydraulic gradient of 12 were used. No backpressure; was applied.
55 ~ ' !
-------
Figure 4.20. Photograph of sliced sludge specimen for hydraulic conductivity testing
in flexible-wall permeameter. . ;
Figure 4.21. Test set-up for hydraulic conductivity testing of compacted paper mill
, sludge specimens.
56
-------
Hydraulic conductivity was measured subsequent to thawing. Termination criteria
similar to those described in Sec. 4.4.2.1 were used. ;
4.4.2.3 Effective Stress Tests \ •
Tests were performed to determine the relationship between hydraulic
conductivity and effective stress for the three paper mill sludges. One specimen of
each sludge was compacted at standard Proctor effort at the molding water content
that yields the lowest hydraulic conductivity for that sludge (low-K water content, Sec.
4.4.2.2). Hydraulic conductivity tests were performed in flexible-wall permeameters at
effective stresses of 7, 35, 46, and 81 kPa and hydraulic gradients ranging from 6 to 9
were used. Termination criteria similar to those described in Sec. 4.4.2.1 were used.
Following the completion of a test, the effective stress was increased by increasing the
cell pressure in the permeameter and the specimen was allowed to consolidate for
about 7 days after which, the hydraulic gradient was applied. Hydraulic conductivity
was measured by monitoring outflow from the specimens. •
4.4.2.4 Long-Term Hydraulic Conductivity Tests !
Biological activity in compacted paper mill sludges has been reported in other
experiments (NCASI 1989, Maltby and Eppstein 1994) and was observed for the three
sludges tested in this project. Therefore, hydraulic conductivity tests were performed
on two specimens of each sludge to determine if the hydraulic conductivity of the three
sludges is affected by time. !
Specimens were compacted at standard Proctor effort. One specimen of each
sludge was compacted at the water content yielding the lowest hydraulic conductivity
(iow-K water content, Sec. 4.4.2.2) and the other specimen was compapted at the as-
received water content (Sec. 4.4.2.2). Double-ring compaction-mold permeameters
were used to measure hydraulic conductivity of the specimens. A lead weight was
placed on the specimens during permeation to simulate an overburden pressure
equal to 7 kPa. A venting tube was attached to the top plate of theipermeameter,
which was used to allow gases produced by biological activity in the sludge to escape.
A hydraulic gradient of 7 was applied across the specimens and termination criteria
similar to those described in Sec. 4.4.2.1 were used.
57
-------
SECTION 5
\
i
RESULTS
CLAYS !
5.1 FIELD TESTS
5.1,1 Freeze-Thaw Monitoring
Temperatures were monitored by CRREL personnel throughout the winter of
1992-93 in the Parkview and Valley Trail test pads constructed for-the COLDICE
project. Freezing of the test pads began in November 1992 and was steady in mid- to
late December 1992. In mid- to late March 1993, thawing became steady. Complete
thaw of the test pads occurred by the first week in April 1993: The test pads .remained
frozen for about 3.5 months (Erickson et al. 1994). ,
Figures 5.1 and 5.2 show frost zones in the test pads over time. Freezing
records show that once steady freezing was established, freeze-thaw cycling only
occurred in the overburden material and not in the compacted clay. It was determined
that the Parkview and Valley Trail test pads were subjected to one freeze-thaw cycle
during the winter of 1992-93 (Figs. 5.1 and 5.2). The maximum depths of frost
penetration in the Parkview and Valley Trail test pads were 0.70 and 0.55 m,
respectively. An average freezing rate of 0.03 mm/hr was measured for each test pad.
5.1.2: In Situ Box Infiltrometers
In situ box infiltrometers were installed in the two test pads constructed with
Parkview clay to measure the hydraulic conductivity of the test pads after exposure to
freeze-thaw. Installation of the infiltrometers and measurement of hydraulic
conductivity were performed by personnel from CH2M Hill, Inc. No box infiltrometers
were Jnstalled in either of the Valley Trail test pads (Erickson et al. 1994).
Erickson et al. (1994) made attempts to measure the hydraulic conductivity of
the Parkview test pad in spring 1993 by filling the PVC riser pipe in the top of the
infiltrometers with water and measuring the water level within the pipe. Any water
which was added to the riser pipe immediately drained such that no measurement of
the water level could be made. The high flow rates were either the result of increased
hydraulic conductivity due to freeze-thaw or a leak in the infiltrometers. The exact
cause of the high flow rates could not be determined (Erickson et al. 1994).
58
-------
o
20 40 60 80
Time (days since December 31, 1992)
100
Figure 5.1. Frost depth vs. time in the Parkview test pad (from Chamberlain 1994,
personal communication). '
o
c
f
20 40 60 80
Time (days since December 14, 1992)
100
Figure 5.2. Frost depth vs. time in the Valley Trail test pad (from Chamberlain 1994,
personal communication).
59
-------
Sand covering one of the infiltrometers was removed and :the top of the
infiltirometer was cut open by personnel from CH2M Hill, Inc. and the University of
Wisconsin-Madison. Shovels were used to excavate and examine the compacted
clay. Figure 5.3 is a photograph of the soil at the surface of the Parkview test pad from
within one of the box infiltrometers. The soil appeared blocky, containing numerous
horizontal and vertical cracks. This type of structure has been observed in other
experiments examining the effect of freeze-thaw on the hydraulic integrity of
compacted clays at field-scale (Benson and Othman 1993, Erickson et al. 1994,
Benson et al. 1995, Chamberlain et al. 1995). .
i .
5.1.3 Laboratory Assessment of Field-Scale Hydraulic Conductivity
Hydraulic conductivity tests were performed on specimens removed from the
Parkview and Valley Trail test pads as blocks, with thin-wall sampling tubes (Shelby
tubes), and as frozen cores. A summary of the results of the hydraulic conductivity
tests are shown in Fig. 5.4. :
5.1.3.1 Block Specimens ;
Hydraulic conductivity of the block specimens removed in June 1993 from
COLDICE test pads PV-3 and VT-4 was measured in flexible-wall permeameters at the
University of Wisconsin-Madison. Table 5.1 is a summary of; the hydraulic
conductivities for specimens removed from the Parkview test pad. All of the specimens
removed before winter had hydraulic conductivities less than 1 x 10'9 m/s, whereas the
specimens removed from the test pad after winter above the maximum;frost depth had
hydraulic conductivities greater than 1 x 10-7 m/s. The specimen removed after winter
from below the maximum frost depth had a hydraulic conductivity of 2.5 x 10-10 m/s
(Fig. 5.4). . j
Similar results were obtained for block specimens removed from;the Valley Trail
test pad (Table 5.2). The specimens removed from the test pad before winter and the
specimen removed from the test pad after winter below the maximum frost depth all
had hydraulic conductivities less than 1 x 10~9 m/s. The hydraulic conductivities for all
of the specimens removed from the Valley Trail pad after winter above the maximum
frost depth were greater than 1 x 10-7 m/s (Fig 5.4). !
After permeation, the block specimens were cut open to examine their structure.
The specimens were opened by pushing a screwdriver into the soil and prying until
60
-------
i-igure 5.3. Photograph of soil inside box infiltrometerrParkview test pad.
61
-------
H
o>
Q
Hydraulic Conductivity (m/s)
Maximum Frost Depth
A
Block-Before F-T
Block-After F-T
Thin-Wall Tube-After F-T
Frozen Core :
Hydraulic Conductivity (m/s)
ID'10 10'9 ID'8
10'7 10'6
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
-| 1 I I HIM 1 1 I I I IMI 1 1 I II
Maximum Frost Depth
Block-Before F-T
Block-After F-T
Thin-Wall Tube-After F-T
Frozen Core
Frigure 5.4. Hydraulic conductivity vs. depth for specimens removed from the
Parkview (a) and Valley Trail (b) test pads.
62
-------
Table 5.1. Results of hydraulic conductivity tests on block specimens removed from
the Parkview test pad.
Specimen1
PV-Block-Before-1 -
PV-Block-Before-2
PV-Block-Before-3
PV-Block-After-i
PV-Block-After-1-R2
PV-Block-After-2
PV-Block-After-3
Sample Depth
(m)
0-0.3
0.3-0.6
0.6-0.9
0-0.3
0-0.3
0.3-0.6
0.6-0.9
Hydraulic Conductivity
(m/s)
1.9xi10-10
2.2x:iO-10
4.5x!lO-10
2.7 x! 10-6
1.1 x10-6
4.4x!10-7
2.5x'lO-10
Notes: ' ,
1. PV = Parkview clay; Block = Block specimen; Before = Sampled before winter; After = Sampled after
winter; -1, -2, -3 = Specimen number. ;
2. R = Replicate specimen from depth of 0-0.3 m. i
Table 5.2. Results of hydraulic conductivity tests on block specimens removed from
the Valley Trail test pad. ; .
Specimen1
VT-Block-Before-1
VT-Block-Before-2
VT-Block-Before-3
VT-Block-After-1
VT-Block-After-1-R2
VT-Block-After-2
VT-Block-After-3
Sample Depth
(m)
0-0.3
0.3-0.6
0.6-0.9
0-0.3
0-0.3
0.3-0.6
0.6-0.9
i
Hydraulic Conductivity
(m/s)
9.8 xlO-11
1.6x10-1°
I.Sx'IO-10
3.3x!10'7
1.8xi10-7
2.6 Xi 10-7
2.2 x:10-10
Notes: ;
1. VT = Valley Trail clay; Block = Block specimen; Before = Sampled before winter; After = Sampled after
winter; -1, -2, -3 = Specimen number. • ;
2. R = Replicate specimen from depth of 0-0.3 m. i
63
-------
the soil cracked. This method prevents smearing of the soil, which-can mask the
internal structure. i
Figure 5.5 contains photographs of specimens from the Parkview and Valley
Trail test pads removed before exposure to freeze-thaw. The photographs in Fig. 5.5
are characteristic of the soil structure existing in both test pads before winter. The soil
is dense and homogeneous and no cracks are evident. This structure is consistent
with the low hydraulic conductivity that was measured. i
. , • ' i •
In contrast, Fig. 5.6 contains photographs of the internal structure of specimens
removed from a depth of 0-0.3 m in the Parkview and Valley Trail test pads after
exposure to freeze-thaw. The specimens appear blocky and contain numerous
horizontal and vertical cracks. Horizontal cracks were most likely the result of the
formation of ice lenses within the soil. Vertical cracks were most likely caused by
desiccation incurred by migration of water during freezing of the soil. The high
hydraulic conductivity of the specimens is attributed to the existence of these
horizontal and vertical cracks. ;
Figure 5.7 contains photographs of the Parkview and Valley Trail specimens
!
removed from the test pads from a depth of 0.6-0.9 m. These sp'ecimens were
collected from below the maximum depth of frost penetration, and thus were never
frozen. None of the cracks that were present in the shallow specimens exist in these
specimens. The soil is a dense, homogeneous mass that appears similar to the soil as
it existed prior to freezing (Fig 5.5). The absence of cracks and the homogeneity of the
soil structure is consistent with the low hydraulic conductivity of these specimens.
5.1.3.2 Specimens Collected in Thin-Wall Tubes
Hydraulic conductivity was also measured on specimens removed in June 1993
from test pads PV-3 and VT-4 using thin-wall sampling tubes (Shelby tubes). The
hydraulic conductivity tests were performed at the University of Wisconsin-Madison.
Results of the tests are summarized in Fig. 5.4 and Table 5.3. All of :the specimens
removed in thin-wall tubes had hydraulic conductivities less than 1 x 10-p m/s.
Hydraulic conductivities measured on specimens removed using thin-wall
sampling tubes after winter were similar to the hydraulic conductivities measured on
block specimens removed before winter, but much lower than;the hydraulic
conductivities of block specimens removed after winter above the maximum frost
depth. The low hydraulic conductivities of these specimens were most ilikely the result
of disturbance of the specimens during sampling and extrusion which masked the
64
-------
Figure 5.5. Interior of block specimens removed before winter from ithe Parkview
(depth = 0-0.3 m) (a) and Valley Trail (depth = 0.6-0.9 m) (b):test pads.
Figure 5.6. Interior of block specimens removed after winter from a depth of 0-0.3 m
from the Parkview (a) and Valley Trail (b) test pads.
65
-------
Park View
(Bottom)
Figure 5.7. Interior of block specimens removed after winter from a depth of 0.6-0.9
m from the Parkview (a) and Valley Trail (b) test pads.
66
-------
Table 5.3.
Results of hydraulic conductivity tests on thin-wall tube specimens
removed from the COLDICE test pads.
Specimen1
PV-TW-1
PV-TW-2
PV-TW-3
PV-TW-4
PV-TW-5
VT-TW-1
VT-TW-2
VT-TW-3
VT-TW-4
VT-TW-5
Sample Depth
(m)
0.10
0.15
0.25
0.45
0.52
0.07
0.15
0.25
0.45
0.52
Hydraulic Conductivity
(m/s)
1.0X'10-9
1.0x10-9
4.5 xlO-10
1.6x 10'10
1.6 x10-10,
4.0 x10-10
1.5x:10-10
1.5x10-10
1.5xi10-10
1.5 x10-10
Note:
' 1. PV = Parkview; VT = Valley Trail; TW = Thin-wall tube specimen; -1, -2, -3, -4, -5 = Specimen number,
67
-------
effects of freeze-thaw. Figure 5.8 shows photographs of specimens removed from the
Parkview and Valley Trail test pads after exposure to freeze-thaw. [^either specimen
contains a crack network similar to those observed in the block specimens. The
absence of cracks is consistent with the low hydraulic conductivity of these specimens.
I
5.1.3.3 Frozen Core Specimens ;
Hydraulic conductivity tests were performed on specimens removed from the
Parkview and Valley Trail test pads in March 1993 while they were frozen. The tests
were performed at the CRREL laboratory by CRREL personnel. Four specimens were
removed from the Parkview test pad at depths of 0.05, 0.1, 0.15, and 0.2 m and three
were removed from the Valley Trail test pad at depths of 0.15, 0.25, and 0.35 m
(Chamberlain et al. 1995). Results of the hydraulic conductivity tests are summarized
in Fig. 5.4 and Table 5.4.
The hydraulic conductivities of the frozen core specimens are similar to the
. hydraulic conductivities measured on the block specimens, and much; higher than the
hydraulic conductivities measured on specimens removed using thin-wall sampling
tubes having the same diameter (71 mm). •
Figure 5.9 is a photograph of a frozen specimen removed from the Parkview test
pad using the CRREL core barrel. The specimen contains cracks land ice lenses
typically observed in compacted clays subjected to freezing. The presence of cracks is
consistent with the high hydraulic conductivities of these specimens.
5.1.3.4 Structure of Compacted Clay Subjected to Freeze-Thaw \
The increase in hydraulic conductivity for the Parkview and Valley Trail clays
exposed to freeze-thaw is related to the soil structure in the test specimens. Test pits
excavated in June 1993 revealed a structure similar to the structure observed in the
box infiltrometers. The soil appeared blocky and contained numerous orthogonal
cracks (Fig. 5.10). Soil within the test pits could be broken apart into chunks and was
no longer a plastic, homogeneous mass as it was before winter. The horizontal cracks
were spaced approximately 5-10 mm apart near the surface of the test pad and
increased with depth. Vertical cracks were spaced approximately 20-30 mm apart.
The increase in hydraulic conductivity observed in block specimens and frozen cores
(Tables 5.1, 5.2, and 5.4) is a direct result of the existence of these! horizontal and
vertical cracks. ;
68
-------
Valley Trail
Parkview
Figure 5.8. Specimens removed after winter from the Parkview and Valley Trail test
pads using a thin-wall tube (a) and their internal structure (b).
69
-------
Table 5.4. , Results of hydraulic conductivity tests on specimens removed from the
COLDICE test pads with CRREL frozen core barrel (from Chamberlain
1994, personal communication). ;
Specimen1 •
PV-CB-1
PV-CB-2
PV-CB-3
PV-CB-4
VT-CB-1
VT-CB-2
VT-CB-3
Sample Depth
(m)
0.05
0.1
0.15
0.2
0.15
0.25
0.35
Hydraulic Conductivity
(m/s)
8.5x10-8
1.3x!lO-6
1.7xi10-7
5.4x:10-7
8.0x!10-7
1.2x;10-8
8.5xi10-9
Note:: i
1. PV = Parkview; VT = Valley Trail; CB = CRREL core barrel specimen; -1, -2, -3, -4 = ;Specimen number.
70
-------
Figure 5.9.
Specimen removed from Parkview test pad while frozen bsing CRREL
core barrel (from Erickson 1994, personal communication).:
Figure 5.10. Test pit excavated in Valley Trail test pad showing blocky
caused by freeze-thaw.
71
soil structure
-------
5.2 LABORATORY TESTS j
5.2.1 Hydraulic Conductivity-Water Content Relationships !
Hydraulic conductivity was measured on specimens of Parkview and Valley
Trail clay compacted to determine compaction curves corresponding to standard and
modified Proctor effort (Sec. 3.1.2). Results of the hydraulic conductivity tests are
shown in Figs. 5.11 and 5.12 and are summarized in Appendix A. !
The hydraulic conductivity of the specimens decreases as the: molding water
content or compactive effort increases (Figs. 5.11 and 5.12). The lowest hydraulic
conductivities occur at molding water contents slightly wet of optimum molding water
content for both compactive efforts (see Figs. 3.3 and 3.4). These results are typical for
compacted clays (Mitchell et al. 1965, Benson and Daniel 1990, Daniel and Benson
1990). :
.... ... ... : .
5.2.2 Standard Freeze-Thaw Tests
Three specimens of each clay were compacted using standard Proctor effort at
the molding water content at which the Parkview and Valley Trail test pads were
constructed (=16 and 21%, respectively). The specimens were subjected to freeze-
thaw using the free-standing procedure (Sec. 4.1.2.2). The hydraulic conductivity of
each specimen was measured after 0, 1, 3, and 5 freeze-thaw cycles.. Results of the
standard freeze-thaw tests are summarized in Figure 5.13 and Table 5.5.
The hydraulic conductivity of both clays increased when subjected to freeze-
thaw (Fig. 5.13). The average initial hydraulic conductivity and average hydraulic
conductivity after five freeze-thaw cycles for the Parkview specimens were 1.9 x 10'10
and ,2.5 x 10'8 m/s, respectively. That is, the average hydraulic conductivity of the
Parkview specimens increased by more than two orders of magnitude1 after exposure
to five freeze-thaw cycles. ;
The average initial hydraulic conductivity and average hydraulic conductivity
after five freeze-thaw cycles for the Valley Trail specimens were 1.3 x J10-10 and 6.4 x
10~9 m/s, respectively, which corresponds to increase in hydraulic conductivity by a
factor of 50 as a result of freeze-thaw. These results are consistent with results
reported in other studies (Chamberlain et al. 1990, Zimmie and LaPlante 1990, Kim
and Daniel 1992, Othman and Benson 1993, Bowders and McClelland; 1994).
72
-------
10"
o
O
o
1
T>
io-
10'
10
"10
5 10 15 20
Molding Water Content (%)
25
Figure 5.11. Hydraulic conductivity vs. molding water content for Parkview clay.
o
3
o
O
o
10'7 r
io
-10
10
11
10 15 20
Molding Water Content (%)
25
Figure 5.12. Hydraulic conductivity vs. molding water content for Valley Trail clay.
73
-------
10
,-7
CO
E
> 10'B r
C
o
O
.y io~* -
CO
10
-9
-10
(a)
Specimen 1
Specimen 2
Specimen 3
0 1 2 3 4 5
Number of Freeze-Thaw Cycles
10
-7
10
-8
o
"I 10
o
O
to
T3
>*
-9
10
-11
Specimen 1
Specimen 2
Specimen 3
0 12 3 4 5
Number of Freeze-Thaw Cycles
6
Figure 5.13. Hydraulic conductivity vs. number of freeze-thaw cycles for Parkview (a)
and Valley Trail (b) clay. • ;
74
-------
Table 5.5. Results of freeze-thaw tests on specimens of Parkview and Valley Trail
clay compacted in the laboratory.
Sample
Number1
PV-3D-1
PV-3D-2
PV-3D-3
VT-3D-1
VT-3D-2
VT-3D-3
Initial
Hydraulic
Conductivity
Ko
(m/s)
2.4 x10-10
2.0 x10-10
1.3x 10'10
2.2 x10-10
7.6x10-"
8.3 x10'11
Hydraulic Conductivity After n Freeze-
Thaw Cycles, Kn
(m/s)
Ki
1.0x10-8
7.4x10-9
6.1 x 10-9
2.2x10-9
3.2x10-9
3.8x10-9
K3
2.4x10-8
' 2.0x10-8
1.9 x 10-8
8.3x10-9
9.1 xlO-9
1.3x10-8
K5
2.6x10-8
2.6x10-8
2.2 x 10-8
3.4x10-9
7.1 xlO-9
8.8x10-9
K5(2) .
KO
108
130
169
15
93
106
Note: • i '
1. PV = Parkview; VT = Valley Trail; 3D = Specimen was frozen and thawed three-dimensionally; -1, -2, -3
= Specimen Number. _,
2. KS/KQ = Hydraulic conductivity of specimen after exposure to 5 freeze-thaw cycles divided by hydraulic
conductivity before exposure to freeze-thaw. • i
75
-------
5.2.3 One-Dimensional Freeze-Thaw Tests
To determine if differences in freezing rate affect the increase in hydraulic
conductivity caused by freeze-thaw, the relationship between hydraulic conductivity
and freezing rate was determined for three specimens of each clay. Methods
described in Sec. 4.1.2.3 were used. The specimens were subjected;to one freeze-
thaw cycle so that they could be compared with the field data. i
Figures 5.14 and 5.15 show the relationship between hydraulic conductivity and
freezing rate for each clay. Laboratory data from the standard freeze-thaw tests and
field data from the COLDICE project are also presented. The hydraulic conductivities
for the standard freeze-thaw tests are the averages of three specimens. No trend
between hydraulic conductivity and freezing rate is evident. ;
5.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS
Summaries of results from the hydraulic conductivity tests performed in the field
and in the laboratory on clays is shown in Tables 5.6. and 5.7. An increase in
hydraulic conductivity occurred as a result of freeze-thaw for both clays regardless of
whether freeze-thaw occurred in the field or was simulated ope- or three-
dimensionally in laboratory tests. ;
Increases in hydraulic conductivity up to 4 orders of magnitude were observed
for specimens removed from the COLDICE test pads as large blpcks. Similar
hydraulic conductivities were measured for specimens removed from the test pads as
frozen cores. However, an increase in hydraulic conductivity was not observed for the
specimens collected in sampling tubes. Othman et al. (1994) and Benson et al. (1995)
report that sampling as block specimens or frozen cores preserves the structure of
compacted clay subjected to freeze-thaw, whereas sampling with thin-wall tubes
disturbs the soil structure which results in a specimen that is not representative of the
field condition. The findings of this study are consistent with those of^ Othman et al.
(1994) and Benson etal. (1995). , .
Hydraulic conductivities of the Parkview and Valley Trail specimens frozen
using the standard three-dimensional freeze-thaw method increased by factors of 130
and 50, respectively, after five cycles of freeze-thaw. The specimens frozen and
thawed using the one-dimensional method increased in hydraulic conductivity by
factors of 62 to 140, respectively, after one freeze-thaw cycle. In contrast, the hydraulic
conductivity of the block specimens (which are assumed to reflect field conditions)
increased in hydraulic conductivity more than three orders of magnitude in a single
76
-------
1 U
5f <
.g. 10"6
&
1
3
1 10'7
o
O
.g
"3
CO . _Jt
•o 10*
I*
•m-9
....,,.,., , .,....,,...,.
O Field
D 1-D Specimens
O Standard Specimens
Hydraulic Conductivity Before
Freeze-Thaw = 2.9 x 10"10 m/s
-
- . D
'
r • . . o
: t
, , , , I , , , , 1 , , , , I , , , , I , I , , I , , ,
=
j
-
E
-
-=
-
;
*
-
—
.
"
10 20 30 40
Freezing Rate (mm/hr)
50
60
Figure 5.14. Hydraulic conductivity vs. freezing rate for Parkview clay.
1 U
(
CO
1. 10'7
1
"D -1 Q-8
0
O
g
"3
1 1°'9
in-10
••••!•••• 1 ' • • • 1 :
o Field I
D 1-D Specimens
O Standard Specimens
Hydraulic Conductivity Before ;
! D Freeze-Thaw = 1.4 x 10"'° m/s \
~ ~
; -
-
• ^ o -
a
- D " -=
— j — i — i — i — I — i — i — i — i — ! — i — i — i — i — 1 — i — i — i — i — I — i — i — i — i — I — i — i — i — i —
0 10 20 30 40 50 60
Freezing Rate (mm/hr)
Figure 5.15. Hydraulic conductivity vs. freezing rate for Valley Trail clay.
77
-------
Table 5.6. Summary of hydraulic conductivity tests performed on Parkview field and
laboratory specimens.
Specimen^)
PV-Block
PV-Block
PV-Block
PV-TW-1
PV-TW-2
PV-TW-3
PV-TW-4
PV-TW-5
PV-CB-1
PV-CB-2
PV-CB-3
PV-CB-4
PV-Lab 3-D
PV-Lab 1-D
Sample Depth
(m)
0-0.3
0.3-0.6
0.6-0.9
0.10
0.15
0.25
0.45
0.52
0.05
0.1
0.15
0.2
N/A
N/A
Initial
Hydraulic
Conductivity*2)
(m/s)
1.9x 10'10
2.2 x 10'10
4.5 x 10'10
2.9x10-1o(4)
2.9x10-1°(4)
2.9x10-1°(4)
2.9x10-10<4)
2.9x10-1°(4)
2.9x10-10<4)
2.9x10-1°(4)
2.9 x 10-1°(4)
2.9 x 10-1°(4)
1 .9 x 1 0'10
1.9x10-10<6)
Final
Hydraulic
Conductivity(2)
(m/s)
1.9x 10'6
4.4 x 10'7
2.5 x 10'10
1.0x 10-9
1.0x 10-9
4.5 x 10-1o
1.6 x10-10
1.6 x10-10
8.5 x 10-8
1.3x 10'6
1.7x 10-7
5.4 x 10'7
2.5x1 0-8 (5)
1.2x10-8(7)
; K{ (3>
i Ki
10,000
2,000
0.56
1 0.35
'• 0.35
, 1.6
1 0.55
' 0.55
: 290
i 4,500
: 590
; 1 ,800
'130
; 62
Notes: \
1. PV = Parkview; Block = Block specimen(s); TW = Thin-wall tube specimen; CB = QRREL core barrel
specimen; Lab 3-D = Laboratory compacted specimens frozen three-dimensibnally; Lab 1-D =
Laboratory compacted specimens frozen one-dimensionally; -1, -2, -3, -4, -5 = Specimen number
2. Hydraulic conductivities are reported as averages for specimens removed from the test pad from a
depth of 0-0.3 m after winter (2 specimens) and for 3D and 1D laboratory compacted specimens (3
specimens). .
3. Change in hydraulic conductivity (Kf/Kj) is defined as the final hydraulic conductivity divided by the
initial hydraulic conductivity.
4. No specimens collected before winter in thin wall tubes or as frozen cores, thus;average hydraulic
conductivity is reported as the average hydraulic conductivity for the block specimens collected
before winter.
5. Average hydraulic conductivity after 5 freeze-thaw cycles.
6. Initial hydraulic conductivity tests were not performed on specimens frozen one-dimensionally. The
initial value is assumed to be similar to the average initial hydraulic conductivity of the specimens
frozen three-dimensionally because they were compacted under identical condtions (compactive
effort and molding water content).
78
-------
Table 5.7. Summary of hydraulic conductivity tests performed on Valley Trail field
and laboratory specimens.
SpecimenO)
VT-Block
VT-Block
VT-Block
VT-TW-1
VT-TW-2
VT-TW-3
VT-TW-4
VT-TW-5
VT-CB-1
VT-CB-2
VT-CB-3
VT-Lab 3-D
VT-Lab 1-D
Sample Depth
(m)
0-0.3
0.3-0.6
0.6-0.9
0.07
0.15
0.25
0.45
0.52
0.15
0.25
0.35
N/A
N/A
Initial
Hydraulic
Conductivity^)
(m/s)
9.8 x10-11
1.6x10-10
1.5 x10'10
1.4x10-10<4)
1.4x10-1°(4>
1.4x10-10<4)
1.4x10-1°(4)
1.4x10-1°(4)
1.4x10-10(4)
1.4x10-1°(4)
1.4x10-1o(4)
1.3 x10'10
1.3x10-1°(6)
Final
Hydraulic
Conductivity^)
(m/s)
2.6 x 10-7
2.6 x 10'7
2.2 x 10'10
4.0 x10'10
1.5X-10-10
1.5 x10"10
1.5 x10'10
1.5x 10'10
1.2x10-8
8.0 x10-7
8.5 x 10-9
6.4 x 10-g(5)
1.8x10-8(7)
I ,
i
1 (V3>
! KI
2,600
. 1 ,600
! 1.5
I 2.8
1.1
; 1.1
. 1.1
'. 1.1
! 86
' 5,700
: 61
' 50
i 140
Motes: ;
1. VT = Valley Trail; Block = Block specimen(s); TW = Thin-wall tube specimen; CB = CRREL core barrel
specimen; Lab 3-D = Laboratory compacted specimens frozen and thawed three-dimensionally; Lab
1-D = Laboratory compacted specimens frozen one-dimensionally; -1, -2, -3, -A, -5 = Specimen
number. !
2. Hydraulic conductivities are reported as averages for specimens removed from the test pad from a
depth of 0-0.3 m after winter (2 specimens) and for 3D and 1D laboratory compacted specimens (3
specimens). !
3. Change in hydraulic conductivity (Kf/Kj) is defined as the final hydraulic conductivity divided by the
initial hydraulic conductivity. :
4. Mo specimens collected before winter in thin wall tubes or as frozen cores, thus: average hydraulic
conductivity is reported as the average hydraulic conductivity for the block specimens collected
before winter.
5. Average hydraulic conductivity after 5 freeze-thaw cycles.
6. Initial hydraulic conductivity tests were not performed on specimens frozen one-dimensionally. The
initial value is assumed to be similar to the average initial hydraulic conductivity of the specimens
frozen three-dimensionally because they were compacted under identical conditions (compactive
effort and molding water content). •
79
-------
cycle of freeze-thaw. Thus, for these soils the increase in hydraulic conductivity that
occurs in the field is an order of magnitude larger than the increase in hydraulic
conductivity measured in the laboratory using one-dimensional, or three-dimensional
freezing methods. Benson and Othman (1993) reported a similar difference in their
field and laboratory tests. '
The cause of this difference is not clear. However, the cracks occurring in the
field were more widely spaced and had larger aperture than those that occurred in
specimens frozen and thawed in the laboratory (Figs. 5.7 and 5.10: vs. Fig. 5.16).
Furthermore, the ice lenses observed in the core specimens (Fig. 5.9) were more
randomly oriented than those observed in the specimens frozen and thawed in the
laboratory (Fig. 5.16). However, without a more detailed study of changes in soil fabric
and structure, the mechanism responsible for the difference in the increase in
hydraulic conductivity occurring in the laboratory and field cannot be determined with
certainty.
80
-------
Figure 5.16. Interior of Parkview specimen frozen and thawed in the laboratory.
81
-------
SECTION 6 ;
RESULTS :
SAND-BENTONITE MIXTURE
i
13.1 FIELD TESTS
6.1..1 Freeze-Thaw Monitoring ; .
Temperatures were monitored by CRREL personnel in the sand-bentonite test
pad constructed for the COLDICE project throughout the winter of 1992-93. Freezing
of the test pad began in November 1992 and was steady in mid- to late December
1992. In, mid- to late March 1993, thawing became steady. Complete thaw of the test
pad had occurred by the first week in April 1993. The test pad remained frozen for
I
about 3.5 months (Erickson et al. 1994).
Figure 6.1 shows a profile of frost penetration over time. Freezing records show
that once steady freezing was established, freeze-thaw cycling only pccurred in the
overburden material and not in the sand-bentonite mixture. Figure 6.1 shows that frost
had fully penetrated the test pad. It was determined that the sand-bentonite mixture
was subjected to one freeze-thaw cycle during the winter of 1992-93 (Fig. 6.1).
6.1 .,2 In Situ Box Infiltrometers
In situ hydraulic conductivity of the sand-bentonite test pad was measured by
personnel from CH2M Hill, Inc. after the test pad had been exposed to!two winters. A
box infiltrometer (Sec. 4.1.1.2) was used. A hydraulic conductivity of 5 k 10'10 m/s was
i
measured in June and July 1994 (Erickson et al. 1994). No measurement of field
hydraulic conductivity was made before the test pad had frozen. Therefore, no
comparison of hydraulic conductivity before and after winter couid be made. However,
visual examination of the soil during disassembly of the infiltrometer revealed that the
mixture was soft and contained none of the crack structures typically seen in
compacted clays exposed to freeze-thaw. Similar structure was observed prior to
winter when the first set of block specimens was collected. Thus, it appeared that the
sand-bentonite mixture had not been affected by, freeze-thaw.
f
Water content of the sand-bentonite mixture was measured at three depths
within the box infiltrometer by personnel from the University of Wisconsin-Madison.
Samples were taken at depths of 0.07 m, 0.14 m, and 0.2 m. The wateY contents were
42.1, 20.4, and 18.2%, respectively. These water contents were all higher than the
82
-------
water content at which the test pad was compacted (=17%). The high: water contents
near the surface of the test pad were probably the result of hydration and swelling of
the bentonite. Also, the presence of high water contents in the upper surface of the
test pad indicates that the sand-bentonite mixture was able to retain water after
exposure to freeze-thaw. I
i
i
6.1.3 Laboratory Assessment of Field-Scale Hydraulic Conductivity
Hydraulic conductivity tests were performed at the University of Wisconsin-
Madison on block specimens removed from the sand-bentonite test pad. Tests were
performed on three specimens removed before winter and three specimens removed
after one winter of exposure. Results of the hydraulic conductivity tests are
summarized in Table 6.1.
Two of the specimens removed before winter (depths = 0-0.3 mjand 0.3-0.6 m)
had high hydraulic conductivity (> 1.0 x 10'8 m/s), whereas the third specimen (depth =
0.15-0.45 m) had very low hydraulic conductivity (3.1 x 10'11 m/s). -The specimen
having low hydraulic conductivity (intermediate depth (0.15-0.45 m) was permeated for
55 days before the test was terminated. The specimen never met the outflow/inflow
criterion, but the hydraulic conductivity was steady for most of the testijig period. The
final outflow/inflow ratio was approximately 0.2. ; ,
The high hydraulic conductivity of the other two specimens was 'caused by loss
of bentonite (i.e., "piping" of bentonite). The effluent burettes for these specimens were
clouded with bentonite soon after testing was initiated.
Piping of bentonite also occurred in all three specimens removed from the
sand-bentonite test pad after winter. Bentonite was also present in the effluent
burettes of all three specimens soon after the hydraulic conductivity testing began.
Hydraulic conductivity tests were also performed at the University of Wisconsin-
Madison on specimens removed from the sand-bentonite test pad in June 1994 using
a thin-wall sampling tube (diameter = 71 mm). Results of the tests are summarized in
Table 6.2. The specimens were removed from a depth of 0-0.6 mj The average
hydraulic conductivity for the four sand-bentonite specimens was 1.1 x 10-10 m/s.
Hydraulic conductivity tests were also performed at CRREUon specimens
removed from the test pad as 71 mm-diameter frozen cores. All of :the specimens
exhibited piping of bentonite. Therefore, no laboratory assessment of post-winter
hydraulic conductivity could be made from the specimens removed as frozen cores or
those removed as blocks. .
83
-------
Table 6.1. Results of hydraulic conductivity tests on block specimens removed from
the sand-bentonite test pad.
Specimen^)
SB-BIock-Before-1
SB-Block-Before-2
SB-Block-Before-3
SB-Block-After-1
SB-Block-After-2
SB-Block-After-3
Sample Depth
(m)
0-0.3
0.3-0.6
0.6-0.9
0-0.3
0.3-0.6
0.6-0.9
Hydraulic Conductivity
(m/s)
1.2x10-7
3.1 x 10-11
2.4x10-8
6.6 xlO'7 :
2.4 x 10'7
5.7x10-8
Notes: !
1. SB = Sand-bentonite; Block = Block specimen; Before = Sampled before winter; After = Sampled
after winter; -1, -2, -3 = Specimen number. ; '
Table 6.2. Results of hydraulic conductivity tests on thin-wall tube specimens
removed from the sand-bentonite test pad. '•
Specimen^)
SB-TW-1
SB-TW-2
SB-TW-3
SB-TW-4
Sample Depth
(m)
0.15
0.25
0.35
0.45
Hydraulic Conductivity
(m/s)
1.3x 10-10
1.8x10-10
8.7 x 10-n
3.0 x 10-n
Note: i
1. SB = Sand-bentonite; TW = Specimen sampled using a thin-wall tube; -1, -2, -3, -4 = Specimen
number. . ,
84
-------
6.2 LABORATORY TESTS
6.2.1 Hydraulic Conductivity-Water Content Relationships i
Hydraulic conductivity was measured for four specimens compacted with
standard Proctor effort and four compacted with modified Proctor effort (Fig. 6.2).
Results of the hydraulic conductivity tests are tabulated in Appendix A. ;
Examination of Fig. 6.2 reveals that the hydraulic conductivity of the sand-
bentonite mixture is less sensitive to molding water content than the hydraulic
conductivity of compacted clays (e.g., Sec. 5.2.1). The results also show that the
hydraulic conductivity of the sand-bentonite mixture is not sensitive; to compactive
effort for. molding water contents dry of optimum, which is in direct icontrast to the
behavior of most compacted clays. '
One specimen (modified Proctor compactive effort, molding Water content =
7.3%) showed anomalously high hydraulic conductivity. Bentonite was noticed in the
effluent burette during permeation. Examination of the specimen after lit was removed
from the permeameter showed a thin zone devoid of bentonite along trie entire side of
the specimen (i.e., "piping" of the bentonite had occurred). This piping explains the
bentonite seen in the effluent and the high hydraulic conductivity that Was measured.
Other than the thin zone low in bentonite, the specimen appeared jo be a dense,
homogeneous, low-conductivity mass. i
6.2.2 Standard Freeze-Thaw Tests !
Three specimens were compacted at standard Proctor effort at the water content
at which the sand-bentonite test pad was compacted for the COLDICE project (=17%,
see Fig. 4.12). The hydraulic conductivity of each specimen was measured after 0, 1,
3, and 5 freeze-thaw cycles. All of the specimens were permeated for a; minimum of 30
days. The hydraulic conductivity for each specimen was steady when the test was
terminated. However, the outflow/inflow criterion was not met for any of the specimens
after any number of freeze-thaw cycles. Results of the freeze-thaw tests conducted on
the sand-bentonite mixture are summarized in Fig. 6.3 and Table 6.3. :
All three specimens exhibited piping of bentonite after three freeze-thaw cycles
and one specimen exhibited piping of bentonite after five freeze-thaw cycles.
Specimens that exhibited piping of bentonite were removed from their! permeameters
and examined. A thin zone (=5 mm wide) appearing low in bentonite was observed
along the side of each specimen, which was continuous from one end of the specimen
85
-------
u.u
01
. I
00
.£.
E"
-o 0.3
nj
~ 0.4
= °-5
£ n R
g. °-6
^ 0,7
0.8
0.9
1 • • i • • • i • • • i • • ' i • ' • i ' ' '
Thaw occurring
from the top of
L the test pad ~^~ i
~\ , downward \
-\ \
_ * — W^^Y Frost Zone t
- V^^, - \ :
\ \
/
Test pad only 0.6 m-thick
— —
, , , 1 . . i 1 . . . ! . . . 1 . . . 1 . . i
i
i "
!
1
,
:
' '
\
i
0 20 40 60 80 100 120 !
Time (days since December 21, 1992) ;
Figure 6.1. Depth of frost penetration vs. time for the COLDICE sand-bentonite test
pad (after Chamberlain 1994, personal communication). ;
• . • - p i
1C'7
5f
.§, 10"8
•>
I
1 1 0'9
0
O
z>
W j _ -10
-o 1 0
' ir^
I
m-11
: ' ' ' 1 ' ' ' 1 ' ' * 1 ' ' ' 1 ' ' ' :
n , ,,
• h- ° Standard
- \ Q Modified
\
• Specimen Piped :
-
- -
• -
.
°" ^^^^—o
r- '"-O • ~
: D ""n -
': , ~~~ D :
, , , i , , , i , , , i , , , i , , ,
i
i
[
i
'!,
I
]-
i
8 12 16 20
Molding Water Content (%)
24
Figure 6.2. Hydraulic conductivity vs. molding water content for;sand-bentonite
mixture.
86
-------
to the other. To correct the piping problem, a bentonite paste consisting of powdered
bentonite mixed with tap water was applied to these zones. The specimens were
placed in their permeameters and permeated again. Addition of the bentonite paste
alleviated the piping problem. One specimen piped again after 5 freeze-thaw cycles.
The same procedure was used to correct the problem with that specimen.
Because zones low in bentonite were only observed on the sides of the
specimens, piping of bentonite is likely a result of particle movement under a high
hydraulic gradient (=30) while thawing and not a direct result of freeze-thaw.
6.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS
A summary of the results of the field and laboratory hydraulic conductivity tests
is shown in Table 6.4. A quantitative comparison of field- and laboratory-scale
hydraulic conductivities for the sand-bentonite is difficult because the quantity of field
data is sparse. All but one of the block specimens removed from the itest pad before
winter piped and all of the block specimens removed from the test pad after winter
piped. Nevertheless, a qualitative comparison between the effect of ifreeze-thaw on
sand-bentonite at field- and laboratory-scales is possible.
Visual examination of the soil within the box infiltrometer revealed that the
hydraulic integrity of the sand-bentonite mixture was not deleteriou;sly affected by
exposure to freeze-thaw. The top of the sand-bentonite within the box infiltrometer
was soft and saturated, whereas soil within the box infiltrometer installed in the
i
Parkview clay test pad was cracked and did not prevent the flow of water (Sec. 5.1.2).
The high water contents, which decreased with depth (Sec. 6.1.2), occurred because
the sand-bentonite mixture restricted infiltration and allowed the berjtonite to swell.
These conditions are consistent with the low hydraulic conductivity that was measured
by Erickson et al. (1994) after two winters of exposure and two freeze-thaw cycles
(Table 6.4). !
The block specimens removed before and after winter exposure iwere examined
following permeation to observe their structure. The specimens were opened by
prying them apart with a long screwdriver. This method prevents smearing of the
specimen, which occurs with a saw or blade. Zones rich in bentonite were found when
the specimens were cut open (Fig 6.4), which are believed to be the result of poor
mixing of the sand and bentonite during preparation. The high hydraulic conductivity
of the block specimens was most likely the result of insufficient mixing of the sand and
bentonite and not freeze-thaw, because high hydraulic conductivities were measured
87
-------
Table 6.3.
Results of freeze-thaw tests on sand-bentonite specimens compacted in
the laboratory. ;
Specimen*1)
SB-3D-1
SB-3D-2
SB-3D-3
Initial
Hydraulic
Conductivity
KO
(m/s)
1.5 x10-11
1.4 x10-11
1.6 x10'11
Hydraulic Conductivity After n Freeze-
Thaw Cycles, Kn
(m/s)
Ki
2.2 x ID'11
1.7 x10-11
3.2 x10'11
K3
1.9X1CH1
1.6x 1CH1
1.8 x10'1-1
K5
5.1 X10'11
3.0 x 1CH1
2.1 x10'11
K5<2>
K0
3.4
2.1
Notes:
1. SB = Sand-bentonite; 3D = Specimen was frozen and thawed three-dimensibnally -1 -2 -3 =
Specimen number.
2. K5/K0 = Hydraulic conductivity of specimen after 5 freeze-thaw cycles divided by the hydraulic
conductivity before exposure to freeze-thaw. ;
Table 6.4. Summary of freeze-thaw test results for sand-bentonite mixture.
Before
Freeze-Thaw
After
Freeze-Thaw
KA/ (6)
/KB
Hydraulic Conductivity, K, by Sampling Method
(m/s) :
Box
Infiltrometers*1)
N/A(5)
<5x10-10
N/A
Block
Specimens*2)
3.1 x10-11
N/A
N/A
Thin-Walled
Tubes*3)
N/A
1.1 x10-10
N/A
Laboratory
Compacted
Specimens*4)
!1.5x 10-11
3.4 x10-11
! 2.3
Notes:
1. , Hydraulic conductivity was only measured after winter using a box infiltrometer (Erickson et al 1994)
2. Elefore freeze-thaw value is for one specimen (depth = 0.15-0.45 m). All of the specimens removed
after winter piped. v
3. Specimens were only collected in thin-wall sampling tubes after winter
4. Values are averages for three specimens. After freeze-thaw value is the average hydraulic
conductivity of three specimens after 5 cycles of freeze-thaw
5. N/A = not available.
6. KA/KB is defined as the ratio of average before freeze-thaw hydraulic conductivity to average after
freeze-thaw hydraulic conductivity. !
88
-------
conductivity
corresponds
than
for specimens removed from the test pads before and after winter.
block specimens also showed that exposure to freeze-thaw did not
the sand-bentonite (Fig. 6.4) as was observed for the Parkview and
(Sec. 5.1.3.4).
The freeze-thaw tests performed on the specimens prepared
also showed that freeze-thaw did not affect the hydraulic
bentonite mixture (Fig. 6.3). A two-tailed t-test was performed to
hydraulic conductivity after five freeze-thaw cycles and the mean
conductivity. A confidence interval of 0.05 was used, which
score (tcr) of 2.13. A t-statistic of 1.89 was calculated, which is less
freeze-thaw had no statistically significant effect on the hydraulic
sand-bentonite mixture. Furthermore, the internal structure of the
specimens showed no signs of cracking in the frozen or thawed states
The specimens collected in thin-wall tubes also had low
One of the specimens was opened to reveal the soil structure after
thaw.. Figure 6.6 is a photograph of the opened specimen. A
is evident; none of the crack structures typically seen in compacted
freeze-thaw exist. Thus, based on similar structures existing in the
block specimens, thin-wall tube specimens, and laboratory compacted
the low hydraulic conductivity measured in both the field (box
laboratory (laboratory compacted specimens), it is inferred that freeze
affect this sand-bentonite mixture.
Examination of the
result in cracking of
Valley Trail clays
in
hydraulic
homogeneo
clays
box
89
the laboratory
of the sand-
compare the mean
initial hydraulic
to a critical t-
tCr- Therefore,
conductivity of the
sand-bentonite
(Fig. 6.5).
conductivity.
exposure to freeze-
ius soil mass
subjected to
infiltrometer,
specimens and
infiltrometer) and
-thaw does not
-------
10*
•52
e,
& m
10
,-10
o
O
4| 10
|.
-11
- Specimen 1
• Specimen 2
• Specimen 3
•| Q I—i—i—i—i—I—i—i—i—i—I—i i i i I
_L
j—i—i—'ii'i'
Figure 6.3.
0 .1 2 3 45 6 ;
Number of Freeze-Thaw Cycles '
i
Hydraulic conductivity vs. number of freeze-thaw cycles for sand-
bentonite. '
Bentonite
CJVfiddle)
Figure 6.4. Photograph of sand-bentonite block specimens before Exposure to
freeze-thaw (a) and after exposure to freeze-thaw (b) showing internal
structure and bentonite-rich zones. .:
90
-------
Sand-Bentonite
Laboratory Compacted
Specimen
(after freeze-thaw)
Figure 6.5. Photograph of internal structure of sand-bentonite specimen compacted
in the laboratory and subjected to freeze-thaw. ;
Figure 6.6.
Sand-Bentonite
Thin-Wall Tube Specimen
(after freeze-thaw)
Photograph of the internal structure of sand-bentonite specimen removed
from the COLDICE test pad after winter using a thin-wafl tube.
91
-------
SECTION 7
RESULTS
GEOSYNTHETIC CLAY LINERS (GCLs)
7.1 FIELD TESTS ;
7.1.1 GCL Test Ponds and Pans \
Temperatures beneath the GCLs were monitored by CRREL personnel during
the winters of 1992-93 and 1993-94 in the test ponds and over the winter of 1993-94 in
the test pans (Fig. 7.1). From the temperature data, it was determined |that the GCLs in
the test ponds underwent 2 freeze-thaw cycles (one each winter) andjthe GCLs in the
test pans underwent 1 freeze-thaw cycle. Hydraulic conductivity of the GCLs was
measured in the GCL test pans before and after the winter of 1993-94 by monitoring
outflow. Hydraulic conductivity of the test ponds was not measured because of
problems caused by leaks (Erickson et al. 1994). Results of the hydraulic conductivity
tests are summarized in Table 7.1. !
Seepage was collected before winter from all but one of the test pans (Erickson
et al. 1994). The GCL test pan from which no seepage was collected apparently was
still hydrating when seepage monitoring was terminated. Analysis of the seepage data
showed that all of the GCLs had low hydraulic conductivity (=1.0 x 1CK10 to 2.8 x 10-"10
m/s).
All of the GCL test pans had measurable amounts of seepage after winter. The
average post-winter hydraulic conductivities of the Bentomat® and Claymax® GCL test
pans were 1.4 x 10-10 and 2.9 x 10-1° m/s, respectively (Erickson et al. 1994).
The effect of freeze-thaw on the hydraulic integrity of seamed GCLs can be
examined by comparing the average-post-winter hydraulic conductivity of specimens
containing seams with the post-winter hydraulic conductivity of the specimen without a
seam for each GCL type. The average hydraulic conductivity of the seamed Bentomat®
specimens after exposure to freeze-thaw was 1.7 x 10'10 m/s, which is slightly higher
than the post-winter hydraulic conductivity of the unseamed Bentomat® specimen (1.0 x
10-''o m/s). The two seamed Claymax® specimens had an average post-winter
hydraulic conductivity of 5.0 x 1Q-10 m/s, whereas the unseamed Claymax® specimen
had a post-winter hydraulic conductivity of 2.8 x 10-10 m/s. Thus slightly higher post-
winter hydraulic conductivities were measured for the seamed specimens for both GCL
types. Nevertheless, the change in hydraulic conductivity is so small that it is safe to
i
- 92 : '
-------
O
o
g>
I
8.
I
6
4
2
0
-2
-4
-6
- J
(a)-
Bentomat*
Claymax®
_L
0 20 40 60 80 100 120
Time (days since December 20, 1992)
O
o
I
o
Q.
E
05
10 20 30 40 50 60 70 80
Time (days since December 19, 1993)
Figure 7.1. Temperature vs. time beneath the GCLs in the COLDICE
the winters of 1992-93 (a) and 1993-94 (b) (from Chamberlain 1994
personal communication).
test ponds for
93
-------
Table 7.1.
Summary of hydraulic conductivity test results for the GCL test pans used
in the COLDICE project (from Erickson 1994, personal communication).
Specimen
Bentomat®, 1 .8 m2
Bentomat®, 0.7 m2
Bentomat®, 0.7 m2
Claymax®, 1 .8 m2
Claymax®, 0.7 m2
Claymax®, 0.7 m2
Seam?
Yes
Yes
No
Yes
Yes
No
Before-Winter
Hydraulic
Conductivity
(m/s)
1.5x10-10
1.0 x 10'10
no outflow
2,8 x 10'10
2.0 x 10'10
2.4 x10-10
After-Winter
Hydraulic
Conductivity
(m/s)
1.9 x 10-10
1.4 x 10'10
1.0 x 10'10
7.0 x 10'10
3.0 x 10'10
2.8 x 10-10
, ' ' " ""'
: KA_ (1)
' KB
' 1.3
1 1.4
N/A(2)
: 25.0
• 1.5
Note: i
1. KA/KB is defined as the ratio of after-winter hydraulic conductivity to before-winter hydraulic
conductivity. ,
2. N/A = Not Applicable ' , ' \
94
-------
conclude that the hydraulic integrity of these GCLs was not affected by exposure to
freeze-thaw. • i
7.1.2 Laboratory Assessment of Field-Scale Hydraulic Conductivity
Hydraulic conductivity was measured on GCL specimens rerpoved from the
COLDICE GCL test ponds after exposure to two winters (i.e., two freeze-thaw cycles).
Methods described in Sec. 4.3.1.2 were followed. The hydraulic conductivity tests
required less than one week to meet the termination criteria. Results:of the tests are
summarized in Table 7.2. ; '
Hydraulic conductivity was measured on two 0.45 m-diameter specimens of
Claymax® and two 0.30 m-diameter specimens of Bentomat®. : Originally, the
Bentomat® specimens had a diameter of 0.45 m. However, very high'flow rates were
measured when they were permeated. To determine the cause of the high hydraulic
conductivity, rhodamine dye was added to the influent water to stain Ithe flow paths.
Examination of the specimens after the permeameter was disassembled showed that
sidewall leakage occurred because of short-circuiting through geotextile? near the edge
of each specimen. To eliminate this problem, the Bentomat® specimens were trimmed
to a diameter of 0.30 m and re-tested.
The hydraulic conductivities of the Bentomat® specimens were 2.5 x 10-8 and 1.7
x 10-1° m/s, whereas the Claymax® specimens had hydraulic conductivities of 3.5 x 10-
1° and 6.3 x 10-1° m/s (Table 7.2). The high hydraulic conductivity for the one 0.30 m-
diameter Bentomat® specimen (BMT-PND-1) was measured after attempts were made
to prevent sidewall leakage. The high hydraulic conductivity of specimen BMT-PND-1 is
attributed to either sidewall leakage that could not be corrected or disturbance during
handling, but not freeze-thaw.
Figure 7.2 is a photograph of the two Bentomat® specimens 'after hydraulic
conductivity testing. The bentonite component of the Bentomat® specimen BMT-PND-1
shows no structural changes often attributed to freeze-thaw (e.g., cracks), and appears
similar to the bentonite component of the Bentomat® specimen having low hydraulic
conductivity (BMT-PND-2). i
< - _- --_ . . - ...,-,...._; .. . - _ L _ ; . . . \
7.2 LABORATORY TESTS i
7.2.1 Laboratory-Scale Hydraulic Conductivity Tests
Hydraulic conductivity tests were performed on three 0.15-m-diam0ter specimens
of Bentofix® and Bentomat® and four 0.15-m-diameter specimens of Claymax®. Initial
i
95 I . •
-------
Table 7.2. Results of hydraulic conductivity tests on GCL specimens removed from
the COLDICE test ponds. . j
Specimen1
BMT-PND-1
BMT-PND-2
CMX-PND-1
CMX-PND-2
Diameter
(m)
0.30
0.30
0.45
0.45
Hydraulic Conductivity
(m/s)
' 2.5x10-8
1.7xiO-1o
3.5 x lO'10
6.3 x 10-10
Note:
1. BMT = Bentomat®; CMX = Claymax®; PND = Specimen removed from test pond; 1-1, -2 = Specimen
number.
96
-------
Figure 7.2. Photograph of the two Bentomat® specimens removed from the GCL test
ponds at the COLDICE field site. :
97
-------
saturation of the specimens required 3 to 4 weeks. During this period, the hydraulic
conductivity decreased and the ratio of outflow to inflow increased, both slowly, as the
bentonite hydrated. Hydraulic conductivity tests conducted after freeze-thaw required
about 1 to 2 weeks before the termination criteria were met. In contrast to the tests
performed before freezing, the hydraulic conductivity was essentially steady from the
onset for the GCLs exposed to freeze-thaw and the outflow/inflow ratio fell between 0.75
and 1.25 immediately following initiation of the tests.
Results of the hydraulic conductivity tests are summarized in Figs. 7.3-7.5 and
Table 7.3. All of the hydraulic conductivities are low, ranging between 2.9 x 1CH1 m/s
and 4.9 x 10-11 m/s for the initial condition and 1.7 x 10-"11 m/s and 3.3 x 1iO-11 m/s for the
specimens exposed to 20 cycles of freeze-thaw. In the initial condition, the Bentomat®
GCLs had slightly lower hydraulic conductivity (3.0 x 10'11 m/s), on average, relative to
the Bentofix® (4.5 x 10'11 m/s) and Claymax® (4.0 x 10-11 m/s) GCLs. After 20 cycles of
freeze-thaw, the Bentomat® GCLs still had slightly lower hydraulic conductivity (1.8 x
10-11 m/s), relative to the Bentofix® (2.6 x 10'11 m/s) and Claymax® (2.8 x 1Q-11 m/s)
GCLs. Furthermore, for all three GCLs, a small decrease in hydraulic conductivity
apparently occurred as a consequence of freeze-thaw (Figs. 7.3-7.5). This decrease in
hydraulic conductivity is likely the result of thaw consolidation (Chamberlain and Gow
1979) or rearrangement of the bentonite particles. \
The slight decrease in hydraulic conductivity that occurs is evident in the ratio
K2o/Ko, which is defined as the hydraulic conductivity of a specimen after 20 cycles of
freeze-thaw divided by its initial hydraulic conductivity (Table 7.2). The ratio varies
between 0.55 to 0.66 for the Bentomat® GCLs, 0.45 to 1.10 for the Bentofix® GCLs, and
0.57 to 0.89 for the Claymax® GCLs. That is, for all but one of the GCL specimens, a
slight decrease in hydraulic conductivity occurred after 20 cycles of freeze-thaw. The
one specimen that exhibited an increase in hydraulic conductivity after 20 freeze-thaw
cycles (Bentofix®-1) showed a decrease in hydraulic conductivity after 1, 3, and 5
freeze-thaw cycles, relative to its initial hydraulic conductivity. i
To determine if the apparent decrease in hydraulic conductivity was statistically
significant, a t-test was conducted to compare the mean initial hydraulic conductivity and
the mean hydraulic conductivity after 20 cycles of freeze-thaw for each GCL. The
comparison was made at a significance level of 0.05, with a corresponding critical t (tcr)
of 1.96. .Results of the test indicate that the reduction in hydraulic conductivity is
significant for each GCL, with t-statistics of 14.7 (Bentomat®), 3.2 (Bentofix®), and 2.2
(Claymax®); i.e., t > tcr for each GCL.
98
-------
- Specimen 1
- Specimen 2
- Specimen 3
-t ID'10 \r
•o
s
— rr11
10
,-12
-i , L
0
5 10 15
Number of Freeze-Thaw Cycles
20
Figure 7.3. Results of freeze-thaw tests on specimens of Bentofix® frozen and thawed
in the laboratory.
10'
.-10
1
1
o
O
.o
-11
10'1
Specimen 1
Specimen 2
Specimen 3
5 10 15
Number of Freeze-Thaw Cycles
20
Figure 7.4. Results of freeze-thaw tests on specimens of Bentomat® frozen and
thawed in the laboratory. ;
99
-------
10
-9
.w
10
TO
o
O
,-10
1-11
~°
x*
10
,-12
Specimen 1
Specimen 2
Specimen 3
Specimen 4
5 10 15
Number of Freeze-Thaw Cycles
: 20
Figure 7.5. Results of freeze-thaw tests of specimens of Claymax® frozen and thawed
in the laboratory. '
100
-------
c:
CD
E
'o
CD
Q.
CO
o
jQ
_CO
_
O
a
CO
a>
N
CD
o
CO
E
E
CO
CO
v*
co
0)
o
>,
O
5
r5
k
N
0)
£
u.
c-ST
£^
<
>,
•*— *
>
o
o
o
o
5
•o
>%
o
«.i
-t— iz
|1"£
°lv^
. *:!*
c
^
ID
i£
co
^
^
>i
>
"o o^
f*t
Q
0
® 0
Q..Q
E E
03 3
CO Z
O
O
X
C\
CO
a
E
p
t
a
0
c
i
X
CO
•s
i
X
0
CO
o
X
O
V
8>
X
2
CD
00
LO
•^
o
o
X
OJ
CM
1
X
c\
1
X
CO
f\
1
X
CD
0
X
CD
^t
CM
i
®
X
5
o
00
LO
"N
0
t
X
in
c\
o
X
CD
CO
1
X
LO
co
o
X
o
X
to
10
CO
8)
X
5
CD
LO
LT
O
o
X
r^
o
X
CO
'•
X
00
r\
i
X
o
CM
0
X
CO
i
g>
a
5
CD
CD
C£
O
O
X
CO
i
X
LO
CM
'•
X
•<3-
r\
O
X
o
X
CO
CM
g)
3
5
CD
CO
CD
CD
0
0
X
a>
o
X
LO
i
X
'St
T"~
1
O
X
CO
*"
o
X
a>
CM
CO
®
co
5
5
en
CM
o
X
00
CO
©
X
CO
•>>
CO
O
CJ
r^<
o
o
X
CM
i
X
CD
CO
'•
X
CM
O
X
NJ-
CM
O
X
O)
CM
CM
i
g>
X
CO
CO
o
r^-
LO
o
i
X
•3-
c\
o
X
CM
CO
X
•sj-
oo
1
X
LO
CO
O
X
CM
••tf
CO
©
X
to
CO
0
sj
o
.
o
X
CO
rf
b
X
•*
•*'
1
o
X
CM
CO
1
o
X
-tf
b
X
O)
•f
*i-
@
X
CO
>,
CO
o
101
-------
7.2.2 Structure of Frozen Specimens
-1 '* -,i K>- ' '
The structure of the GCLs was examined to determine why the hydraulic
conductivity of GCLs does not change when they are frozen and thawed, whereas an
increase in hydraulic conductivity .is generally observed for compacted clays subjected
to freeze-thaw. The increase in hydraulic conductivity occurring in compacted clays is
attributed primarily to cracking (e.g., Fig. 5.10) that occurs during freezing (Chamberlain
et al. 1990, 1994, Othman and Benson 1993).
A band saw was used to cut open hydrated specimens of each GCL while they
were frozen. Two cuts were made; the first cut was made along the diameter of the
specimen to obtain a vertical cross-section, the second cut was macje parallel to the
geosynthetic layers to obtain a horizontal cross-section. A razor blade was used to
remove smeared soil that formed during sawing. The exposed frozen faces were
examined and photographed.
Small ice crystals in a random orientation existed in the face of each specimen
(Fig. 7.6). However, large ice lenses and distinct cracks, such as those observed in
frozen compacted clays by Othman and Benson (1993) and Chamberlain et al. (1990,
1994), did not exist in the GCLs. The lack of such cracks is consistent with the low
hydraulic conductivity measured after thawing. '-.
Figure 7.7 is a photograph of two hydrated GCL specimens. One of the
specimens has been exposed to 20 cycles of freeze-thaw and the other has never been
exposed to any freeze-thaw (initial saturation condition). Both specimens in Fig. 7.7
have similar structure. Their nearly identical structure, which is devoid of cracks, is
consistent with the similarity of their hydraulic conductivities. Apparently, the voids
containing ice lenses collapse as the bentonite returns to its soft, highly plastic condition
during thawing. i
I
7.2.3 Bentonite in Effluent
Bentonite was observed in the graduated cylinders used to collect effluent during
hydraulic conductivity testing, Bentonite appeared from the Bentofix® specimens after 3
,freeze-thaw cycles, from the Bentomat® specimens after 5 freeze-thaw cycles, and from
the Claymax® specimens after 20 freeze-thaw cycles. Migration of'bentonite from
Claymax® and Bentomat® GCLs has also been reported by Estornell and Daniel (1992)
in large bench-scale tests. In this study, migration of the bentonite probably occurred as
a consequence of particle movement that occurred during freezing and thawing.
102
-------
Clay ma
Figure 7.6. Photograph of section of frozen Claymax® GCll
After 20
Freeze-Thaw
Cycles
Figure 7.7. Photograph of hydrated Bentomat® specimens before and after freeze-
thaw. '•
103
-------
The presence of the bentonite in the effluent was a concern, because if a
sufficient quantity of bentonite was lost, an increase in hydraulic conductivity could
result. However, the quantity of expelled bentonite was small (about 0.4% of the mass
of the effluent, <0.01% of the mass of the GCL specimen) and had no influence on the
overall hydraulic integrity of the GCLs (Figs. 7.3-7.5). j
7.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS
In general, the hydraulic conductivities measured in the field (or in the laboratory
with specimens removed from the field) were greater than those measured in the
laboratory, regardless of whether the tests were conducted before or after freezing. The
higher hydraulic conductivities observed in the field were likely the result of different test
conditions. For example, the effective stress and hydraulic gradient were lower in the
field, which has been shown to result in higher hydraulic conductivity of GCLs (Estornell
and Daniel 1992). These conditions are difficult to simulate in flexible-wall
permeameters. More important, however, is that no significant increase in hydraulic
conductivity was observed in the field or laboratory tests. The lack of change in
hydraulic conductivity is consistent with the lack of change in structure bbserved in the
field and laboratory specimens (Figs. 7.2 and 7.7). That is, no cracks were present in
any of the GCLs that had been subjected to freeze-thaw.
The GCL specimens tested in the field that contained seams had slightly higher
hydraulic conductivities than the GCLs without seams, both before arid after freeze-
thaw. The higher hydraulic conductivity of the seamed specimens may have been
caused by the construction methods that were used or freeze-thaw. Additional research
regarding how freeze-thaw affects GCL seams should be conducted. ;
104
-------
SECTION 8
RESULTS
PAPER MILL SLUDGES
8.1 FIELD TESTS
8.1.1 Compaction of Pipe Specimens
Pipe specimens were constructed by compacting paper mill sludge in large
PVC pipes (diameter = 0.35 m, length = 0.6 m). Six specimens were constructed (two
specimens per sludge) using a compactive energy equal to standarjj Proctor effort
(592.5 kJ/m3). Molding water contents of 90, 120, and 130% were used for sludges A,
B, and C, respectively. These water contents correspond to the water contents
yielding the lowest hydraulic conductivities for specimens used to'determine the
hydraulic conductivity-water content relationship for each sludge (see Sec. 8.2.1).
Ensuring uniform water content within each specimen was difficult. The paper
mill sludges were dried to the molding water contents from their as-|received water
contents in a large, forced-air walk-in oven in the Structures and Materials Testing
Laboratory at the University of Wisconsin-Madison. The oven ternperature was
approximately 70°C. Approximately 100 kg of each sludge was dried at one time
using large pans. Mixing of the sludge to ensure uniform drying was difficult because
of the volume of sludge being dried. To alleviate this problem, the! partially dried
sludge was homogenized in sealed high density polyehtylene. (HOPE) drums for 48
hours prior to compaction. This ensured a more uniform distribution of water in the
sludge; however, the water content still varied within each specimen by up to 20%.
Nonetheless, the water content of each lift was greater than the target water content
corresponding to low hydraulic conductivity of the laboratory-compacted specimens
(see Sec. 8.2.1), which also results in low hydraulic conductivity (< 1 x 10-9 m/s).
8.1.2 Freeze-Thaw Monitoring of Field Specimens
The field specimens were buried in the ground outside the Environmental
Geotechnics Laboratory at the University of Wisconsin-Madison on December 1, 1993.
Air temperature, relative humidity, and temperatures within the specimens were
recorded every hour throughout the winter of 1993-94 using a Campbell Scientific
CR10 Datalogger. A photograph of the monitoring system is shown in Fig. 4.18. The
105
-------
other three specimens (control specimens) were permeated in the laboratory using the
procedure described in Sec. 4.4.1.3. ',
Freezing began in mid-December 1993 and was steady later that same month.
Thawing began in early February 1994 and complete thaw of the ^specimens had
occurred by mid-March 1994. Figures 8.1-8.3 show the temperature distribution in
each of the sludge specimens over time. One freeze-thaw cycle occurred in each
specimen, and frost penetrated the entire thickness of each specimen (Figs. 8.1-8.3).
The pipe specimens were removed from the ground immediately 'after thaw was
complete in spring 1994.
8.1.3 Hydraulic Conductivity of Pipe Specimens
Two methods were used to measure the hydraulic conductivity of each pipe
specimen (field and control specimens). First, the specimens were permeated in the
PVC pipes, as if the pipes were rigid-wall permeameters. Permeation of the control
specimens was conducted immediately following compaction, whereas permeation of
the field specimens was conducted soon after they were removed from the ground.
Second, after permeation, specimens were removed from the pipes in slices and
permeated in flexible-wall permeameters. Details of the testing procedures are
described in Sec. 4.4.1.3. Results of hydraulic conductivity tests performed on the pipe
specimens are summarized in Table 8.1. \
8.1.3.1 Hydraulic Conductivity Measured in Pipes
For each sludge, higher hydraulic conductivities were measured for the control
specimens than for the specimens exposed to freeze-thaw. The control specimens
had hydraulic conductivities ranging from 1.2 x 10'9 to 8.2 x 10-9 m/s, whereas the field
specimens exposed to one cycle of freeze-thaw had hydraulic conduptivities ranging
from 3.0 x 1CH0 to 4.1 x 10'10 m/s. That is, the hydraulic conductivity1 of each sludge
decreased by at least one-half order of magnitude after exposure to freeze-thaw.
These results suggest that the hydraulic integrity of these paper mill sludges is not
deleteriously affected by freeze-thaw.
8.1.3.2 Hydraulic Conductivity Measured in Flexible-Wall Permeameters
!
After permeation, each pipe specimen was sliced horizontally so that two
specimens were obtained for hydraulic conductivity testing in flexible-wall
106
-------
10
o
o
2
I
I
I
-5
-10
Depth=0.1 m
- Depth=0.2 m
- Depth=0.3 m
- Depth=0.4 m
• Depth=0.5 m
_L
20 40 60 80 100
Time (days since December 15, 1993)
120
Figure 8.1. Temperature vs. time at various depths within field specimen consisting
of paper mill sludge A. . i •
10
o
8.
-5 -
-10
- Depth=0.2 m
— — - Depth=0.3 m
Depth=0.4 m
Depth=0.5 m
0 20 40 60 80 100
Time (days since December 15,1993)
Figure 8.2. Temperature vs. time at various depths within field
of paper mill sludge B.
107
120
specimen consisting
-------
10
o
o
I
i
o
0.
e
-5 -
-10
— - Depth=0.3 m
Depth=0.4
Depth=0.5 m
20 40 60 80 100
Time (days since December 15, 1993)
Figure 8.3. Temperature vs. time at various depths within field specimen consisting
; of paper mill sludge C.
108
120
-------
permeameters. The specimens were either 0.30 m or 0.15 m in diameter (Table 8.2).
Details of the testing procedure are summarized in Sec. 4.4.1.4. ;
Results of the hydraulic conductivity tests are summarized in Table 8.2. The
hydraulic conductivities varied widely, ranging from 7.3 x 10"11 m/s (sludge A, no
exposure to freeze-thaw) to 6.2 x 10'9 m/s (sludge A, exposed to freeze-thaw).
Furthermore, the hydraulic conductivities of the specimens exposed to:freeze-thaw are
higher than those for the corresponding control specimens, which is in direct contrast
to results of the hydraulic conductivity tests performed directly on the;pipe specimens
(Sec. 8.1.3.1). Also, all of the control specimens tested in flexible-wal permeameters
have lower hydraulic conductivities than the hydraulic conductivities measured directly
in the pipes, whereas the hydraulic conductivities of the specimens exposed to freeze-
thaw that were tested in flexible-wall permeameters are higher than the hydraulic
conductivities measured directly in the pipes from which they were rernoved. The only
exception is sludge C, where the hydraulic conductivities are essentially the same.
Two hypotheses are provided for this behavior. First, disturbance may have
occurred when pipe specimens exposed to freeze-thaw were removed from the
ground or sliced for testing in flexible-wall permeameters. That is, ^cracks or other
macroscopic defects may have been opened when the specimens were transported
inside the laboratory or when specimens were removed from the pipes. Second, the
effective stress in the pipes (1.6 kPa) was lower than the effective; stresses in the
specimens tested in the flexible-wall permeameters (18 kPa). Thus, the pipe
specimens should have higher hydraulic conductivity (see Sec. 8.2.3). However, this
second hypothesis only explains the behavior of the control specimens.
As a result of these ambiguities, the results of the small-scale field tests are
inconclusive, in that exposure to freeze-thaw resulted in a decrease in hydraulic
conductivity for the pipe specimens and resulted in an increase in hydraulic
conductivity for specimens tested in flexible-wall permeameters. The results of the
tests conducted directly in the pipes are probably more representative of the field
condition. Nonetheless, a larger field study is needed to adequately address how
freeze-thaw affects the in situ hydraulic conductivity of paper mill sludges.
109
-------
Table 8.1. Results of hydraulic conductivity tests on paper mill sludge conducted in
the pipes. i
Sludge
A
B
C
Hydraulic conductivity of
Control Specimen
(no exposure to
freeze-thaw), Kc
(m/s)
1.2x10-9
8.2x10-9
1.2x10-9
Hydraulic Conductivity of
Field Specimen
(exposure to
freeze-thaw), Kf
(m/s)
3.5 x10-10
3.0 x10'10
4.1 x10-10
Change in
Hydraulic
Conductivity^1)
! Kf
i Kc
0.29
' 0.037
; 0.34
Note: I
1. Change in hydraulic conductivity is defined as the hydraulic conductivity of the^field specimen (Kf)
divided by the hydraulic conductivity of the control specimen (Kc).
Table 8.2. Results of hydraulic conductivity tests on sludge specimens tested in
flexible-wall permeameters.
Specimen^)
SLA-Control-T
SLA-Control-B
SLB-Control-T
SLB-Control-B
SLC-Control-T
SLC-Control-B
SLA-Field-T
SLA-Field-B
SLB-Field-T
SLB-Field-B
SLC-Field-T
SLC-Field-B
Specimen
Diameter
(m)
0.30
0.30
0.30
0.30
0.30
0.30
0.15
0.15
0.30
0.15
0.30
0.30
Hydraulic
Conductivity,
Kfiex
(m/s)
8.9 x10'11
7.3 x10'11
2.8 x10-10
3.1 x10-10
1.0 x10-10
1.6 x10-10
6.2x10-9
5.0x10-9
1.1 X10-9
1.2x10-9
5.5 x10'10
1.8 x10-10
Hydraulic :
Conductivity of
Parent Pipe
Specimen, Kp !
(m/s)
1.2x10-9 ;
8.2x10-9 ;
. 1.2x10-9
3.5x10-1°
i
3.0 xio-1° :
4.1 x 10'10 !
Kflex(2)
• KP
0.07
0.06
0.03
0.04
0.08
0.13
18
14
3.7
4.0
1.3
0.4
Notes: |
1. SLA = Sludge A; SLB = Sludge B; SLC = Sludge C; Control = Specimen taken from control pipe
specimen (no freeze-thaw exposure); Field = Specimen taken from field pipe specimen (exposure to
freeze-thaw); T= Specimen was taken from top of pipe specimen (depth < 0.3 m); B = Specimen was
taken from bottom of pipe specimen (depth > 0.3 m).
2- Kf|ex/Kp = Hydraulic conductivity of specimen tested in flexible-wall permeameter divided by hydraulic
conductivity of parent pipe specimen (taken from Table 8.1)
110
-------
8.2 LABORATORY TESTS
8.2.1 Hydraulic Conductivity-Water Content Relationships
Hydraulic conductivity tests were performed on the specimens compacted to
!
obtain the compaction curves described in Sec. 3.3.2. Figures 8.4-8.6 show the
relationship between hydraulic conductivity and molding water content for sludges A,
B, and C, respectively. Individual results of the hydraulic conductivity tests are
tabulated in Appendix A. The tests required up to 45 days to reach steady conditions.
The hydraulic conductivity of each sludge is sensitive to molding water content
(Figs. 8.4-8.6); i.e., the specimens compacted at molding water contents dry of
optimum have high hydraulic conductivities, whereas specimens compacted wet of
optimum have low hydraulic conductivities. Furthermore, wet of ioptimum water
content, hydraulic conductivities less than 1 x 10'9 m/s were obtained for each sludge.
This behavior was also observed for the two clays used in this study (Sec. 5.2.1) and is
characteristic of most clayey soils. The high percentage of fines in paper mill sludge
(Sec. 3.3.1) is likely responsible for the similarity in hydraulic behavior of paper mill
sludge and compacted clay.
8.2.2 Standard Freeze-Thaw Tests
- i
Hydraulic conductivity tests were performed on six specimens of sludge A and
five specimens of sludges B and C after a various number of freeze-thaw cycles.
Specimens of each sludge were compacted using standard Proctor effort at water
contents yielding the lowest hydraulic conductivities ("low-K" water content) and at the
water contents at which the sludges were received at the University "of Wisconsin-
Madison ("as-received" water contents). Four specimens of sludge A and three
specimens of sludges B and C were compacted at their low-K water contents and two
specimens of each sludge were compacted at their as-received water contents.
Results of the hydraulic conductivity tests on these specimens are i summarized in
Table 8.3 and Figs. 8.7-8.9. 5
8.2.2.1 Low-K Molding Water Contents j _
Some of the low-K specimens of sludge A showed an increase in hydraulic
conductivity after exposure to freeze-thaw, whereas others did not (Fig. 8.7). The two
low-K specimens of sludge A frozen and thawed five times without permeation after
each freeze-thaw cycle showed an increase in average hydraulic conductivity
hydraulic conductivity (27 times), whereas the two identical specimens permeated
111
-------
""
o
o
10"
10
10'
10"
10"
10
50 100 150
Molding Water Content (%)
200
Figure 8.4. Hydraulic conductivity vs. molding water content for paper mill sludge A.
ID"
« 10
1
.£•
%
o
O
o
I
•o
10-
1C'7
10'8
10'9 b-
10
-10
"As-Received" Water Content -
< i , , i , , , i i , Q i , i , i , , i i , , i i
50 100 150 200 250
Molding Water Content (%)
300
Figure 8.5. Hydraulic conductivity vs. molding water content for paper
mill sludge B.
112
-------
0
]5
•o
o
O
I
I
10
-2
10
,-4
10
-5
10
10"
-7
0
"As-Received" Water Content:
50 100 150 200
Molding Water Content (%)
250
Figure 8.6. Hydraulic conductivity vs. molding water content for paper mill sludge C.
113
-------
CO
CD
CO
0)
i
8
0
CO
E
E
u
CO
CO
00
^)
JO
c
"XI
V)
£
O
(X)
JZ
1™
CD
N
CD
CD
ul
C "7n
Jl
<
^
>
tS
•o
O
O
.g
CO
|
g
<0 m
•*- »;:
"X
c\T
*1*°
1C
k:
•^j-
*:
£
CM
^
X.
£.
> ^
'•tt f.'tn
f *£
O
O
cu *""
"o. CD
C JD
cB £
wi
00
CO
c*5"
Q.
o
r-
6
X
r-.
c\i
o
T—
6
X
o>
CO
CD
6
X
CM
T~
o>
CD
X
CD
CO
T—
T—
6
X
CM
l<
T_
^
5
CO
CD
CO
CD
n
z
o
6
X
o
CO
o
T—
6
X
•*
CD
6
X
CO
T~
0
6
X
o>
^
o
T—
6
X
-a-
00
CM
*
5
CO
o
CO
O)
6
X
CO
00
n
z:
Q_
z
1-
n
z:
n
0
r—
b
X
00
cvi
CO
X.
_j
5
CO
in
CM
o>
6
X
CO
00
n
z
Q_
Z
h-
n
z
n
z
0
T—
b
X
CO
CO
M-
*1
5
CO
OJ
o
0
b
X
b
X
CO
cvi
T—
CC
<
3
CO
CO
0
T—
b
X
m
co'
Q.
z
Q_
Z
n
z
o
T—
b
X
in
o
0
X
0
•*
b
X
p
CM
^
_l
CQ
_l
CO
CD
^f
op
b
X
h»
•^l-
Q.
Z
h-
op
b
X
CM
*"
oo
O
X
^
T~
0
b
X
"3-
CO
o
b
X
CO
o>
CO
^
CD
_J
CO
00
O)
op
b
X
to
•<*
00
b
X
m
CD
00
b
X
"
b
X
co
^
,_
CC
<
CO
CO
m
O5
op
b
X
O)
CO
op
b
X
m
CO
00
b
X
CO
CM
oo
O
X
T~
q>
b
X
o>
CO
en
b
X
^
CM
CC
<
CD
CO
CM
O)
op
b
X
CM
T~
°P
O
X
CM
oo
b
X
m
CM
oo
O
X
h-
CM
Op
b
X
•*
T~
o>
b
X
CO
T_
X.
3
CO
CO
co
b
X
CM
°P
O
X
CO
'•"
00
b
X
N.
1—
co
O
X
(•>-
CM
Op
b
X
o
T~
O)
b
X
CO
(M
X.
O
CO
0
1"~
O>
b
X
O)
o
b
X
O)
Tt
n
Z
H
oo
b
X
CO
r~
o>
b
X
0)
CO
0
T—
b
X
CD
00
CO
^
_l
O
CO
CD
CM
eo
b
X
00
CM
op
b
X
CO
"•"
05
b
X
CO
N-
CO
o
X
m
CM
9
o
X
m
CM
o
i—
b
X
00
CD
T—
CC
<
3
CO
00
co
b
X
•*
CM
CO
b
X
CO
op
b
X
CO
CM
00
O
X
m
-------
10
-11
Specimen-1 o Specimen-3
Specimen-2 A Specimen-4
0 1 2 3 4 5 :
Number of Freeze-Thaw Cycles
Specimen-1
Specimen-2
1 2 3 4
Number of Freeze-Thaw Cycles
Figure 8.7. Hydraulic conductivity vs. number of freeze-thaw cycles for low-K
specimens (a) and as-received specimens (b): paper mill sludge A.
115
-------
1 2 3 4
Number of Freeze-Thaw Cycles
Specimen-1
Specimen-2
0 1 23 4 5
Number of Freeze-Thaw Cycles
Figure 8.8. Hydraulic conductivity vs. number of freeze-thaw cycles
specimens (a) and as-received specimens (b): paper mill
116
for low-K
sludge B.
-------
I I——I—1—I—1——1—I—1—I——I—I—i—i——i—i—i—i—:
Specimen-1
Specimen-2
- Specimen-3
0 12 3 4
Number of Freeze-Thaw Cycles
I—1——I—I—I—I——i—i—I—I——i—i—i—I——I—i—I—i—
Specimen-1
Specimen-2
1 2 3 4
Number of Freeze-Thaw Cycles
5 :
Figure 8.9. Hydraulic conductivity vs. number of freeze-thaw cycles for low-K
specimens (a) and as-received specimens (b): paper mill sludge C.
117
-------
after each freeze-thaw cycle showed a slight decrease in average hydraulic
conductivity after exposure to freeze-thaw. :
All of the low-K specimens of sludge B showed an increase in hydraulic
conductivity of approximately 2 orders of magnitude after being exposed to freeze-
thaw (Fig. 8.8). One of the low-K specimens was frozen and thawed five times without
intermediary permeation. The hydraulic conductivity of this specimen also increased
approximately two orders of magnitude. Similar results were obtained for the other
two low-K specimens of sludge.B that were permeated after 0, 1, 2, 3, and 5 freeze-
thaw cycles. :
All of the low-K specimens of sludge C showed an increase in hydraulic
conductivity after exposure to freeze-thaw (Fig. 8.9). All of the low-K sludge C
specimens were permeated after each freeze-thaw cycle, up to fivejexcept for one
specimen which was not permeated after its third cycle. The average hydraulic
conductivity of the three low-K sludge specimens increased approximately one order
of magnitude as a result of freeze-thaw.
8.2.2.2 As-Received Molding Water Contents
One of the specimens of sludge A compacted at its as-received water content
showed a decrease in hydraulic conductivity after exposure to freeze-thaw. This
specimen was permeated after each freeze-thaw cycle. The other specimen of sludge
A was permeated after 0, 1, and 5 freeze-thaw cycles and showed a slight (1.6 times)
increase in hydraulic conductivity as a result of freeze-thaw.
The two specimens of sludge B compacted at the as-received water content
were permeated after each freeze-thaw cycle. These specimens, had a higher
average initial hydraulic conductivity than the low-K specimens, but had a similar
hydraulic conductivity (-4.0 x 10-8 m/s) after five freeze-thaw cycles. i '
Both of the specimens of sludge C compacted at the as-received water content
were permeated after each freeze-thaw-cycle. The initial hydraulic conductivity and
the hydraulic conductivity after five freeze-thaw cycles was essentially the same.
However, at an intermediate number of freeze-thaw cycles (1, 2, Is, and 4) the
hydraulic conductivities of the specimens varied; i.e., one specimen showed an
increase in hydraulic conductivity after two freeze-thaw cycles and no increase after
subsequent cycles, whereas the other specimen showed a steady increase in
hydraulic conductivity with each freeze-thaw cycle (Fig. 8.9). ]
118
-------
8.2.3 Effective Stress Tests I
Hydraulic conductivity were performed on one specimen of each sludge
compacted with standard proctor effort at the low-K water content (Sec. 8.2.1). The
specimens were placed in flexible-wall permeameters and permeated at effective
stresses of 7, 35, 46, and 81 kPa. Approximately one week of consolidation occurred
between each stage of permeation. Results of the effective stress tests are
summarized in Fig. 8.10. • ;
The hydraulic conductivity of all three sludges decreased' as a result of
increasing the effective stress. However, after an effective stress!of 46 kPa, the
subsequent decrease in hydraulic conductivity was small (Fig. 8.10). Similar behavior
(decreasing hydraulic conductivity with increasing effective stress) was observed for a
paper mill sludge tested by Zimmie et al. (1994).
The hydraulic conductivities of sludges A and C were essentially the same at
each effective stress, whereas the hydraulic conductivity of sludge B was slightly
higher. The hydraulic conductivity of each sludge was less than 1 x ib-9 m/s at every
effective stress tested, except for the specimen of sludge B permeated at an effective
stress of 7 kPa. ;
l
8.2.4 Long-Term Hydraulic Conductivity Tests I
Hydraulic conductivity was measured on two specimens of each paper mill
sludge compacted at standard Proctor effort over a period of at least 90 days to
determine whether or not biological decay of the paper mill sludge would adversely
impact hydraulic conductivity. One specimen of each sludge was compacted at the
low-K water content and the other was compacted at the as-received water content
(Sec. 4.4.2.2). Results of the long-term hydraulic conductivity tests are shown in Figs.
.8.11-8.13.
The two specimens of sludge B had higher hydraulic conductivity than the
specimens of sludges A and C compacted at their respective low-K and as-received
water contents. Examination of Figs. 8.11-8.13 shows that the hydraulic conductivity of
each specimen decreased slightly over the time interval during which the tests were
performed. Thus, biological decay appears to have no adverse impact on hydraulic
conductivity, at least for the test period used in this study.
119
-------
10"
C
o
o
I 10-10
10'1
Sludge A
Sludge B
Sludge C
J , . i
15 30 45 60
Effective Stress (kPa)
75
90
Fig. 8.10. Hydraulic conductivity vs. effective stress for paper mill sludges A, B, & C.
10"
•f 10*
1
c
o
O
I io-10
10'1
• "Low-K" Water Content
• "As-Received" Water Content
30
60 90
Time (days)
120
150
Figure 8.11. Hydraulic conductivity vs. time for paper mill sludge A.
120
-------
o
o
10
10-*
-7
10
,-10
"Low-K" Water Content
"As-Received" Water Content
' ' 1 ' ' i 1 1 1 1 L-
30
60 90
Time (days)
120 150
Figure 8.12. Hydraulic conductivity vs. time for paper mill sludge B.
1Q-8 F
i
f 10'9
•a
o
O
= ID'
10'1
10
30
"Low-K" Water Content
"As-Received" Water Content
60 90
Time (days)
120 150
Figure 8.13. Hydraulic conductivity vs. time for paper mill siujdge C.
121
-------
Hydraulic conductivity tests performed on specimens compacted in compaction
molds to construct the compaction curves for each sludge indicate that low hydraulic
conductivities (< 1 x 10'9 m/s) can be achieved for each sludge. ! Based on this
behavior, paper mill sludge could be used as an alternative to compacted clays for use
in landfill final covers. However, results of the hydraulic conductivity tests performed
on the specimens compacted and frozen in the laboratory indicate that the hydraulic
conductivity of some paper mill sludges (e.g., sludges B and C) can be affected by
exposure to freeze-thaw. Nevertheless, it appears that some sludges (e.g., sludge A)
may be frost resistant, depending on the hydrologic conditions to iwhich they are
exposed. For example, the hydraulic conductivity of sludge A depends on whether or
not it is permeated after each freeze-thaw cycle. That is, a decrease in hydraulic
conductivity occurs when the specimen is permeated, whereas an increase in
hydraulic conductivity occurs if it is not permeated. !
Effective stress tests performed on the three paper mill sludges indicate that the
hydraulic conductivity of the sludges decreases with increasing effective stress up to a
certain effective stress, after which no significant decrease in hydraglic conductivity
occurs. Hydraulic conductivities as low as 1 x 1CH0 m/s were observed for specimens
of sludges A and C at effective stresses greater than 46 kPa. Thus, extra overburden
placed to protect paper mill sludge from frost penetration used in final covers may also
reduce the hydraulic conductivity of the sludge substantially.
Long-term hydraulic conductivity tests performed on the1 three sludges
compacted at two different water contents indicate that the hydraulic conductivity
slightly decreased over the test period. The decrease in hydraulic conductivity is likely
the result of consolidation under the 7 kPa overburden stress induced by the lead
weight. Maltby and Eppstein (1994) observed consolidation and; a decrease in
hydraulic conductivity over time in their large-scale field study. Based on the results
obtained in this project and the experiments of others, paper mill sludges are likely to
perform well for a long period of time provided they are protected against detrimental
stresses such as freeze-thaw.
8.3 COMPARISON OF FIELD AND LABORATORY TEST RESULTS
Because the results of the hydraulic conductivity tests performed on the pipe
specimens and the specimens removed from the pipes were inconclusive, no direct
comparison between field and laboratory test results can be made.i However, the
122
-------
results of the field and laboratory tests performed on the paper mill: sludges warrant
discussion. !
The structure of the sludge within the pipe specimens was noted during slicing
and trimming of the specimens for hydraulic conductivity testing in flexible-wall
permeameters. All of the sludges that were exposed to freeze-thaw (field specimens)
were less homogeneous and had a more gravel-like structure than;the sludges that
had not been exposed to freeze-thaw (control specimens). The specimens would fall
apart when being trimmed for flexible-wall tests. Thus, based on visual observation,
freeze-thaw had affected the structure of the compacted paper mill sludges. However,
given the results of the hydraulic conductivity tests performed in the pipes after freeze-
thaw, it is not clear whether these structural changes, which were iobserved during
trimming, had adversely affected hydraulic conductivity. Thus, to ascertain the true
impact of freeze-thaw on the field hydraulic conductivity of paper mill sludges, field
tests similar to the one described by Maltby and Eppstein (1994) need to be
conducted. :
123
-------
SECTION 9
SUMMARY AND CONCLUSIONS
The objective of this study was to evaluate and compare the effect that freezing
and thawing has on the hydraulic conductivity of two clays and three alternative barrier
materials (a sand-bentonite mixture, three geosynthetic clay liners (GCLs), and three
paper industry sludges) under field and laboratory conditions. A battery of hydraulic
conductivity tests was performed on specimens prepared in the laboratory under
conditions which yield low hydraulic conductivity. The hydraulic conductivity of each
specimen was measured before and after exposure to a number of freqze-thaw cycles.
Results from the laboratory tests on the clays, sand-bentonite mixture, and GCLs were
then compared to hydraulic conductivities measured during the GOLDICE/CPAR
project, a recent large-scale field study conducted by the U. S. Army Corps of
Engineers Cold Regions Research and Engineering Laboratory (CRREL), CH2M Hill,
Inc. and a suite of industrial partners. Results of hydraulic conductivity tests performed
on the paper mill sludges were compared to results obtained from a small-scale field
study performed as part of this project.
!
. i
9.1 COMPACTED CLAYS j
Increases in hydraulic conductivity of two orders of magnitude;were observed
for clay specimens compacted and frozen and thawed in the laboratory, whereas
increases in hydraulic conductivity up to four orders of magnitude were observed for
specimens removed from the COLDICE test pads after winter. The observed increase
in hydraulic conductivity of the field and laboratory specimens was; attributed to a
macroscopic network of cracks caused by the formation of ice lenses land desiccation
induced by freezing. Extensive cracking of the soil was observed in the field and
laboratory experiments. In contrast, no increase in hydraulic conductivity was
observed for specimens removed from the tests pads after winter using thin-wall
sampling tubes. Apparently, disturbance during sampling and extrusion masked the
crack structure caused by freeze-thaw. !
Two factors that potentially could have caused the increase in hydraulic
conductivity to be larger in the field than in the laboratory were: (1) the dimensionality
of freezing (e.g., one-dimensional vs. three-dimensional) and (2) that the in situ
freezing rate was smaller than that used in the standard (three-dimensional) laboratory
124
-------
tests. Thus, specimens were frozen in the laboratory one-dimensionally at controlled
freezing rates to assess whether dimensionality of freezing or freezing rate affected the
increase in hydraulic conductivity. Results of the tests showed that there was no
relationship between hydraulic conductivity and freezing rate. Also, the freezing rates
occurring in the field were smaller than the lowest freezing rates; that could be
simulated in the laboratory, whereas the increase in hydraulic conductivity measured
in the field was two orders, of magnitude greater than the increase in hydraulic
conductivity measured on specimens frozen in the laboratory. Thus, dimensionality of
freezing and freezing rate were probably not the cause of the discrepancy between the
laboratory- and field-measured hydraulic conductivities. Differences in soil structure
prior to freezing may have been responsible for the differences in the field and
laboratory behavior. However, sufficient data were not available tq evaluate this
hypothesis.
9.2 SAND-BENTONITE MIXTURE . !
No increase in hydraulic conductivity was observed for the specimens of sand-
bentonite compacted and frozen -three-dimensional.!/ in the laboratory. Low post-
winter hydraulic conductivities were also observed in the field and measured in the
laboratory for specimens removed from the test pad using thin-wall sampling tubes.
The field hydraulic conductivity and the hydraulic conductivities of the specimens
collected in sampling tubes were slightly higher than the hydraulic conductivity of the
laboratory-compacted specimens. The difference in hydraulic conductivity was
attributed to heterogeneities in the sand-bentonite mixture and, for the specimens
collected in tubes, to disturbance during extrusion.
In contrast, high hydraulic conductivities were measured for all but one of the
block specimens removed from the sand-bentonite test pad before winter and on all of
the block specimens removed after winter. Examination of the structure of the block
specimens revealed bentonite-rich zones. Based on this observation, the high
hydraulic conductivities of the block specimens were attributed to insufficient mixing
and not to environmental stresses, such as freeze-thaw. \
Examination of the soil within the box infiltrometer installed in the sand-
bentonite test pad showed that the soil had retained water and that a network of cracks
was non-existent. Similarly, examination of the internal structure iof specimens
removed from the test pad after winter using tubes and the laboratory-compacted
specimens exposed to freeze-thaw revealed none of the cracks typically observed in
125 . :
-------
compacted clays exposed to freeze-thaw. The absence of cracks is consistent with the
low hydraulic conductivities measured for the sand-bentonite specimens.
9.3 GEOSYNTHETIC CLAY LINERS (GCLs)
Laboratory tests performed on three GCLs (Bentofix®, B^entomat®, and
Claymax®) showed that their hydraulic conductivity was low (typically: about 2 x 10'11
nn/s) and that the hydraulic conductivity decreased slightly as a result of freeze-thaw.
Results of t-tests indicated that the slight decrease in hydraulic conductivity was
statistically significant.
Low hydraulic conductivity was also measured on large undisturbed specimens
removed from the COLDICE GCL test ponds after exposure to two winters of freeze-
thaw. Hydraulic conductivities less than 1 x 10-9 m/s were measured for all but one
specimen. The high hydraulic conductivity of that specimen was attributed to
disturbance during handling. Similar results were reported by Erickson et al. (1994)
for the GCLs tested in the COLDICE test pans. For each of the COLDICE tests,
hydraulic conductivities less than 1 x 10-9 m/s were measured before and after
exposure to freeze-thaw. Erickson et al. (1994) did report, however, that the field
hydraulic conductivities of the seamed GCLs were slightly higher than those for the
unseamed GCLs. No seamed specimens were tested in this study. Additional study of
the hydraulic integrity of GCL seams exposed to freeze-thaw is recommended.
9.4 PAPER MILL SLUDGES !
Results of hydraulic conductivity tests performed on paper mill sludges showed
that conductivities less than 1 x 10-9 m/s can be achieved for each type of sludge when
compacted wet of optimum water content. Slightly lower hydraulic conductivities were
obtained for sludges A and C (the two combined sludges) relative to sludge B (a
primary sludge). The lower hydraulic conductivities of sludges A and C :were attributed
to the existence of additional biological material from secondary wastewater treatment
processes.
Freeze-thaw affected each sludge differently. For specimens of sludge A
compacted at the low-K water content, no change in hydraulic conductivity was
observed for specimens permeated after each freeze-thaw cycle, whereas an increase
in hydraulic conductivity was observed for specimens that were not permeated
between freeze-thaw cycles. Also, for specimens of sludge A compacted at the as-
126
-------
received water content, no increase in hydraulic conductivity was observed after
exposure to freeze-thaw. . ''• • '
Increases in hydraulic conductivity of approximately one order of magnitude
were observed for all specimens of sludge B, regardless of the molding water content
or whether the specimen was permeated between freeze-thaw cycles.; This increase
in hydraulic conductivity is similar to the increase measured for the compacted clays
studied in this project. Similar behavior was observed for sludge C. :
Results of the hydraulic conductivity tests conducted on the pipe specimens
were inconclusive; a decrease in hydraulic conductivity after freeze-thaw was
observed when the sludges were permeated in the pipes, whereas ian increase in
hydraulic conductivity was observed in the slices tested in flexible-wall permeameters.
Examination of the sludge in the pipes revealed that exposure to freeze-thaw resulted
in a blocky structure that fell apart easily during trimming. Nevertheless,' it could not be
determined whether formation of the blocky structure had an adverse impact on the
hydraulic conductivity of the sludges. Large-scale field tests including morphological
investigations are recommended to fully assess the impact of freeze-thaw on the
hydraulic conductivity of paper mill sludges.
127
-------
REFERENCES
^^ o,
^
J. Trautwein, Eds., Amer. Soc. for Test, and Mat'ls., Philadeiphfa PA P
Chamberlain, E. J., Erickson, A. E.( and Benson, C. H. (1995) "Effects
'
128
-------
Estorne'1' P- M- 0990), "Compilation of information on alternative
EPA 600^002,^ Environment
Mad,son Waste Conference, Sept. 21-22, Dept. of Engineering ProfesS
Development, University of Wisconsin-Madison, Madison Wl. ng rr°Tessional
oK^
solWsoonnrMTh and 9< *•
sows content, M. S. Thesis, University of Wisconsin-Madison, Madison, Wl.
f Bentomat'"
i
HaU9> ^n?" *dtW°n9' L c- (1992), "Impact of molding water content on hydraulic
conductivity of compacted sand-bentonite," Can. Geotech. J., Vol. 29, pj ,. Iss?
Kenney, TO. van Veen, W. A., Swallow, M. A., and Sungalia, M. A. (1992) "Hydraulic
pp 364-374 C0mpacted bentonite-sand mixtures," Can. Geotech. 'j. Vol 29?
Kim, W. and Daniei/D. E. (1992), "Effects of freezing on the hydraulic oonductivitv of
compacted clay,'' J. of Geotech. Engrg., ASCE, Vol. Ill, X 7 ^ PP 1083 ?1097
Maltby ,C. V,, and I Eppstein L K. (1993), "A field-scale study of the! use of paper
industry sludges as hydraulic barriers in landfill cover systems " Hvdrauffc
ConductMty and Waste Contaminant Transport in Soils^ASTM STP1142
E J- Tr Py
Mitchell, J. K., Hooper, D. R., and Campanella, R. G (1965) "
' " Mechanics and
129
-------
NCASI (1989), "Experience with and laboratory studies of the use of pulp and paper
mill solid wastes in landfill cover systems," National Council of the Pulp and
Paper Industry for Air and Stream Improvement (NCASI) Technical Bulletin No.
559, New York, NY.
NCASI (1992), "Chemical composition of pulp and paper industry landfill leachates,"
National Council of the Pulp and Paper Industry for Air and Stream
Improvement (NCASI) Technical Bulletin No. 643, New York, NY.
Othman, M. A. (1992), "Effect of freeze-thaw on the structure and hydraulic conductivity
of compacted clays," Ph.D. Dissertation, University of Wisconsin-Madison,
Madison, Wl. '
Othman, M. A., and Benson, C. H. (1991), Influence of freeze-thaw on the hydraulic
conductivity of a compacted clay," Proc., 14th International Madison Waste
Conference, Sept. 25-26, Dept. of Engineering Professional Development,
University of Wisconsin-Madison, Madison, Wl.
Othman, M. A., and Benson, C. H. (1993a), "Effect of freeze-thaw on the hydraulic
conductivity of three compacted clays from Wisconsin," Transportation Research
Record, No. 1369, Transportation Research Board, Washington, DC., pp. 118-
125.
Othman, M. A., and Benson, C. H. (1993b), "Effect of freeze-thaw o!n the hydraulic
conductivity and morphology of compacted clay," Can. GeotechJ J., Vol. 30, No.
2, pp. 236-246.
Othman, M. A., Benson, C. H., Chamberlain, E. J., and Zimmie; T. F. (1994),
"Laboratory testing to evaluate changes in hydraulic conductivity caused by
freeze-thaw: State-of-the-art," in Hydraulic Conductivity and Waste Contaminant
Transport in Soils, ASTM STP 1142, David E. Daniel and Stephen J. Trautwein,
Eds., Amer. Soc. for Test, and Mat'ls, Philadelphia, in press. :
Robert L. Nelson and Associates, Inc. (1993), "Report of Bentomat® freeze/thaw test
results," report to CETCO, Inc. by Robert L. Nelson and Associates, Inc.,
Schaumburg, IL. • ; •
Shan, H. Y. (1990), "Laboratory tests on bentonitic blanket," M. S. Thesis, University of
Texas, Austin, TX.
Shan, H. Y., and Daniel, D. E. (1991), "Results of laboratory tests on a
geotextile/bentonite liner material," Geosynthetics '91, Industrial Fabrics
Association International, St. Paul, MN, Vol. 2, pp. 517-535.
Taylor, K. R., Hansen, J. S., and Andrews, D. W. (1994), "The potential
use of pulp and
paper mill sludge in landfill closure," Proc., Practical Application of Soil Barrier
Technology, Maine Section ASCE 1994 Technical Seminar.
130
-------
Wong, L., and Haug, M. (1991), "Cyclical closed-system freeze-thaw permeability
testing of soil liner and cover materials," Can. Geotech. J., Vol. 28, No. 6, pp.
784-793. ;
Zirnmie, T. F., and LaPlante, C. (1990), "The effect of freeze-thaw cycles on the
permeability of a fine-grained soil," Proc., 22nd Mid-Atlantic Industrial Waste
Conference, Philadelphia, PA, July 24-27, pp. 580-593.
Zirnmie, T., Moo-Young, H., and LaPlante, K. (1993), "The Use of Waste Paper Sludge
for Landfill Cover Material," Green '93: Waste Disposal by Landfill, An
International Symposium on Geotechnics related to the Environment, Bolton
Institute, Bolton, UK., Vol. 2, pp. 11-19.
131
-------
APPENDIX A
1 • - i
RESULTS OF COMPACTION AND HYDRAULIC CONDUCTIVITY TESTS
PERFORMED ON LABORATORY-COMPACTED SPECIMENS
132
-------
Table A.1. Results of compaction and hydraulic conductivity tests performed on
Parkview clay. :
Specimen^)
PV-S-1
PV-S-2
PV-S-3
PV-S-4
PV-S-5
PV-S-6
PV-S-7
PV-S-8
PV-S-9
PV-S- 10
PV-S-11
PV-S- 12
PV-M-1
PV-M-2
PV-M-3
PV-M-4
PV-M-5
PV-M-6
PV-M-7
PV-M-8
PV-M-9
PV-M-10
PV-M-1 1
Molding Water
Content
(%)
10.5
10.6
10.7
12.1
13.4
13.4
15.3
16.1
16.3
17.7
19.6
20.1
4.1
4.5
5.3
7.4
7.9
10.9
12.8
13.9
14.0
18.3
19.1
Dry Unit
Weight
(kN/m3)
17.06
17.30
16.81
17.37
18.52
18.69
18.25
18.34
18.33
17.72
17.07
17,01
19.45
19.27
19.63
20.70
20.47
20.63
19.73
19.37
19.63
17.48
17.34
Hydraulic
Conductivity
;(m/s)
2.1 x 10-7
7.6 x 10'7
1.7 x10'7
2.1 x10-7
2.0x1 0'8
2.1 x10'8
4.5 X10'9
2.4 x10-10
2.6 x10-10
4.9 x10-10
1.1 x10-9
3.9 x10-10
TNP(2)
^TNP
5:1 X10'9
:TNP
1.4x10-9
5.7 x10-11
' TNP
5.1 x10-10
TNP
TNP
3.1 x10-10
Notes: „
1. PV = Parkview clay; -S = Compacted at standard Proctor effort; -M = Compacted jat modified Proctor
effort;-1..-12 = Specimen number.
2. TNP = Test not performed. . <
133
-------
Table A.2. Results of compaction and hydraulic conductivity tests performed on
Valley Trail clay.
SpecimenO)
VT-S-1
VT-S-2
VT-S-3
VT-S-4
VT-S-5
VT-S-6
VT-S-7
VT-S-8
VT-S-9
VT-S-10
VT-S-11
VT-S-1 2
VT-S-1 3
VT-S-1 4
VT-S-1 5
VT-M-1
VT-M-2
VT-M-3
VT-M-4
VT-M-5
VT-M-6
VT-M-7
VT-M-8
VT-M-9
Molding Water
Content
(%)
10.4
12.9
14.4
14.7
16.1
16.1
19.0
19.3
19.4
20.2
20.2
20.7
20.8
22.7
23.3
7.8
11.3
11.4
13.4
13.5
13.7
17.0
17.8
18.9
Dry Unit
Weight
(kN/m3)
16.18
15.77
16.37
16.66
17.12
16.71
16.87
16.82
17.18
17.04
17.03
16.54
16.68
16.13
15.85
18.39
18.63
18.69
18.97
19.01
19.05
18.20
18.22
17.86
Hydraulic
Conductivity
, (m/s)
1:6x10-8
9.5x10-8
3.4x10-8
6.9x10-8
7.5 x 10-8
1'.4x10-8
1.6x10-9
1;.8x10-9
7.|2x10-11
7.6 x10-11
8.3 x10-11
1.2 x10-10
2:2 x10-10
2J5x10-10
3.9 x10-10
4.8x10-9
2.6 x 10-9
TNP(2)
TNP
TNP
2.1 x10-10
8.1 x10-11
TNP
TNP
Notes:
1. VT = Valley Trail clay; -S = Compacted at standard Proctor effort; -M = Compacted at modified Proctor
effort; -1 ..-15 = Specimen number.
2. TN P = Test not performed.
134
-------
Table A.3. Results of compaction and hydraulic conductivity tests performed on
sand-bentonite. :
Specimen^)
SB-S-1
SB-S-2
SB-S-3
SB-S-4
SB-S-5
SB-S-6
SB-S-7
SB-S-8
SB-S-9
SB-S-1 0
SB-S-11
SB-S-12
SB-S-13
SB-S-1 4
SB-S-1 5
SB-S-1 6
SB-S-1 7
SB-S-1 8
SB-S-19
SB-S-20
SB-M-1
SB-M-2
SB-M-3
SB-M-4
SB-M-5
SB-M-6
SB-M-7
SB-M-8
SB-M-9
SB-M-1 0
SB-M-1 1
SB-M-12
SB-M-1 3
SB-M-1 4
Molding Water
Content
(%)
7.9
8.1
8.4
11.4
12.7
13.9
15.1
15.9
16.5
16.7
16.7
16.7
16.7
16.7
16.7
16.7
16.7
19.3
19.7
20.3
7.1
7.1
7.3
10.0
10.6
10.7 •
12.3
12.4
13.4
15.4
16.3
16.5
17.9
18.4
Dry Unit
Weight
(kN/m3)
17.05
16.88
16.99
17.17
17.06
17.26
17.23
17.09
17.31
17.50
17.26
17.31
17.37
17.47
17.28
17.39
17.50
16.81
16.70
16.62
18.01
18.71
18.57
18.79
18.75
18.63
18.85
18.85
18.68
18.13
17.84
18.09
17.36
17.14
Hydraulic
Conductivity
! (m/s)
1.7x10'10
TNP(2)
TNP
iTNP
iTNP
1.7X10'10
TNP
TNP
1.2 x10-10
:TNP
TNP
TNP
TNP
TNP
iTNP
TNP
TNP
1.0 x10-10
:TNP
TNP
'TNP
'TNP
5.3x10-8
TNP
!TNP
6.9 x10'11
TNP
:TNP
5.9 x10-11
TNP
4.3 x10-11
TNP
TNP
TNP
Notes:
1. SB = Sand-bentonite; -S = Compacted at standard Proctor effort; -M = Compacted at modified Proctor
effort; -1 ..-20 = Specimen number.
2. TNP = Test not performed.
135
-------
Table A.4. Results of compaction and hydraulic conductivity tests performed on
paper mill sludge A.
SpecimenO)
SLA-S-1
SLA-S-2
SLA-S-3
SLA-S-4
SLA-S-5
SLA-S-6
SLA-S-7
SLA-S-8
SLA-S-9
SLA-S-10
SLA-S-11
SLA-S-1 2
SLA-S-1 3
SLA-S-1 4
SLA-S-1 5
SLA-S-1 6
SLA-S-1 7
SLA-S-1 8
SLA-S-1 9
SLA-S-20
SLA-S-21
SLA-S-22
SLA-S-23
SLA-S-24
SLA-S-25
SLA-S-26
SLA-S-27
Molding Water
Content
(%)
17.0
31.4
43.6
47.2
58.7
72.4
72.8
82.8
85.6
88.3
89.0
89.6
90.4
91.2
92.3
93.1
93.3
94.8
95.9
97.4
102.1
122.1
149.2
150.0
152.8
185.1
185.3
Dry Unit
Weight
(kN/m3)
8.32
8.68
8.66
8.54
8.41
7.53
7.61
7.22
7.25
6.73
6.68
6.67
6.68
6.62
6.60
6.56
6.54
6.48
6.63
6.41
6.41
5.61
4.78
4.78
4.72
3.98
3.96
Hydraulic
Conductivity
: (m/s)
6.6x10-6
3.7X10-6
TNP(2)
1,3x10-6
1.0 x10-9
7.2 x10-10
2.5x10-9
6.2 x10-10
5.5 x10-10
'TNP
TNP
,TNP
TNP
TNP
TNP
2.8 x 10-™
7.0 x10-10
TNP
8.0 x10-10
3.3 x10-10
4.4 x TO-10
1.2x10-9
2.3x10-9
4.0 x10-10
2.3 x 10-9
TNP
,TNP
Notes:
1. SLA = Paper mill sludge A; -S = Compacted at standard Proctor effort; -1 ..-27 = :
2. TNP = Test not performed.
Specimen number.
136
-------
Table A.5. Results of compaction and hydraulic conductivity tests performed on
paper mill sludge B. :
Specimen^)
SLB-S-1
SLB-S-2
SLB-S-3
SLB-S-4
SLB-S-5
SLB-S-6
SLB-S-7
SLB-S-8
SLB-S-9
SLB-S-1 0
SLB-S-11
SLB-S-12
SLB-S-13
SLB-S-14
SLB-S-15
SLB-S-16
SLB-S-17
SLB-S-1 8
SLB-S-1 9
SLB-S-20
SLB-S-21
SLB-S-22
SLB-S-23
SLB-S-24
SLB-S-25
Molding Water
Content
(%)
13.3
57.3
86.7
89.8
115.9
116.9
117.7
119.0
121.1
121.7
123.7
123.9
124.1
124.5
124.7
126.8
126.8
200.1
239.5
252.9
255.9
256.0
259.7
259.9
262.7
Dry Unit
Weight
(kN/m3)
5.60
5.83
5.91
5.94
5.77
5.67
5.74
5.60
5.56
5.58
5.49
5.47
5.47
5.49
5.52
5.45
5.41
3.73
3.27
3.13
3.11
3.16
3.10
3.11
3.10
Hydraulic
Conductivity
;(m/s)
1.9x 10-5
1.5x10-5
1.0x 10-9
3.2 x 10-9
TNP(2)
1.4x 10-10
TNP
TNP
TNP
TNP
1.0x10-9
9.6 x10-10
TNP
TNP
3.0x10-9
TNP
5.9 x 10-9
1.8x 10-9
TNP
4.1 x 10-9
4.6 x 10-9
5.0 x 10-9
TNP
TNP
TNP
Notes:
1. SLB = Paper mill sjudge B; -S = Compacted at standard Proctor effort- -1 -25 = :
2. TNP = Test not performed.
Specimen number.
137
-------
Table A.6. Results of compaction and hydraulic conductivity tests performed on
paper mill sludge C. ; •
Specimen^)
SLC-S-1
SLC-S-2
SLC-S-3
SLC-S-4
SLC-S-5
SLC-S-6
SLC-S-7
SLC-S-8
SLC-S-9
SLC-S-10
SLC-S-1 1
SLC-S-12
SLC-S-13
SLC-S-14
SLC-S-15
SLC-S-16
SLC-S-1 7
SLC-S-18
SLC-S-19
SLC-S-20
SLC-S-21
SLC-S-22
Content
(%)
5.8
33.9
55.4
97.3
120.9
121.0
121.7
124.4
126.4
132.2
135.2
135.5
135.9
145.0
156.0
159.0
165.0
189.9
191.6
207.5
226.2
227.9
Dry Unit
Weight
(kN/m3)
5.64
6.24
6.13
5.31
5.49
5.47
5.47
5.36
5.30
5.22
5.00
5.12
5.11
4.78
4.51
4.64
4.37
• 3.84
3.77
3.54
3.40
3.36
Hydraulic
Conductivity
: (m/s)
1.4x10-5
1 iO x 1 0'5
8.7x10-6
7iOx10-7
TNP(2)
TNP
ITNP
TNP
TNP
4.6 x10'10
8.9 x10-10
;TNP
: TNP
9.4 x10'10
8.9 x10-10
1.2x10-9
1.6x 1C-9
1.3x1 0-9
9.8 x 10-10
TNP
7.0 X10'10
TNP .
Notes:
1 ' ™£ = £aper mil1 sludge C; ~S =
2. I NP = Test not performed.
at standard Proctor effort; -1 ..-22 = Specimen number
'
138
-------
APPENDIX B
GRAPH OF AIR TEMPERATURE VS. TIME FOR THE WINTER OF
1993-94 FOR THE BURIED SMALL-SCALE FIELD PIPE SPECIMENS
139
-------
20.0
O
o_
2
I
-30.0 -
-40.0
20 40 60 80 100
Time (days since December 15, 1993)
120
Figure B.1. Air temperature vs. time during freezing of buried small-scale field pipe
specimens. ;
140
------- |