Citation
Volume controlled compatibility measurements in a triaxial system

Material Information

Title:
Volume controlled compatibility measurements in a triaxial system
Creator:
Allerton, David Kelley
Publication Date:
Language:
English
Physical Description:
xiv, 174 leaves : illustrations (some color) ; 29 cm

Subjects

Subjects / Keywords:
Sanitary landfills -- Linings ( lcsh )
Clay soils ( lcsh )
Hazardous wastes ( lcsh )
Clay soils ( fast )
Hazardous wastes ( fast )
Sanitary landfills -- Linings ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 168-174).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Civil Engineering.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by David Kelley Allerton.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
34889253 ( OCLC )
ocm34889253
Classification:
LD1190.E52 1995m .A45 ( lcc )

Full Text
VOLUME CONTROLLED COMPATIBILITY MEASUREMENTS
IN A TRIAXIAL SYSTEM
fey
David Kelley Allerton
B.A., University of Colorado at Denver, 1981
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering
1995


This thesis for the Master of Science
degree by
David Kelley Allerton
has been approved
by
Date


Allerton, David Kelley (M.S., Civil Engineering)
Volume Controlled Compatibility Measurements in a Triaxial System
Thesis directed by Professor N.Y. Chang
ABSTRACT
A compatibility testing apparatus was developed for investigating low
hydraulic conductivity engineered barriers. This apparatus was developed utilizing
current flow pump technology, but including a system for fluid exchange. The
apparatus is capable of continuously and simultaneously monitoring hydraulic
conductivity and effective stress at the base of a soil specimen under constant volume
and constant stress conditions. In addition, flow rates can be controlled, hydraulic
gradients and intrinsic permeability determined. Effluent samples can be removed
with minimal sample disturbance, and chemical concentrations determined in order
to examine sorption properties of natural and admixed soils.
These capabilities were demonstrated on two types of clay soils. First, the
clay mineral palygorskite was admixed with a sand fill material and permeated with
demineralized water followed by acetone, methanol, or industrial waste leachate.
Secondly, a natural soil was admixed with a clinoptilolite (a zeolite mineral) rich rock
to enhance its cation exchange capacity and permeated with the industrial waste
leachate to examine its sorption characteristics.
The palygorskite admixtures were most affected by acetone permeation,
followed by methanol, and the industrial waste leachate had the least effect. The
natural soil admixed with clinoptilolite rich rock was effective in adsorbing Ca, K,
and Mg while desorbing Na. The effectiveness of the admixture for capturing heavy
metals could not be measured because their concentrations in the leachate were below
detection limits. Scanning electron microscopy of freeze dried soil specimens
m


supports the assertion that channel flow existed during testing.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
IV


DEDICATION
I would like to dedicate this work to my lovely wife Sarah, who encouraged
and supported me throughout the research and writing process of this thesis. Also,
to the memory of my grandfather, Earl Monroe Kelley, whose attic laboratory,
complete with corroding door knobs and window screens, aroused my interest in
minerals and science.


CONTENTS
Page
Abstract.................................................................iii
Dedication.................................................................v
Tables.....................................................................x
Figures...................................................................xi
Symbols................................................................ xiii
Acknowledgements .........................................................xv
Chapter
1. Introduction ...........................................................1
1.1 Problem statement ..................................................1
1.2 Significance of investigation ......................................1
1.3 Objectives .........................................................3
2. Review of Laboratory Methods of
Hydraulic Conductivity Measurements....................................4
2.1 Introduction ...................................................... 4
2.2 Darcys Law.........................................................4
2.3 Falling Head Method.................................................5
2.4 Constant Head Method................................................8
2.5 Flow pump Method....................................................8
2.6 Permeameters ......................................................11
3. Review of Scanning Electron Microscopy
of Clay-Soil Textures.................................................13
3.1 Introduction ......................................................13
3.2 Scanning Electron Microscopy.......................................13
3.2.1 Energy Dispersive X-ray Spectrometry...............................16
vi


3.2.2 Sample Preparation.....................................................19
3.2.2.1 Air Drying...........................................................19
3.2.2.2 Critical Point Drying................................................19
3.2.2.3 Freeze Drying........................................................22
3.2.2.4 Impregnation.........................................................23
3.2.2.5 Surfaces for Study ..................................................23
3.2.3 SEM Preparation Techniques.............................................24
3.2.3.1 Mounting.............................................................24
3.2.3.2 Coating Materials....................................................24
3.3 Clay Mineralogy .......................................................25
3.3.1 Kaolinite group .......................................................28
3.3.2 Smectite group.........................................................28
3.3.3 Palygorskite group ....................................................29
3.4 Clay microtexture......................................................32
4. Clay Liners in Landfill Applications......................................37
4.1 Introduction ..........................................................37
4.2 Design of Clay Liners..................................................37
4.2.1 Soil liner composition ................................................39
4.3 Double layer theory ...................................................40
4.4 Effects of chemicals on clay liners....................................43
4.4.1 Acidic conditions......................................................43
4.4.2 Alkaline conditions.................................................. 45
4.4.3 Neutral inorganic and organic fluids...................................45
5. Materials and Methods.....................................................47
5.1 Introduction ..........................................................47
5.2 Materials............................................................. 47
5.3 Mineralogical Analyses ................................................48
5.4 Geotechnical Index Tests ..............................................48
5.5 New Compatibility Testing Method.......................................49
5.5.1 Sample preparation.....................................................49
5.5.2 Compatibility Testing apparatus .......................................50
5.5.3 Compatibility Testing procedure........................................54
5.6 Freeze Drying Method...................................................57
5.6.1 Freeze drying apparatus................................................57
5.6.2 Freeze drying procedure................................................57
5.7 Scanning Electron Microscopy...........................................60
5.7.1 Scanning Electron Microscope and equipment.............................60
vii


5.7.2 Scanning Electron Microscopy procedure..................................60
5.8 Chemical Analyses.......................................................61
5.9 Viscosity and Specific gravity .........................................62
6. Results ...................................................................63
6.1 Introduction ...........................................................63
6.2 Mineralogical Analyses .................................................63
6.3 Geotechnical Index Properties .........................................64
6.4 Hydraulic conductivity of volume control specimens................67
6.5 Compatibility Tests.....................................................69
6.5.1 Scope of compatibility tests............................................69
6.5.2 Compatibility test data for palygorskite admixtures.....................69
6.5.3 Compatibility test data for natural soil/zeolite admixture..............74
6.6 Scanning Electron Microscopy Analyses ..................................77
7. Discussion of Test Results.................................................79
7.1 Introduction ...........................................................79
7.2 Mineralogical Analyses .................................................79
7.2.1 Palygorskite admixture..................................................79
7.2.2 Natural soil/zeolite admixture..........................................79
7.3 Geotechnical Index Tests ...............................................80
7.3.1 Palygorskite admixtures.................................................80
7.3.2 Natural soil/zeolite admixture......................................... 80
7.4 Hydraulic conductivity of volume control specimens................81
7.5 Compatibility Tests.....................................................82
7.5.1 Palygorskite admixtures.................................................82
7.5.1.1 Acetone...............................................................83
7.5.1.2 Methanol............................................................. 84
7.5.1.3 Industrial waste leachate.............................................85
7.5.2 Natural soil/zeolite admixture..................................... . 86
7.5.2.1 Industrial waste leachate.............................................86
7.6 Scanning Electron Microscopy............................................89
8. Conclusions and Recommendations ...........................................90
Appendixes.....................................................................94
A. Compatibility Test Data.....................................................94
Vlll


B. Scanning Electron Microscope Photomicrographs......................... 141
C. Job Safety Analysis................................................... 160
Procedure for Compatibility Testing Apparatus
D. Job Safety Analysis................................................... 164
Preparation of Soil Samples for the Scanning
Electron Microscope by Freeze Drying
References ............................................................... 168
IX


TABLES
Table Page
2.1 Advantages and disadvantages of rigid and flexible
wall permeameters ....................................................12
3.1 Shrinkage comparisons.................................................20
4.1 Liner/Industrial waste compatibility..................................41
6.1 Geotechnical index properties for the sand fill and
palygorskite admixtures ..............................................65
6.2 Geotechnical index properties for the Highlands Ranch clay
and the Highlands Ranch clay plus
South Dakota zeolite admixture........................................66
6.3 Properties of permeant fluids.........................................69
6.4 Summary of compatibility test data for the 10%
palygorskite admixed soil.............................................72
6.5 Summary of compatibility test data for the 20%
palygorskite admixed soil.............................................73
6.6 Compilation of chemical data for the HRC + 10%
SDA permeated with industrial waste leachate..........................76
6.7 Net gains (+) and losses (-) of elements for the
top, middle, and bottom segments of the
soil specimen ........................................................77
x


FIGURES
Figure Page
2.1 Diagrammatic representation of Darcys Law.................................6
2.2a Diagrammatic representation of the falling head test.......................6
2.2b Diagrammatic representation of the constant head test......................6
2.3 Diagrammatic representation of a simple constant flow rate test.........10
3.1 Schematic diagram of SEM/EDX system ......................................14
3.2 Electron transitions in an atom...........................................17
3.3 Critical point drying process ............................................21
3.4 Kaolinite structure ......................................................26
3.5 Pyrophyllite structure....................................................26
3.6 Montmorillonite structure.................................................30
3.7 Palygorskite structure ...................................................31
3.8 Clay fabrics ........................................................ 33-36
4.1 Single clay liner system..................................................38
4.2 Double liner with leak detection system...................................38
4.3 Typical variation in k for sample of compacted clay
permeated with concentrated acid ........................................44
5.1 Compatibility testing apparatus...........................................51
5.2 Photographs of compatibility testing apparatus.
A) Complete system showing pressure panel, compatibility
xi


testing apparatus, and data acquisition system, B) Compatibility
testing apparatus, C) Close up of flow pump with infuse/withdraw
actuator, bladder accumulators, D) Close up of triaxial cell
and differential pressure transducers................................52-53
5.3 Freeze drying apparatus.................................................58
5.4 Photograph of freeze drying apparatus ................................59
6.1 Plot of hydraulic conductivity and effective stress versus time for
the 30% palygorskite admixture under volume control conditions .... 67
6.2 Plot of hydraulic gradient and flow rate versus time for the 30%
palygorskite admixture under volume control conditions................68
6.3 Plot of hydraulic gradient versus flow rate for the 30%
palygorskite admixture under volume control conditions..................68
6.4 Hydraulic conductivity and effective stress for the 10%
palygorskite admixture permeated with acetone under
"constant volume" conditions............................................70
6.5 Hydraulic gradient and flow rate for the 10%
palygorskite admixture permeated with acetone under
"constant volume" conditions............................................71
6.6 Intrinsic permeability for the 10% palygorskite
admixture permeated with acetone under
"constant volume" conditions............................................71
6.7 Hydraulic conductivity and effective stress versus time
and pore volumes for the HRC + 10% SDA permeated with
industrial waste leachate...............................................74
6.8 Breakthrough curves for Na, Ca, K, and Mg .......................75
6.9 Breakthrough curves for Fe, As, Pb, Zn, and Cu......................75
6.10 Photomicrographs of palygorksite admixed soil after
permeation with demineralized water, frozen in liquid
nitrogen, and freeze dried..............................................78
xii


SYMBOLS
Symbol
Q flow rate
k hydraulic conductivity
i hydraulic gradient
A area
Ah change in head
AL change in length
K intrinsic permeability
7 unit weight of fluid
V- absolute or dynamic viscosity
P mass density of liquid
g acceleration due to gravity
U kinematic viscosity
a area of column
V volume
1 length
Tw unit weight of water
x2 chi square analysis
y unknown intensity
X standard intensity
d double layer thickness
o dielectric constant of vacuum
D dielectric constant of fluid
k Boltzman constant
T temperature
X1U


nc electrolyte concentration
e electronic charge
v cation valence
C concentration of cation in effluent
CD concentration of cation in influent
k baseline hydraulic conductivity of demineralized water
kf hydraulic conductivity of permeant
K,, intrinsic permeability of demineralized water
Kf intrinsic permeability of permeant
aj initial effective stress
Of final effective stress
i0 initial hydraulic gradient with demineralized water
if final hydraulic gradient with permeant
xiv


ACKNOWLEDGEMENTS
I wish to express my gratitude to Dr. N.Y. Chang of the University of
Colorado at Denver (UCD) for his assistance throughout the preparation of this
thesis. Special thanks are extended to Dr. Harold W. Olsen of the United States
Geological Survey (U.S.G.S.) for help with the initial design of the compatibility
apparatus, and numerous discussions thereafter. I would also like to acknowledge the
following people: Mike Higgins of Joe Caesar and Assoc., for collecting the natural
soil used in this investigation; Ken Esposito for his help in the Clay Mineralogy
laboratory of the U.S.G.S.; Dr. George A. Desborough of the U.S.G.S. for his
assistance in the selection of the most appropriate zeolite deposit to use; Sarah
Allerton of the U.S.G.S. for assistance in proofreading and editing the manuscript;
and Mike Gansecki of the United States Environmental Protection Agency for
supplying me with contacts for potential sources of landfill leachate.
This project was primarily funded through the University of Colorado at
Denver. Special thanks to Dr. Gary Winkler, Branch Chief, Branch of Central
Mineral Resources of the U.S.G.S. for believing in, and providing partial funding for
the project. The Center for Environmental Geochemistry and Geophysics of the
U.S.G.S. also provided partial funding. I would also like to thank the Milwhite
Mining Co. in Attapulgus, Georgia for donating the palygorskite used in the project.


1. Introduction
1.1 Problem Statement
Clay soils used in liner systems for municipal landfills and hazardous waste
impoundments are exposed to numerous types of chemicals. These liners must not
only achieve a hydraulic conductivity of 1 x 10'7 cm/sec, but also be capable of
containing any leakage of leachate which may occur through the primary containment
liner system. Chemical reactions between the clay liner material and the leachate
thus play an important role in how effective the overall liner system will function.
It is for this reason that many studies have been undertaken to determine the
effects of different chemicals on both geosynthetic and clay liners. These studies will
help to determine the most suitable material available for use in a clay liner system
in order to mitigate environmental damage from potential leachate leakage.
1.2 Significance of Present Investigation
Numerous studies have been performed to determine effects of chemicals on
engineering properties of different types of soils and soils amended with various clay
minerals. The majority of clays studied in this context include bentonite (Kenney,
etal, 1992; Alther, etal, 1985; Anderson, etal, 1985), kaolinite (Olsen, 1966; Acar,
et al, 1985), and natural clay soils (Fernandez and Quigley, 1985). The
Environmental Protection Agency has suggested that the clay mineral palygorskite
(synonomous with attapulgite) may be suitable for use in clay liners due to its
insensitivity to attack by waste fluids (EPA/625/4-89/022). It is for this reason that
palygorskite is chosen for this investigation. The palygorskite used in this study is
a commercially available palygorskite drilling mud obtained from the Milwhite
Mining Company in Attapulgus, Georgia. The palygorskite was mixed with
commercially available sands and silts to where the only clay material present is that
1


from the palygorskite drilling mud. Three different percentages by weight of the
palygorskite were tested.
Comparisons of hydraulic conductivity tests
performed using flexible and rigid wall permeameters have been performed (Daniel,
et al., 1985). It has been shown that rigid wall permeameters tend to overestimate
hydraulic conductivity due to chemical-clay interactions resulting from shrinkage and
cracking. Also, flexible wall permeameters tend to underestimate hydraulic
conductivity due to the closing of these cracks and fissures (Mitchell and Madsen,
1987).
The constant rate flow pump was initially introduced as a method for
hydraulic conductivity testing in the early 1960s (Olsen, 1961, 1962). In recent
years, advances in flow pump technology include improvement of permeant delivery
systems, such as the infuse-withdraw actuator of Olsen, et al, 1991. The flow pump
is able to evaluate hydraulic conductivity of fine grained materials in considerably
less time than conventional methods such as falling head and constant head methods.
One component of the investigation presented in this Masters thesis includes the
development of a new laboratory technique utilizing present flow pump technology,
but including a system for permeant exchange. The laboratory apparatus utilizes a
flow pump equipped with an infuse-withdraw actuator (Olsen, et al., 1991), bladder
accumulators as permeant reservoirs, a triaxial cell utilized as a flexible wall
permeameter, and differential pressure transducers to simultaneously and continuously
monitor hydraulic conductivity and effective stresses.
Laboratory testing of hydraulic conductivity and chemical compatibility is
extremely important in estimating how a clay liner may function in the field. Testing
the clay liner material with chemicals to which the liner may be exposed is an
important aspect of the laboratory testing program. The following organic permeant
fluids were chosen for this investigation: demineralized water, acetone, methanol,
xylene, and unleaded gasoline. These permeant fluids display a wide range of
2


dielectric constants. By changing the dielectric constant of the permeant fluid, the
diffuse double layer surrounding the clay particles may expand and contract, thus
inciting a change in hydraulic conductivity of the clay soil.
The texture and fabric of a clay soil also play an important role in the
measured hydraulic conductivity. Flocculated and dispersed textures determine sizes
of pore spaces between soil particles, and thus affect hydraulic conductivity. The
present investigation studies changes in fabric induced by the different permeant
chemicals by using scanning electron microscopy (SEM) to examine the soil particles
subsequent to permeation by different organic fluids. Since the electron beam of the
SEM must operate in a vacuum, it is necessary to dry samples prior to use in the
SEM. A second component of this investigation is to design, construct, and use a
freeze drying apparatus to prepare soil samples for study in the SEM. Freeze drying
can preserve the texture and fabric of soil specimens.
1.3 Objectives
The objectives of this research are three-fold: 1) to expand on existing
designs of flow pump systems for hydraulic conductivity measurements, 2) to
measure the hydraulic conductivities of soils containing three different weight
percentages of palygorskite using five different permeant fluids, and 3) to observe
the changes in soil fabric and texture using scanning electron microscopy methods.
3


2. Review of Laboratory Methods of Hydraulic Conductivity Measurements
2.1 Introduction
Hydraulic conductivity measurements are used for evaluating the flow of fluids
through porous media. By measuring the hydraulic conductivity of porous materials
in the lab, it is possible to determine whether a given material will be suitable for use
in low permeability engineering applications. Chemical compatibility of clay liner
materials evaluated using different permeant fluids may also be determined by using
hydraulic conductivity tests. This chapter reviews Darcys Law for flow through
porous media, and determination of hydraulic conductivity using the falling head,
constant head, and flow pump methods. In addition, the use of flexible and rigid
wall permeameters for hydraulic conductivity measurements will be discussed.
2.2 Darcys Law
Darcys Law states that the rate of flow through porous media is proportional
to the hydraulic gradient. Darcys original work (Darcy, 1856) was performed using
clean sands, and is given by the equation
Q = kiA = kA (2.1)
A L
where
Q = flow rate (cm3/sec)
i = hydraulic gradient (cm/cm)
Ah = change in head (Pa)
AL = change in length (cm)
A = cross-sectional area (cm2)
and k = Darcy coefficient of permeability or hydraulic conductivity (cm/sec).
4


Figure 2.1 shows a diagrammatic representation of Darcys Law.
The hydraulic conductivity, k, is a function of both the properties of the soil
and the properties of the liquid permeating through the soil. It is therefore necessary
to define intrinsic permeability, K, which takes into account the properties of the
fluid. The relationship between k, the hydraulic conductivity, and K, the intrinsic,
absolute, or specific permeability, is given by
k = K 1 = K SJS. = K £ (2.2)
H H U
where
k = hydraulic conductivity or coefficient of permeability (length/time)
K = intrinsic, absolute, or specific permeability (length2)
7 = unit weight of the liquid (mass/length2/time2)
fi = absolute or dynamic viscosity of the liquid (poise = mass/length/time)
p = mass density of the liquid (mass/length3)
g = acceleration due to gravity (length/time2)
and U = kinimatic viscosity of the liquid (length2/time).
It can be seen from equation 2.2 that a change in the hydraulic conductivity can be
affected by a change in the ratio of y/p of the permeant fluid.
2.3 Falling head method
In the falling head test (see Figure 2.2a), a column of water is connected to
the base of a soil specimen. The upper level of the column of water is situated above
the top of the soil specimen in order to induce a head gradient across it. The
permeability test is performed by opening a valve located between the column of
water and the soil specimen. This allows water to permeate through the soil
specimen under the induced head. By continuity the flow into the sample must equal
the flow out of the sample. The flow into the sample is given by
5


Figure 2.1 Diagrammatic representation of Darcys Law (Holtz and Kovacs, 1981)
-at t = t.
h
1
2.
Soil
I
T
l
dh
-I at t = t,
h2
(b)
Figure 2.2 Diagrammatic representation of a) constant-head test, b) falling-head test
(Holtz and Kovacs, 1981).
6


0-3)
and the flow out of the sample is given by
Qca = kui = kA (2.4)
Ld
Since the flow into and the flow out of the soil specimen must be equal, the equations
are combined to give
- a ^ = kA. (2.5)
dt L
After separating the variables the following integrals are obtained
- a (*-k[dt
i h L J
After performing the integration, the following formula is obtained
aL.
k =
-In
ALt K
or
k
2.3
aL
ALt
where
a = area of column
A = area of soil sample
L = length of soil sample
At = time for column head to decrease from 1^ to h2.
(2.6)
(2.7)
(2.8)
7


Falling head tests have typically been used for low permeability soils.
2.4 Constant head method
In the constant head permeability test (see fig 2.2b), a constant level of fluid
is maintained above the top of a soil sample. The volume of water that has
permeated through the sample is given by
V = volume
A = cross-sectional area of sample
k = coefficient of permeability
Ah = change in head
A1 = change in length
t = time.
By rearranging terms, the coefficient of permeability is then calculated from the
relation
Constant head tests have typically been used for high permeability soils.
2.5 Flow pump method
The flow pump method of hydraulic conductivity measurements was
introduced in the mid 1960s (Olsen, 1966). Olsens paper demonstrated that the
relationship between hydraulic gradient and flow rate in saturated kaolinite with
various porosities was linear. In order for Darcys Law to be valid, this relationship
(2.9)
where
V A l
Ah A t
(2.10)
8


must be true. Hydraulic conductivities obtained for the saturated kaolinite ranged
from 2.31 x 10'5 cm/sec to 2.88 x 10'8 cm/sec. Since that time, numerous
investigators have developed constant flow laboratory apparatuses to perform
hydraulic conductivity testing (Olsen et al, 1985; Fernandez and Quigley, 1985;
Olsen et al, 1991; Gill et al, 1991; Aiban and Znidarcic, 1989; Schackleford and
Glade, 1994), compatibility testing for engineered low hydraulic conductivity
containment systems (Quigley et al, 1988; Quigley and Fernandez, 1988;
Schackleford and Glade, 1994;), and evaluation of flow pump apparatus and testing
(Aiban and Znidarcic, 1989; Morin and Olsen, 1987).
The flow pump method uses a constant rate flow pump to induce a constant
flow rate of permeant through a soil specimen. The head difference induced across
the two sides of the sample is measured with a pressure transducer. The hydraulic
conductivity can then be calculated from
- QAh- (2.11)
Ai AAh
where
k = hydraulic conductivity
Q = constant flow rate
A = cross sectional specimen area
i = hydraulic gradient
Al = specimen length
Ah = induced head difference
yw = unit weight of water (or other permeant fluid).
The intrinsic permeability K, can then be calculated from equation 2.2. A typical
schematic diagram for performing constant flow rate hydraulic conductivity tests is
shown in figure 2.3.
9


Figure 2.3 Diagrammatic representation of a simple constant flow rate test (Olsen
et al, 1985).
10


2.6 Permeameters
Two types of permeameters are typically used for hydraulic conductivity tests,
rigid-wall and flexible wall.
Three typical kinds of rigid-wall permeameters include the compaction-mold
permeameter, the consolidation-cell permeameter, and the sampling-tube
permeameter. The soil to be tested is directly compacted into the compaction-mold
permeameter, and then tested for hydraulic conductivity. The consolidation-cell
permeameter can be used in two different ways (Olson and Daniel, 1981; and Olson,
1986): the soil can be consolidated and the hydraulic conductivity calculated from
the rate of consolidation, or the soil can be permeated directly. "Undisturbed" soil
samples can be acquired by pushing a sample tube directly into the soil being
investigated. End caps are then placed on the end of the tube and the sample is
permeated while contained in the sampling tube.
Flexible wall permeameters and triaxial cells used as flexible wall
permeameters consist of a sample confined by porous stones and end caps on the top
and bottom and jacketed with a latex (or other material) membrane. Back and
confining pressures can be controlled by means of burettes connected to an air
compressor.
Both rigid and flexible wall permeameters have advantages and disadvantages
as summarized in Table 2.1.
It is important to maintain chemical compatibility between the permeant fluids
being used and all parts of the rigid or flexible wall permeameter. Testing of the
components should be performed prior to beginning a hydraulic conductivity test.
11


Table 2.1 Advantages and Disadvantages of Rigid and Flexible Wall Permeameters
(from Daniels, 1994).
Type of Cell Principal Advantages Principal Disadvantages
Rigid Wall Permeameter Simplicity of construction and operation of cell Low cost of cell Very large permeameters can be constructed fairly conveniently Wide range of materials can be used (including chemically resistant materials) Unrestrained vertical swelling can be allowed Zero vertical stress can be applied, if desired Sidewall leakage is possible No control over horizontal stress If specimen shrinks, sidewall leakage is virtually guaranteed Cannot confirm saturation via B- coefficient measurement Cannot conveniently back pressure saturate the test specimen Longer testing time for low- hydraulic conductivity material
Flexible Wall Permeameter Can backpressure saturate the specimen Can confirm saturation via B- coefficient measurement Can control principal stresses Sidewall leakage is highly unlikely, even for test specimens with rough sidewalls Faster testing times for low hydraulic conductivity materials due to capability for rapid saturation via backpressure High equipment cost for cell Requires 3 pressure positions (cell pressure, influent pressure, effluent pressure) Problems with chemical compatibility of membrane with certain chemicals and waste liquids More complicated operation of cell compared with rigid wall cell Difficult to perform test with extremely low compressive stress
12


3. Review of Scanning Electron Microscopy of Clay-soil Textures
3.1 Introduction
Clay minerals influence the geotechnical properties of soils. A program for
the investigation of the geotechnical properties of clay-soils should include the
identification of the clay minerals present. One important aspect of clay-soils are
their textures and any possible change in texture due to the environment to which
they are exposed. Scanning electron microscopy and the accompanying energy
dispersive x-ray analyzer can make a significant contribution to the understanding of
clay textures.
This chapter reviews scanning electron microscopy, energy dispersive x-ray
analysis, clay-soil textures and fabric, and clay mineralogy.
3.2 Scanning Electron Microscopy
The scanning electron microscope (SEM) is a powerful instrument that permits
the observation and characterization of heterogeneous organic or inorganic material
on the micrometer or submicrometer scale. The material to be examined is
bombarded with a finely focused electron beam, which can be static or swept in a
raster across the surface of the specimen. The types of signals which are produced
by this electron beam interacting with the specimen include secondary electrons,
backscatter electrons, Auger electrons, characteristic x-rays, and photons of various
energies. The electrons of interest here are the secondary electrons (used to produce
the image on a screen and the SEM photomicrograph) and the characteristic x-rays
(used to obtain an energy dispersive x-ray (EDX) spectrum).
The SEM consists of an electron optics column and an electronics console (see
Figure 3.1). The electron column consists of an electron gun with a filament (usually
tungsten or lanthanum hexaboride) and two or more electron lenses, operating in a
13


Scanning Electron Microscope
and
Energy Dispersive X-Ray Spectrometer
cDX System >1
SEM
->1
Electron Ootics Column
Figure 3.1 Schematic diagram of SEM/EDX system (Welton, 1984)
14


vacuum (approximately 2 x 10-6 torr). The electron gun is a source of electrons
which are focused by electron condenser lenses into a small beam onto the specimen.
The electron beam emerges from the final electron lens and interacts with the near
surface region of the specimen to a depth of approximately 1 micrometer and
generates signals to produce an image.
To produce an image, a scanning system must be used which constructs the
image point by point. A deflection system (rastor) is utilized to scan the electron
beam along a line and then displace the line position for the next scan. The
secondary and the backscattered electrons produced by the beam interaction with the
specimen are collected by a secondary electron detector or a backscatter electron
detector. In this way a rectangular raster is formed on the specimen and the viewing
screen. The magnification of the specimen image is the ratio of the linear size of the
viewing screen (cathode ray tube) to the linear size of the raster on the specimen.
Therefore, the smaller the raster width on the specimen, the higher the magnification
of the specimen.
The electron gun produces a large stable current in a small electron beam.
Electron guns consist of a filament which emits electrons due to a high negative
potential induced by a high voltage supply. There are three types of filaments
commonly used: tungsten, lanthanum hexaboride, and field emmission. The
tungsten filament is the most commonly used and the least expensive. The lifetime
of a tungsten filament is about 45 hours. The cost of the lanthanum hexaboride
filament is much higher, has an operating life of approximately 700-1000 hours, and
produces better resolution in photomicrographs. Field emmision guns are generally
made from tungsten, but have a much sharper tip than the regular tungsten filament,
and produce a brighter SEM image.
As the electron guns are emitting electrons, the electrons are accelerated
through the column, demagnified and focused through a series of electromagnetic
lenses into a finely focused beam which bombards the sample. Electrons are thus
15


emitted and detected, and it is the electrical signals created from this bombardment
that allow the specimen to be studied.
3.2.1 Energy Dispersive X-ray Spectrometry
Chemical analysis of the specimen loaded in the SEM sample chamber can be
performed by Energy Dispersive X-ray Spectrometry (EDX). The EDX system is
connected to the SEM as shown in Figure 3.1. The EDX system utilizes the primary
scanning electron beam to excite the atoms in the specimen. The interaction of the
electron beam with the specimen results in an electron being ejected from an atoms
inner shell. This leaves a void in the inner shell which is filled by an electron from
its outer shell (Figure 3.2). This transition, from outer to inner shell, releases energy
in the form of characteristic x-rays. The energy of each x-ray is used to identify
which element is present in the specimen. In computer operated systems, the output
is shown as a graph of intensity versus energy level.
Most EDX systems consist of a Si(Li) x-ray detector, multichannel pulse
height analyzer, data processor, image enhancer, TV display (cathode ray tube), and
a plotter (see Figure 3.1). Computer software is available to reduce the data obtained
into a quantitative chemical analysis with accuracy and precision approaching 1 %
(Goldstein, et al, 1992; and J. Nishi, oral communication).
The reduction of the data obtained from the EDX system requires several
steps: a chi square analysis, and Z, A, and F corrections.
The chi square analysis compares the intensity (in counts per second) of the
element in the specimen to the intensity that would be expected from a pure sample
of that element. The formula for the chi square analysis is:
X2 = £ [y~x)l (3.1)
16


Electron Transitions in an Atom (modified from
Goldstein and Yakowitz, 1978). When an orbiting
electron is ejected from the K shell by the SEM
electron beam, to regain stability an electron from
the L shell fills the vacancy. The amount of x-ray
energy released during this transition is termed the
Ka x-ray. If an electron from the M shell fills the
vacancy, the energy released is termed K/3 etc.
3.2 Electron transitions in an atom (Welton, 1984)
17


where x = standard intensity, and y = unknown intensity. A value of chi square
approaching 1 means that a very good fit for the data was achieved (within 1 %). For
values of chi square > > 1, a poor fit has been obtained, and a new spot on the
specimen should be analyzed. Factors which can affect the quality of the EDX data
include count time (i.e. dead time the time interval after a pulse is recorded during
which the system cannot respond to another pulse should be between 20 and 40 %),
specimen topography (holes in the specimen should be avoided), detector geometry
(direction the detector is located compared to the spot to be analyzed), and the
specimen coating.
Z, A, and F corrections are for atomic number, absorption, and fluorescence,
respectively. If the effect of atomic number is not corrected for, analyses of heavy
elements in a light element matrix will generally yield values which are too low, and
analyses of light elements in a heavy element matrix will generally yield values which
are too high (Goldstein, et.al., 1992). Absorption is corrected for because the
characteristic x-rays are generated at some non-zero depth in the specimen and
interact with the various atoms in the sample, thus decreasing the intensity of the
measured characteristic x-rays (Goldstein, et.al., 1992). The fluorescence correction
factor is necessary because the energy of the x-ray peak from one element can be
sufficient enough to excite x-rays secondarily from another element. This would
result in more x-rays from the first element than would have been produced by
electron excitation alone (Goldstein, et.al., 1992).
After these corrections have been performed, the results of the standardless
semi-quantitative analysis are tabulated in weight percents summing to exactly 100
percent. A standardless analysis means that the elements in the spectrum of the
unknown sample are compared to that obtained from a spectrum of each pure element
present. The spectrum of each respective element analyzed for is programmed into
the computer software utilized for the analysis.
18


3.2.2 Sample Preparation Techniques
Many techniques are available for sample preparation for study in the SEM.
However, some of these procedures will induce changes in the specimen to be studied
(artifacts of the drying process). The techniques to be mentioned will be those most
commonly used in the study of fine grained materials, including clay textures. These
techniques include air drying, critical point drying, freeze drying, and impregnation.
Following the discussion of these techniques, methods and materials for
coating samples with electrically conductive materials (required for SEM studies) will
be examined.
3.2.2.1 Air Drying
The technique of air drying is to allow the specimen being studied to dry
naturally at ambient temperatures. One can also speed up this process by placing the
sample under a heat lamp or in an oven. These different air drying techniques will
affect the specimens differently. Table 3.1 shows the shrinkage effects due to air
drying quickly, air drying slowly, and oven drying (Smart and Tovey, 1982).
The effects of surface tension and interfacial contact angles of the pore fluid
with the clay or soil particles are to decrease the void ratio of the specimen.
However, at higher temperatures the surface tension of the fluids are lower, and less
shrinkage should occur. As can be seen in Table 3.1, a vertical shrinkage of 28%
occurs during slow air drying. This would be unacceptable for the study of soil
microstructure. Nonetheless, EDX analysis could still be performed, so that this
technique would still have some merit. If one is to study the texture and fabric of
the clays, another method would need to be chosen.
3.2.2.2 Critical Point Drying
The critical point of a volatile liquid is defined as that point on the phase
diagram where both the liquid phase and the vapor phase of the liquid are in
19


Table 3.1 Shrinkage during preparation, % corrected for initial swell, after Tovey
1970. Horizontal and vertical are indicated by h and v.
Drying method sh Sv
Oven drying 6.7 23.5
Air-drying, quick 7.7 22.6
Air-drying, slow 12.5 28.5
Freeze-drying -0.63 -.78
Substitution drying
methanol 3.8 6.8
ether 3.2 9.8
isopentane 2.5 5.9
acetone 2.8 8.5
cellusolve -1.2 6.6
Impregnation
Carbowax variable 5.7
Araldite -0.4 -9.5
Vestopal -0.5 -9.5
Durcapan 0.3 -11.0
equilibrium. Surface tension at this point is virtually nonexistent. This prevents or
reduces artifacts of the drying process within the sample. A flow chart of steps for
the critical point drying process is shown in Figure 3.3. Percent shrinkage in critical
point dried samples for different replacing fluids are given in Table 3.1.
The sample is placed in a critical point drying apparatus (bomb) and subjected
to the temperature and pressure required to keep the pore fluid at its critical point.
Ideally, water would be the the pore fluid to use. However, the critical temperature
of water is 375 C, and would effectively cook the sample. In order to alleviate such
a high temperature, the pore fluid in the sample is replaced in a series of steps until
a pore fluid with a lower critical temperature is left saturating the sample. For soil
samples, the fluids exchanged for the interstitial water are, in order; ethyl alcohol,
amyl acetate, and liquid carbon dioxide. The critical temperature and pressure of
20


SAMPLE PREPARATION TECHNIQUES FOR MICROFABRIC ANALYSIS
DEHYDRATION
STEP TECHNIQUE INSTRUMENTATION OR DESCRIPTION

auPeampilng of
clay apecimen
ana trimming

wire knife
raolaeamant of SESsST
Incaratitial water
wtta atnyl
alconol
ailvar nitrate tast
aaeied glaaa
containers
3
replacement of
atnyl aieonot
wiin amyl
acetate_______
HH3
m
protaetad apaelman
in lana oaoar or
Kerfaratad aampia
oidar
raolaeamant of amyl ecetaim with liquid CO, F
4 0
E "I
S critical point drying witlt CO( 0
placement and atoraga of
6 Individual specimen In deeteeaiora
critical point
apparatua
critical point
apparatua
mamelaa In vacuum
daaieeatad or aaaiad
jar with daaiccam
TEM ONLY
6(catto cue or
atm capsule
protective lana
papar removed
vacuum daaiecatoi
Spurr
poxy roain
glaaa knife or.
manually with i
razor blade
diamond knife
microtome
copper grid
SEM ONLY
7S
SS
9S
10S
fractured
ipeeiman
mounted
apecunsn
aiactroaUtlcally
cam fractured
Kirtaed
axrttar coating
vnth
pailadium/geld
wire
aluminum stub
CAS plastic tuba
aampia
T?
a
vacuum evaporator
11S
Surface faaura
pitommicrcacopy
nn
SSML
Figure 3.3 Critical point drying process (Baerwald et al, 1991)
21


liquid carbon dioxide are 31.3 C and 72.9 atm. This technique has been shown to
work well with no artifacts from the drying process on submarine sediments
(Baerwald, et al, 1991).
3.2.2.3 Freeze Drying
Freeze drying involves the freezing of the interstitial pore fluid in the sample
and subsequent sublimation of the frozen fluid under a low pressure vacuum. It has
been shown (Smart and Tovey, 1982) that faster rates of freezing of the interstitial
pore fluid result in less artifacts of the drying process. Table 3.1 gives the
percentage of shrinkage during freeze drying. Very rapid freezing is best because
it will solidify the pore fluid instantaneously without disruptive crystal growth. Low
pressure vacuum is required to keep the frozen fluid from melting and creating
surface tension forces within the microfabric of the soil.
The saturated specimen is plunged into a bath of liquid coolant. Liquid
coolants utilized for this purpose include: pentane, isopentane, liquid propane, and
liquid nitrogen. A mixture of 2-butoxyethanol and C02 (dry ice) can also be used.
Due to the extremely flammable nature of the first three fluids, liquid nitrogen is the
favorable coolant. Additionally, if soil samples will be saturated with fluids other
than water different techniques need to be employed. Carasso and Favard (1961)
mention the possibility of explosion when isopentane is used, since liquid oxygen
collects on its surface. Also, soils saturated in organic compounds will react
violently when exposed to oxidizing compounds such as liquid oxygen. The mixture
of 2-butoxyethanol and C02 (dry ice) will not freeze liquid oxygen from the
atmosphere, since its sublimation temperature is -78C. This mixture will result in
longer freezing times, and more potential for freeze dry artifacts. However, the
safety considerations make it preferable for soils saturated in organic fluids.
After the sample is frozen, it is carefully and quickly placed in a drying vessel
and put under a low pressure vacuum (about 1 x 10-4 torr). The frozen liquid is
22


allowed to sublimate until the sample is dried.
3.2.2.4 Impregnation
Impregnation is the solidification of the sample by filling its pores with a
substance which is subsequently hardened. Impregnation is most commonly used to
reinforce clay soil structures for study in the Transmission Electron Microscope
(TEM). The TEM requires the use of a polished thin section, that is, a section of
the clay (or rock) mounted on a glass slide, finely ground to approximately 0.02 mm
thick, and then polished. Another use of impregnation is for surface studies,
whereby the structure of porous material can be seen in a section cut through an
arbitrary plane of the sample. Extreme care must be taken during the impregnation
process so as not to disturb the structure of the specimen.
Wet and dry impregnation techniques have been used where the sample is
immersed in the impregnating compound under a vacuum. The details of these
techniques can be found in Smart and Tovey (1982). Heating of samples on a hot
plate in order to lower the viscosity of the impregnating fluid has also been used.
Common impregnating compounds include (Smart and Tovey, 1982):
vestopal, epoxy resins, lakeside, methacrylate, bakelite, gelatine, agar, and carbowax.
Spurr resin, and LR White resin have been used following critical point drying and
freeze drying and have been shown to give similar results in TEM analysis (Albert
Yeung, Texas A & M University, oral communication).
3.2.2.5 Surfaces for study
When the drying process is complete, it is necessary to prepare a surface of
the specimen for study. At this point, the samples are extremely fragile (unless
impregnated) and must be handled delicately. The most common way of preparing
a surface is by fracturing. The specimen is literally broken in two pieces, thus
allowing the fractured surface to be studied. Other methods include peeling (after the
23


specimen is glued onto an SEM pedestal), where silver dag (Tovey, 1970), sellotape,
or epoxy is used to adhere to the surface of the specimen. It is subsequently
removed, and takes a "peel" of the specimen off and exposes the inner portion of the
material. As one might expect, specimens must be worked with carefully in this
process.
3.2.3 SEM Preparation
Once the specimen is dried and the surface of the specimen is ready, the
sample must then be prepared for the SEM. This includes mounting the specimen
on an SEM pedestal and coating it with an electrically conductive material.
3.2.3.1 Mounting
The specimen is mounted on an aluminum pedestal designed specifically for
the SEM to be used. Duco cement works well for glueing cohesive samples to the
pedestal and dries relatively quickly. For the study of single grains, double sided
sticky tape can be used, but it does not conduct electricity across the specimen as
well, and charging may occur (this will be discussed in the next section).
Additionally, small grains can be covered with a drop of an electrically conductive
fluid designed specifically for this purpose (available from SEM supply companies).
3.2.3.2 Coating materials
To study materials in the SEM, it is necessary to make the surfaces
electrically conductive. When the electron beam interacts with the specimen, some
of the charge injected is emmitted as backscattered and secondary electrons. That
which is not emmitted needs to flow electrically to a ground. If the material is a
nonconductor, this excess charge will buildup within the specimen (charging). The
result will be image aberrations, poor EDX data, and the sample may tend to move.
Which makes study of the specimen impossible. To remediate this problem, the
24


specimen should be coated with an electrically conductive material. Many different
coating materials can be used, but the most common ones are carbon, gold, gold-
palladium, and copper. Carbon is the best conducting material to use, but gold
allows for the best photomicrographs. It is possible to coat with carbon first, then
with a layer of gold. This has proven to work well for clays and clay soils.
If EDX is an important aspect of the study being performed, one must choose
that coating material which will not have overlapping spectra with the elements of
interest in the specimen.
3.3 Clay Mineralogy
Clay minerals are classified as phyllosilicates and have a diverse range of
properties due to their chemical and physical structures. Clay minerals may be
subject to volume change due to such processes as adsorbtion, absorbtion, and ion
exchange. These properties make them useful materials in a wide variety of
applications. Knowledge of the physical and chemical mineralogy of the main clay
mineral groups is vital to an understanding of how these minerals may behave when
used in different applications.
There are two main structural units comprising the clay minerals, namely, the
octahedral and tetrahedral sheets. Tetrahedral sheets consist of connected silica
tetrahedra. The silica tetrahedra are comprised of silica (Si4+) cations, each
surrounded by four oxygen anions. Aluminum (Al3+) may substitute for the silicon,
as may iron (Fe3+). This sheet may be pictured as extending infinitely in two
directions, with each tetrahedron lying on one of its faces. Each tetrahedron will
then share oxygens with three neighboring tetrahedra, and have one oxygen pointing
upwards (Figure 3.4).
The octahedral sheet consists of two planes of closest packed oxygen ions,
with cations occupying the octahedral sites formed between the two planes (Figure
3.4). These octahedra are connected by their edges to neighboring octahedra while
25


Figure 3.4 Structure of Kaolinite (Berry and Mason, 1959)
6 0
4 Si
z (OH)+4.0
4 Al
2. (OH)+ 4-0
4 Si
6 O
Figure 3.5 Structure of Pyrophyllite (Berry and Mason, 1959)
26


sharing oxygen ions. The most common cations found in the octahedral sites of this
sheet are aluminum (Al3+), magnesium (Mg2+), and iron (Fe2+ or Fe3+), but others
such as the transition elements may also occur (Berry and Mason, 1959).
Clay minerals may also be described as dioctahedral or trioctahedral
depending on the valence of the cations contained in the octahedral sites and the
number of cations present. A trioctahedral clay is one containing divalent cations in
all the octahedral sites. This will maintain electrical neutrality. In dioctahedral
clays, some of the octahedral sites are filled with trivalent cations. To maintain
electrical neutrality in this situation some of the octahedral sites must be vacant.
To join the octahedral and tetrahedral sheets together, it is necessary for the
apical oxygen ion of the tetrahedral sheet to be shared within the structure of the
octahedral sheet. Such a clay structure made up of one octahedral sheet and one
tetrahedral sheet is called a 1:1 layer silicate structure. Similarly, a 2:1 layer silicate
structure is comprised of two tetrahedral sheets surrounding an octahedral sheet.
Both apical oxygens of the tetrahedral sheets face inward towards the octahedral
sheet. This assemblage essentially makes a tetrahedral-octahedral-tetrahedral
sandwich (Figure 3.5).
The distance between these sheets is controlled by which cations are present
in the structure. Different cations have different radii and charges. The different
charges on the cations present may result in electrical inequality for the mineral. The
result is a net negative charge present on the surface of the clay. This negative
charge results in attractions for other cations, and cation exchange may occur.
Cation exchange alters the interplanar spacing within the crystal lattice, and this
interplanar spacing may be measured using an x-ray diffractometer (XRD).
The cation exchange capacity (CEC) is a measure of the ability of a clay to
have one cation on its surface exchanged for another cation. In general, the ease
with which one cation will replace or exchange another may be given as
27


Na+ < Li+ < K+ < Rb+ < Cs+ < Mg2+ < Ca2+ < Ba2+ < Cu2+
< Al3+ < Fe3+ < Th4+ < NH4+
i.e., Ca2+ is more stable, or more firmly fixed, in the interlayer space of the clay
than is Na+. The hydrogen ion may also be substituted in acidic conditions (Moore
and Reynolds, 1989). The CEC depends on the surface charge of the clay mineral
involved.
The main clay mineral groups kaolinite, smectite, and palygorskite can
now be discussed.
3.3.1 Kaolinite group
Kaolinite is the most common of the kaolinite group of clay minerals, all
having similar physical and chemical composition. Other minerals in this group
include: dickite, nacrite, anauxite, halloysite, and metahalloysite. A rock or
aggregate which contain these minerals is called "kaolin", or sometimes "china clay".
Kaolinite is commonly found in soils and permeable bedrock in warm, moist
climates, forming as a weathering product of other aluminosilicates, such as feldspar.
It can also be formed from the hydrothermal alteration of the same feldspar minerals.
The chemical composition of kaolinite is Al4(Si4O10)(OH)g. Kaolinite is a 1:1
layer silicate mineral (Figure 3.4) and consists of alternating tetrahedral and
octahedral sheets. It has essentially no cation substitution, so it is an electrically
neutral clay. It therefore has little or no cation exchange capabilities.
3.3.2 Smectite group
Smectites are 2:1 layer silicate type clay minerals. The smectite group
minerals include montmorillonite, beidellite, nontronite, hectorite, and saponite.
These clay minerals can be either dioctahedral or trioctahedral and display the
property of being able to expand and contract their structure while maintaining their
crystallographic integrity. Expansion and contraction takes place when water or
28


organic fluids enter the interlayer space of the clay mineral. A schematic drawing
of montmorillonite is shown in Figure 3.6. This drawing shows the tetrahedral-
octahedral-tetrahedral layering of montmorillonite and where water and organic fluids
may be adsorbed between the layers.
An ideal formula for montmorillonite would be Al2Si4Oio(OH)2* xH20.
However, ions of Na+, Ca2+, and K+, are present as adsorbed ions, and substitution
of Al3+ for Si4+ may also take place. The layer charge on these clays will vary
depending on the amount and types of different cations present. For montmorillonite,
the layer charge will originate from substitution of cations in the octahedral sheet;
for beidellite and nontronite, the layer charge is the result of substitution in the
tetrahedral sheet (Moore and Reynolds, 1989).
There are several theories on the origin of smectite minerals. First, smectites
are known to form from the alteration of volcanic glass, forming a poorly indurated
rock that is termed bentonite. Second, smectite precipitates directly in pore spaces
of sandstones and apparently forms in environments of very slow water movement
in swampy lowlands or in arid to semi-arid regions (Berner, 1971). Third, smectite
forms as a weathering product of kaolinite, chlorite, and illite.
3.3.3 Palygorskite group
The two minerals which make up this group are palygorskite (also known as
attapulgite) and sepiolite. They are both 2:1 layer silicates with the tetrahedral sheets
linked essentially infinitely in two dimensions. They are structurally different from
the other clay minerals in two ways. First, the octahedral sheets are continuous in
only one direction, and second, the tetrahedral sheets are divided into ribbons by
inversion, but are still linked (Moore and Reynolds, 1989). This creates an open
channel (Figure 3.7) within the crystallographic structure of the mineral. These
channels may then contain water and exchangeable cations. The ideal formula for
palygorskite is (Mg,Al)5(Si,Al)8O20 8H20. Sepiolite is similar in composition to
29


nHjO layers and exchangeable cations
Figure 3.6 Structure of Montmorillonite (Holtz and Kovacs, 1981)
30


Crystal structure of palygorskite, c axis vertical (Zoltai and Stout, 1985).
Figure 3.7 Palygorskite structure (Jones and Galan, 1988)
31


palygorskite but has more magnesium. Although palygorskite and sepiolite are
somewhat rare, they are quite abundant in the Mediterranean and Middle East. They
appear to be products of weathering, and may form from smectite or convert to
smectite in marine environments that are sub- or hypersaline. Palygorskite is a
common constituent of desert soils (Moore and Reynolds, 1989). In the United
States, palygorskite is mined near Attapulgus, Georgia (the origin of the term
attapulgite), and is interpreted as having formed in a shallow marine depositional
environment (Van Olphen and Fripiat, 1979).
3.4 Clay Microstructure
Clay microstructure refers to the orientation and arrangement or spatial
distribution of the solid particles and their particle to particle relationships. Since the
clays have a net negative charge on their surfaces, electrical attractions and repulsions
may occur between particles. This will result in various fabrics due to different
particle arrangements. These different fabrics are shown in Figure 3.8. Different
particle arrangements such as end to end, face to face, and face to end can be
present. Also, samples may be dispersed or flocculated into groups.
Mitchell (1993) defines three levels of scale for the fabric of a soil. 1)
Microfabric The regular aggregations of particles and the very small pores between.
Typical fabric units are up to a few tens of micrometers across. 2) Minifabric The
aggregations of the microfabric and the interassemblage pores between them.
Minifabric units may be a few hundred micrometers across. 3) Macrofabric -
Macrofabric contains cracks, fissures, root holes, laminations, etc.
32


Various modes of panicle association (after van Olpken 1963): fa) Dispersed and defloc-
culated. (b) Aggregated but deflocculated, (c) EF flocculated but dispersed, (d) EE flocculated but dispersed,
(e) EF and FF flocculated and aggregated, (f) EE flocculated and aggregated, (g) EF and EE flocculated and
aggregated.
Figure 3.8 Clay textures and fabrics (Bennett and Hulbert, 1986)
33


TurfroscroRc jtRia
(redruu'n from A>(more and Quirk
I9o0).
Perfect jwck (redrawn
fnm Sides and Barden I9i).
Bookkouse or book
structure (after foil 1964; modified
fnm Sloan and Kell 1966).
I
r
r
U)
Domnin structures
(redrawn fnm Moon i972): (al Book
(redrawn fnm SZoan and Kell 1966)
and (b, c) Stepped face-to-faze struc-
tares (redrawn fnm Smalley and
Cabrera 1969).
C
1
(b)
(c)
Figure 3.8 (cont.)
34


Tcictoiti structure
indrawn from Ingles 1966).


CVirdhouse structure
(redrawn from Ingles 1968).
Stairstep stnicture oj
kaolinite minerals (redrawn from
OBrien 1971).
Stairstep structure of il-
lite mineral (redrawn from O'Brien
1971).
35


Flocculated
Dispersed
Consolidated
(c). (d)
Proposed schemes of panicle arrangement in clay sediments (redrawn from Moon 1972): (a)
Open, random arrangement of domains of 2 to 3 panicles per packet, (b) Parallel or subparallel anangement
of domains of 2 to 3 particles per packet, fc) Increased parallelism of domains and moTe particles incor-
porated into each domain than in (z) (mudstone), (d) Complete parallelism and more panicles per packet
than in (b) (shale).
Figure 3.8 (cont.)
36


4. Clay Liners in Landfill Applications
4.1 Introduction
Clay liners in present landfill design represent the last barrier of a liner system
to inhibit leachate from contaminating local aquifers. This chapter reviews the basic
design of liner systems used in landfills, mineralogical composition of clay liners,
chemical effects on clay liners, and double layer theory.
4.2 Design of clay liners
Typical designs of clay liners for a single clay liner system and a double liner
with a leak detection system are shown in Figures 4.1 and 4.2, respectively. A
single liner system is typically comprised of a native soil foundation, a compacted
clay liner, a flexible membrane liner, a drainage system, and a filter medium.
Double liner systems consist of (from bottom to top) a native soil foundation, a
compacted clay liner, a flexible membrane liner (FML), a secondary leachate
collection and removal system with a sump, another flexible membrane liner, a
primary leachate collection and removal system, and a filter medium.
Materials to be used in soil liners must meet the following requirements (EPA,
1988):
A field hydraulic conductivity of 1 x 10'7 cm/sec when compacted
Sufficient strength after compaction to support itself and the overlying
materials without failure
Compatibility with hazardous wastes or leachate to be contained at the
site.
Soil liner material may be derived from the landfill site, or may need to be
hauled in from a nearby borrow site. If local site or borrow material are inadequate
to properly construct a clay liner, additives such as bentonite or other clay materials
37


Schematic of a Single Clay Liner System tor a Landfill
Protective Filter Medium
(Not to Scale)
Figure 4.1 Single clay liner system (EPA/625/6-88/018, 1988)
Schematic of a Double Liner and Leak Detection System for a Landfill
Filter Medium
Top Liner
(FML)
Bottom Composite
Liner
Primary Leachate
Collection and
Removal System
Secondary Leachate
Collection and
Removal System
Being Proposed as the
Leak Detection System
Upper
Component
(FML)
Native Soil Foundation
Leachate
Collection
System
Sump
Lower Component
(compacted soil)
Figure 4.2 Double liner with leak detection system (EPA/625/6-88/018, 1988)
38


may be introduced to enhance the performance characteristics of the soil.
A soil with the following specifications are suitable for liner construction, and
will be able to obtain the required hydraulic conductivity of 1 x 10'7 cm/sec (EPA,
1994):
Plasticity index >10%
Percent fines (#200 sieve, > 30 %)
Percent gravel (#4 sieve, < 20% to 50%)
Maximum particle size (< 25 mm to 50 mm).
4.2.1 Soil liner composition
To fully understand the chemical reactions taking place within a given system,
it is necessary to have knowledge of the materials involved in the reactions.
Although the chemicals found in landfill facilities may not be precisely known, one
can make assumptions as to what chemicals may be present, and combined with
knowledge of the mineralogy of the clay liner, reactions can be inferred and studied.
In addition, leachate may be obtained from existing landfill facilities and tested with
known clay liner materials.
From a geotechnical engineering standpoint, the term clay refers to clay soil.
The clay soil can be classified either according to the Unified Soil Classification
System (USCS), or the American Association of State Highway and Transportation
Officials (AASHTO) system. In either case, the clay soil is defined and classified
by the percent passing the no. 200 sieve and the Atterberg limits of the soil using
Casagrandes plasticity chart.
In addition to the physical characteristics of the clay soil as defined by its
classification, it is also composed of several different mineral constituents. It is these
mineral constituents that may react with the chemicals within the landfill and have an
effect on the strength and hydraulic conductivity of the clay liner system.
39


Most soils will largely consist of the following minerals (and possibly many
others in smaller amounts), but will be dependent upon the source of the material:
Quartz Si02
Feldspars_ (K,Ca,Na)AlSi3Og
Carbonates__ Calcite, CaC03; siderite, FeC03, etc.
Evaporites__ Halite, NaCl; Gypsum, Anhydrite, CaS04; etc.
Clay minerals___ formed by the weathering of the feldspar minerals
above. The clay mineral groups were discussed in Chapter 3
Other minerals__ such as olivine, amphiboles, pyroxenes, hematite.
This is not meant to be an exhaustive list, and are certainly not the only minerals
found in soils. However, these are the most important minerals (and mineral groups)
which affect a soils chemical, physical, and geotechnical properties.
Materials available for use as liners include clayey soils, admixes of asphalt
and cement with soils, and polymeric substances. Table 4.1 shows some examples
of compatibilities of clay liner systems and industrial waste. It can be seen from this
table that clay soils and admixes of soils are not as reliable as polymeric materials.
Thus a composite liner system of both a clay liner and a flexible membrane liner will
outperform a flexible membrane liner or a clay liner alone.
4.3 Double Layer Theory
Before discussing the effects chemicals have on the mineral constituents of
clay liners, it is first necessary to have an understanding of double layer theory.
The double layer (equivalent to "diffuse double layer") theory relates a
particles charged surface and the distributed charge in the adjacent phase. Pertaining
to clay liners, this relates the net negative charge of the clay mineral phase to the
fluid phase within the pores of the compacted clay liner. The net negative charge on
40


Table 4.1 Liner/Industrial Waste Compatibilities (Dawson and Mercer, 1986).
F = fair, P = poor, and G = good.
Liner material Strong Caustic Strong Acid Organic Solvent Oily Waste Exchang. Cations
Soils
Compacted clay F P P-G G P
Soil-bentonite F P P-G G P
Admixes
Asphalt-concrete G F P P G
Asph-membrane G G P P G
Soil asphalt F P P P F
Soil cement F P P-G G F
Polymeric membranes
Butyl rubber G G P P G
Chlorinated G G G P G
polyethylene
Polypropylene G G G G G
Ethylene propy- G G P P G
lene rubber
Polyethylene G P-G F-G G G
(low density)
Polyvinyl G G G G G
chloride
the clay minerals is derived from the isomorphous substitution taking place between
the cations of the octahedral and tetrahedral layers. Changes in the cation valence,
temperature, electrolyte concentrations, and dielectric constants (permittivity)
represent possible variations within the fluid phase.
Four idealized assumptions are made in the quantification of double layer
models (Mitchell, 1993):
Ions in the double layer are point charges
Charge on the particle surface is uniformly distributed
The particle surface is a plate that is large relative to the thickness of
the double layer
The dielectric constant of the medium is independent of position.
41


An approximate value of the thickness of the double layer can be obtained from the
following equation (Mitchell, 1993):
d = (
e.DkT \
----)2
(4.1)
where
d = double layer thickness
e = dielectric constant of vacuum
(8.8542 x 1012 Coulomb2/Joule-meter)
D = dielectric constant of fluid
k = Boltzmann constant
(1.38 x lO'23 Joules/K)
T = temperature
nG = electrolyte concentration
e = electronic charge
(1.602 x 1019 coulomb)
and v = cation valence.
Equation 4.1 shows that the thickness of the double layer will vary with changes in
the pore fluid. The thickness varies inversely with the valence and the square root
of the concentration. Also, the thickness varies directly with the square root of the
dielectric constant and temperature. Pertaining to clay liners, an increase in cation
valence (as from monovalent to di- or trivalent cations) and/or an increase in
electrolyte concentration will result in a decrease in the thickness of the double layer
around the clay particles. A decrease in cation valence and/or a decrease in
electrolyte concentration will result in an increase in the double layer thickness. If
42


the dielectric constant of the fluid or the temperature were to decrease, the thickness
of the double layer would also decrease. The converse is also true.
4.4 Effects of Chemicals on Clay Liners
The compatibility of a specific waste chemical with the clay liner material is
of prime importance in the selection of the type of clay liner material to use. The
principal waste properties that may adversely affect a landfill liner are (Matrecon,
1980):
Acidic conditions, pH < 3.5
Alkaline conditions, pH > 10
Organic compounds
Exchangeable cations such as Ca+, Na+, Mg2+, Fe2+, Fe3+, etc.
Mitchell and Madsen (1987) give an excellent review of chemical effects on clay
hydraulic conductivity.
4.4.1 Acidic conditions
Depending on what acids are present in the system, the minerals in a clay
liner may or may not react. However, under certain conditions most minerals can
be dissolved by acids. Hydrofluoric acid will dissolve silicate minerals, hydrochloric
acid will dissolve iron oxides, some metals, and carbonates, sulfuric acid will
dissolve sulfates, nitric acid will dissolve some sulfides, and acetic and oxalic acids
will dissolve carbonates. Thus, acidic conditions can effectively dissolve the minerals
in the clay liner resulting in channel formation and increases in the hydraulic
conductivity of the liner. Daniel (1987) shows a typical plot (see Figure 4.3) of
hydraulic conductivity versus pore volumes of flow for a compacted clay permeated
with concentrated acid. Daniel attributes the initial decrease in hydraulic conductivity
to the precipitation of solid matter from the permeating liquid as the acid is neutralized
43


Hydraulic Conductivity (cm/s) pH of
Effluent
Pore Volumes of Flow
Figure 4.3 Typical variation in hydraulic conductivity for sample of compacted clay
permeated with concentrated acid (Daniels, 1987).
44


by the dissolved soil. The precipitates initially plug the soil pores and decrease the
hydraulic conductivity. With continued permeation, fresh acid redissolves the
precipitates and eventually causes an increase in hydraulic conductivity. Other
investigations of low pH hydraulic conductivity tests include Nasiatka et al, 1981;
Peterson and Gee, 1986; and Wright and Shackleford, 1995.
4.4.2 Alkaline conditions
Tests performed with high pH solutions have showed decreases in hydraulic
conductivity using flexible wall permeaters (Lentz et al, 1985), and increases in
hydraulic conductivity using rigid wall permeaters (DAppolonia, 1980). However,
bases should be viewed as having the same detrimental effects on clay liners as acids.
More compatibility studies need to be performed using high pH conditions.
4.4.3 Neutral Inorganic and Organic Fluids
Non-acidic and non-basic liquids can change hydraulic conductivity of soils
in other ways. Clayey soils are made up of colloidal particles that have negative
charges along their surfaces. Water is a polar molecule and has its atoms aligned
asymetrically. This allows the water to be attracted to the negatively charge soil
particles. Also, any cations in the water will be attracted to the soil particles. This
leads to a zone of water and ions surrounding the clay particles known as the double
layer.
The water and ions in the diffuse double layer are attracted so strongly
electrochemically to the clay particles that they do not conduct fluids. Fluids
permeating through the clay liner must go around the soil particles and this double
layer. The hydraulic conductivity of the clay liner is then controlled very strongly
by the thickness of the double layer. When the double layer shrinks, it opens up
flow paths resulting in increases in hydraulic conductivity. When they swell, they
constrict flow paths, resulting in decreasing hydraulic conductivity.
45


The dielectric constant represents the effect of the dielectric material (the
chemical in solution, or leachate), in decreasing the force between charges (the
negative charge of the clay and the positive charge of the dielectric material).
Aqueous solutions with few electrolytes, such as distilled water, tend to
expand the double layer and to produce a low hydraulic conductivity. Solutions with
monovalent cations such as Na+, tend to produce lower hydraulic conductivities than
those with polyvalent cations like Ca+2. A strong solution containing polyvalent
cations tends to produce the largest hydraulic conductivities.
The dielectric constant of water is 80. The dielectric constants of most
organic chemicals are much smaller. This will result in a smaller value of thickness
for the diffuse double layer. This means that the hydraulic conductivity of the clay
liner will increase with the presence of organic chemicals due to the shrinking and
cracking of the clay.
It has also been shown (Daniel, 1987; Bowders, 1989) that dilute organic
chemicals do not alter the hydraulic conductivity significantly. If a small amount of
low-dielectric constant organic chemical is mixed with water, the dielectric constant
of the mixture is only slightly less than water. Tests have shown that the dielectric
constant must be less than 30 50 for the hydraulic conductivity to increase.
Two criteria must be met if the organic chemical is not to adversely affect the
clay liner: (1) the solution contains at least 50 % water, and (2) there is no
separation of phases, as in immiscible fluids (U.S.E.P.A., August 1989).
46


5. Materials and Methods
5.1 Introduction
The experimental approach of this investigation focuses on a design for a
compatibility testing apparatus with capabilities of continuously and simultaneously
measuring the hydraulic conductivity and effective stresses in a soil sample.
Additionally, permeant fluid can be exchanged, flow rates controlled, hydraulic
gradients determined, and effluent samples removed for chemical analysis. The
experimental system described is capable of performing volume controlled, as well
as stress controlled experiments. A freeze drying method and apparatus is described
which was designed and constructed in order to preserve the microfabric of the soil
specimens for scanning electron microscopy analysis.
This chapter describes the new compatibility testing apparatus, the freeze
drying apparatus, and the procedures for using them. The clay and soil materials,
and the permeant fluids used to demonstrate the capabilities of the new compatibility
apparatus are also described, as well as procedures for scanning electron microscopy
investigations and chemical analyses.
5.2 Materials
Three admixtures of a sand fill material and the clay mineral palygorskite
were used for the initial hydraulic conductivity and effective stress experiments under
volume controlled conditions. The sand fill used in this investigation was purchased
from Ainsworth Rock Sales, Inc., Thornton, Colorado. The palygorskite material
was a palygorskite drilling mud produced by the Milwhite Mining Co., Attapulgus,
Georgia, and donated for this investigation. The three admixtures consisted of the
sand fill material plus ten, twenty, and thirty percent palygorskite by weight.
After these initial tests, the ten and twenty weight percent mixes were used
47


for compatibility testing under both volume and stress controlled conditions. The
permeants used were demineralized water, acetone, methanol, and an industrial waste
leachate.
In addition to these tests, the industrial waste leachate was permeated through
a natural clay soil (HRC) from the Highlands Ranch area admixed with ten percent
by weight clinoptilolite (a zeolite group mineral, < 4.75mm) rich rock (CRR) from
South Dakota (SDA). The effluent was monitored for changes in major element and
trace metal chemistry.
A total of fourteen tests were performed to demonstrate the capabilities of the
new compatibility testing apparatus.
5.3 Mineralogical Analysis
The mineralogy of the sand fill material and thepalygorskite drilling mud was
determined by x-ray diffraction (XRD). A Philips 1830 x-ray generator with a
Philips 1840 microprocessor controller was used.
The sand fill material was pulverized to -400 mesh in a Spex 8500 shatterbox.
A backpack powder mount was then made of the pulverized material, and the mount
loaded in the XRD unit. The palygorskite drilling mud was prepared using the
Millipore Filter Transfer Method of Moore and Reynolds (1989). The palygorskite
was then analyzed before and after treatment with ethylene glycol to determine if
smectite was present.
5.4 Geotechnical Index Tests
The geotechnical index properties of the sand fill material, the three
admixtures of sand fill plus palygorskite, the natural clay, and the natural clay
admixed with zeolite were determined using the following ASTM methods: ASTM
D 4318 Standard Test Method for Liquid Limit, Plastic Limit, and Plasticity Index
of Soils; ASTM D 854 Standard Test Method for Specific Gravity of Soils; ASTM
48


D 422 Standard Test Method for Particle-Size Analysis of Soils; ASTM D 2487
Standard Test Method for Classification of Soils for Engineering Purposes; ASTM
D 698 Standard Test Method for Moisture-Density Relations of Soils and Soil-
Aggregate Mixtures Using 5.5-lb (2.49-kg) Rammer and 12-in. (305-mm) Drop.
5.5 New Compatibility Testing Method
5.5.1 Sample Preparation
The palygorskite soil specimens were compacted 2% wet of optimum in a
standard compaction mold. A two inch diameter brass sampling tube was then
pressed into the mold until a specimen of approximately one and a half inches in
height was obtained. The specimen was then extruded. Filter paper, stainless steel
porous stones, and two-inch plastic cylinders with holes drilled into them were placed
on each end (these allow air to escape during saturation) of the specimen. The
specimen was then wrapped in two layers of teflon tape, jacketed with a 1.875 inch
diameter latex membrane, and o-rings were placed on the plastic cylinders. The
specimens were then placed into a vacuum desiccator filled with deaired
demineralized water and placed under a water aspirator generated vacuum for five
days to saturate. The samples were then placed into the compatibility testing
apparatus.
The natural clay admixed with zeolite was prepared differently. The
cylindrical soil sample consisted of three segments. The top and bottom segments
consisted of Highlands Ranch Clay (HRC) and the middle segment consisted of HRC
plus 10 % South Dakota zeolite (SDA). Schleicher and Schuell sharkskin filter paper
was placed between the three segments to make it easier to remove each separately
for chemical analysis following the test.
The soil sample was 1 15/16 inches in diameter and 1 inch long. Each
segment was compacted statically about 2% wet of optimum (25% water content).
Filter paper was placed on the top and bottom of the sample, and then capped with
49


stainless steel porous stones. The plastic cylinders described above were placed on
each end, the sample wrapped with two layers of teflon tape, jacketed in a 1.875 inch
diameter latex membrane and placed in the vacuum desiccator to saturate for five
days. It was then loaded into the compatibility apparatus.
5.5.2 Compatibility Testing Apparatus
The design of the compatibility testing apparatus is shown in Figure 5.1.
Photographs of the system and its components are shown in Figure 5.2. This new
experimental system consists of the following components: a Harvard Apparatus
model 909 flow pump (P) equipped with an infuse-withdraw actuator (after Olsen et
al., 1991); four Geostore model S-470 bladder accumulators (A, B, C, and D)
equipped with impermeable membranes (viton or buna) compatible with the permeant
being used; a Geotest triaxial cell (S) fitted with a teflon base and top cap; two
Validyne model DP-15 differential pressure transducers (M and N) driven by a
Validyne model UPC-601-L LVDT and VRDT sensor interface card in an IBM 286
personal computer. All plumbing lines, fittings, and valves are constructed of
stainless steel. Data were collected using the Validyne Easy Sense software package.
The compatibility testing system allows both volume controlled and stress
controlled experiments to be performed. Stress controlled experiments can be
performed with the base pressure of the test specimen controlled by a pressure
regulator connected to an air compressor. In this type of experiment, the effective
stress at the base of the specimen should remain constant.
Volume controlled experiments (constant void ratio) can be performed by first
bringing up the effective stress to the desired pressure using a pressure regulator
connected to an air compressor. The base of the specimen is then isolated from the
pressure regulator. The confining pressure remains connected and open to a constant
pressure. Any alteration in effective stress is then attributed to soil texture and fabric
changes. However, small quantities of undissolved air in the sample may also
50


Air
Figure 5.1 New compatibility testing apparatus. A, B, C, and D are bladder accumulators. M and N are
differential pressure transducers. S is the soil sample in a flexible wall permeameter (triaxial cell).
P is a constant rate flow pump equipped with an infuse-withdraw actuator. Ah = head difference,
a' = effective stress at base of specimen.




t
I
I
Figure 5.2 Photographs of compatibility testing apparatus. A) Complete system
showing pressure panel, compatibility testing apparatus, and data acquisition
system, B) Compatibility testing apparatus.
52


Figure 5.2 (cont.) C) Close up of flow pump with infuse withdraw actuator,
bladder accumulators. D) Close up of triaxial cell, differential pressure
transducers.
53


contribute to effective stress changes. Olsen, et al, (1989) saw a steady increase in
effective stress from about 140 KPa (20.3 psi) to over 400 KPa (58 psi) in a test
specimen of kaolinite when permeated with distilled water followed by 1 M NaCl,
distilled water adjusted to pH 10 with NaOH, and 0.5 M CaCl2. This increase in
effective stress was attributed to undissolved air in the system.
A temperature controlled room is the ideal environment for using the
apparatus. Changes in temperature can induce shrinkage and expansion of the
stainless steel components of the system, with subsequent changes in measured
pressures. A temperature controlled room was not available for this investigation.
However, the system was located in a laboratory fume hood for safety concerns. The
sash on the hood remained down while running tests, and the hood motor was turned
off. This isolated the apparatus from major temperature changes within the room,
so that tests were not significantly altered by temperature variations.
Both volume and stress controlled compatibility tests were continued until a
baseline permeability with demineralized water was obtained. The fluids were then
exchanged in the testing apparatus, followed by permeation with the second fluid until
the permeability again reached equilibrium, and at least two pore volumes of
permeant fluid had been leached through the sample. At this point the test was
ended.
For the compatibility test that included chemical analyses, testing was to
continue until breakthrough of analyzed elements was complete.
5.5.3 Compatibility testing procedure
Before beginning a compatibility test, it is necessary to check the calibration
of the differential pressure transducers. To calibrate the Validyne DP-15 differential
pressure transducers, it is necessary to first fill both sides of the transducer with
fluid. A known pressure is then placed on one side of the transducer while
measuring the differential pressure with the UPC601 computer controlled data
54


acquisition card. This should be performed for several different pressures and will
result in a linear distribution of data points. The slope of this line will be the
transducer coefficient used to convert the electrical signal obtained in millivolts to
pounds per square inch. During permeability testing, the transducer may be
calibrated by short circuiting (isolating it from the system), thus creating equal
pressure on both sides. The calibration can be performed while the permeability test
is in progress, and will not disturb the results. The reading should drop to zero psi.
If not, adjust the offset factor to give a reading of zero. Any aberrations in
transducer calibration can be explained by electrical drift, temperature change, or
diaphragm failure.
For transducers fitted with high pressure diaphragms, the calibration can be
performed with a dead weight system or with compressed air. When using
compressed air, the pressure gauges do not always give accurate readings. However,
the line generated will still be linear, and the slope of the line obtained will still
represent the transducer coefficient. For transducers fitted with low pressure
diaphragms, the calibration must be performed with a column of water. Again, both
sides of the transducer must be filled with fluid. The column of water is raised a
known amount, thus increasing the differential pressure between the sides of the
diaphragm. This pressure is monitored on the computer, and a straight line is
obtained. The slope of the straight line represents the transducer coefficient.
With the transducers calibrated, the procedure for performing a compatibility
test follows:
1. Fill bladder accumulators with the desired fluid for each. Water, dump, and
chamber bladder accumulators (A, B, D) should be filled with water. The
permeant bladder accumulator (C) should be filled with the desired permeant
fluid. (The permeant bladder accumulator may be filled with the desired
chemical or leachate just prior to permeation. This will reduce the contact
time between the chemical and the bladder.)
2. Saturate all plumbing lines with deaired water. This is best achieved in a
systematic fashion by saturating, in order, the effective stress tranducer lines,
55


the change in head transducer lines, the base and top cap of the triaxial cell,
the bladder accumulator lines, and the infuse/withdraw actuator lines. Use
the water and dump bladder accumulators for sources of water.
3. Load the prepared sample carefully into the triaxial cell. Put the cylinder and
the top piece of the triaxial cell into place. Tighten into position.
4. Fill the chamber of the triaxial cell with water.
5. Increase the chamber pressure to the desired pressure.
6. Increase the chamber pressure and the base pressure to the desired values,
maintaining the desired effective stress at the base of the specimen.
7. Check the latex membrane for leaks. Isolate the base pressure from the
pressure regulator while maintaining the chamber pressure. If the effective
stress at the base of the specimen is constant, there are no leaks. If the
effective stress decreases, water from the chamber is leaking into the sample.
Other possible causes of a change in effective stress: consolidation, sample
not saturated, plumbing leaks.
8. Check zero balance on transducers. Isolate each differential pressure
transducer from the system and short circuit each. Pressure reading should
be zero. If not, reset the offset factor within the Easy Sense software to
obtain a reading of zero.
9. Begin the test. Set the flow pump gear to a previously determined setting.
Activate data logging system. Infuse the actuator.
10. When the permeability has equilibrated with the initial permeant fluid, fill the
permeant bladder accumulator with the next fluid to be tested. Stop flow
pump and isolate the sample from the system. Exchange fluids in the actuator
by simultaneously pushing fluid from the back of the actuator (the effluent)
into the dump bladder accumulator and refilling the front of the actuator with
the second permeant fluid.
NOTE: If the effluent is to be collected, push the fluid from the back of the
actuator (effluent) out the effluent collection line. The effluent can be
collected in a beaker as the effluent collection line is not pressurized.
11. Exchange permeant fluids across the base of the specimen.
12. Begin test with second permeant fluid. Continue until equilibrium is reached
with the second permeant fluid, or approximately two pore volumes of
permeant have leached through the soil specimen.
13. Repeat steps 9 through 12 until desired permeants have been tested for the
specific soil specimen.
14. Terminate the test by shutting off the flow pump and deactivating the data
acquisition system.
15. Drain the chamber of the triaxial cell and remove the soil specimen.
56


5.6 Freeze Drying Method
5.6.1 Freeze drying apparatus
The process of freeze drying inhibits the formation of capillary forces between
soil particles due to liquids contained in soil pores and reduces soil microfabric
disturbances.
The design of the freeze drying apparatus is shown in Figure 5.3. A
photograph of the freeze drying apparatus is shown in Figure 5.4. This apparatus
consists of the following components: a Welch Duo Seal Vacuum Pump Model 1405,
a vacuum desiccator, a vacuum trap, and a McLeod Gauge. Not shown is a dewar
filled with cold liquid (liquid nitrogen or other liquid) into which the vacuum trap is
immersed.
A frozen soil sample is placed into the vacuum desiccator. The vacuum trap
is placed into a cold temperature dewar and the vacuum pump is turned on. The
apparatus then sublimates off the frozen liquid. The gas formed during sublimation
will be frozen in the vacuum trap, thus recovering the frozen liquid and keeping
water or organic chemicals out of the vacuum pump.
5.6.2 Freezing drying procedure
Freeze drying a soil specimen is done as an attempt to retain the microfabric
and microstructure of the soil. This is best done by freezing the soil sample as
quickly as possible (Smart and Tovey, 1982). When freezing specimens saturated
with various chemicals, caution must be observed in order not to create dangerous
or explosive situations. It was observed in this investigation that when using liquid
nitrogen to freeze soil specimens saturated with water, and contained in test tubes,
that air was liquifying in the process and collecting in the test tubes. This also
occurred in the vacuum trap when just pumping down the apparatus. The liquid
formed was probably liquid oxygen and liquid nitrogen. In freezing soil specimens
saturated with organic liquids, this creates a high potential for explosions as organic
57


Figure 5.3 Freeze drying apparatus


1
Figure 5.4 Photograph of freeze drying apparatus.
59


liquids are incompatible with oxidizing materials, such as liquid oxygen. For this
reason, liquid nitrogen was not used for freezing the soil specimens when saturated
with organics. Instead, a mixture of C02 (dry ice) and 2-butoxyethanol was used
giving a mixture with a temperature of -78C. Therefore, freezing time was
increased as a tradeoff for a safer freezing procedure. The freezing and freeze drying
procedure follows:
1. Remove the soil specimen carefully from the triaxial cell and latex membrane.
2. Carefully cut the soil into small sticks approximately 3-5 mm x 10 mm. Place
the small sticks into a small test tube.
3. Immerse the test tube into the slurry mixture of dry ice and 2-butoxyethanol.
Leave in the slurry for three to fifteen minutes.
4. Remove the test tube from the slurry mix and remove the frozen specimen.
5. QUICKLY fracture the specimen and place it in the vacuum desiccator.
6. Turn on vacuum pump. Sublimate the frozen liquid from the specimen while
capturing the liquid in a vacuum trap immersed in the slurry of dry ice and
2-butoxyethanol.
7. The soil specimen should dry in about 3-5 hours.
8. When the specimen is dry, isolate the vacuum desiccator and turn off the
vacuum pump. Slowly bleed air back into the system.
9. Slowly bleed air into the vacuum desiccator and remove the specimen.
10. The soil specimen is now ready to be prepared for scanning electron
microscopy analysis.
5.7 Scanning Electron Microscopy
5.7.1 Scanning electron microscope and equipment
The scanning electron microscope used was a Cambridge Stereoscan 250 Mark
II fitted with a Tracor Northern 5500 energy dispersive spectrometer. An SPI gold
sputter coater and carbon coater were used to make the soil specimens electrically
conductive.
5.7.2 Scanning electron microscopy procedure
1. Remove freeze dried sample from vacuum desiccator of freeze dryer and glue
to aluminum SEM mount with Duco cement. Allow cement to dry.
60


2. Place sample into SPI carbon coater and coat the sample with carbon.
3. Place sample into SPI gold sputter coater and coat the sample with gold. (It
was found that a coating of carbon followed by a coating of gold produced
better electrical conductivity of the sample. In this way, better SEM
photomicrographs were obtained than using a single coating of either carbon
or gold.)
4. Place the sample into the sample chamber of the SEM.
5. Adjust the SEM to give the best photomicrographs. (The spot size of the
electron beam was the most useful adjustment in obtaining good
photomicrographs. Occasionally, decreasing the accelerating voltage worked
also. Each sample acts a little differently, so adjustments need to be
performed for each sample individually.)
6. Take a photomicrograph at various magnifications.
5.8 Chemical Analyses
Chemical analyses for the permeation of industrial waste leachate through the
Highlands Ranch clay and the Highlands Ranch clay admixed with the South Dakota
zeolite were obtained in order to demonstrate the capabilities of the compatibility
apparatus for plotting breakthrough curves. Breakthrough curves represent the value
of C/Cc versus time or pore volumes of influent. C represents the concentration of
cation in the effluent, and C represents the initial concentration of cation present in
the influent. Typical breakthrough responses plot as S-shaped curves. Complete
breakthrough is achieved when C/C0 = 1, with the advective front occurring at a
value of C/C0 = 0.5. Cation exchange and dispersion processes result in attenuation
and retardation of C/C0 values.
The initial demineralized water, initial leachate, and the effluent samples were
analyzed for Na, Ca, K, Mg, Fe, Pb, Cu, As, and Zn using atomic absorption (AA)
methods at Skyline Laboratories, Inc., Wheatridge, Colorado. All water samples
were acidified with nitric acid to prevent the precipitation of heavy metals. The
initial and final clay samples were also analyzed for Na, Ca, K, and Mg by
inductively coupled plasma spectrometry (ICP) at Skyline Laboratories, Inc. The
elements Fe, Pb, Cu, As, and Zn in the initial and final clay samples were analyzed
61


using energy dispersive methods with a Cd109 radioisotope excitation source and Li-
drifted Si detector. The metal concentrations were obtained from spectra using the
SuperXap program of Yager and Quick (1992).
The pH of the effluent was measured with an Oakten pHTestr 1 digital pH
meter calibrated using the Oakten pH calibration kit. pH measurements of the clay
slurries were performed by mixing 2g of clay soil with 30ml of demineralized water
(pH = 7). The pH meter was then immersed in the slurry until a stable reading was
obtained.
5.9 Viscosity and Specific Gravity
The viscosities of fluids were determined using a Cannon-Fenske routine type
viscometer. Specific gravity of the fluids were determined using a Christian Becker
specific gravity balance.
62


6. Results
6.1 Introduction
The testing program for the volume controlled test specimens and the
compatibility test specimens include: mineralogy, geotechnical index properties,
compatibility tests with distilled water, acetone, methanol, industrial waste leachate,
and scanning electron microscopy of soil/clay microfabric.
This chapter presents the results of the mineralogical analysis by x-ray
diffraction, the geotechnical index properties, and the compatibility tests obtained for
demineralized water and the various permeant liquids used. Also, scanning electron
photomicrographs of the microtextures and microfabrics of the soils are presented.
6.2 Mineralogical Analysis
The x-ray diffraction (XRD) analysis of the sand fill material showed it to be,
in order of decreasing abundance: quartz and potassium feldspar (probably
microcline), plagioclase feldspar, and a trace of mica. XRD of a random mount of
the -200 mesh fraction showed no trace of clay minerals. The mineralogy of the sand
fill suggests that it came from a typical granitic material.
XRD of the palygorskite drilling mud showed it to be, in order of decreasing
abundance: predominantly palygorskite, trace of sepiolite, and minor smectite.
Sepiolite is a Mg-rich palygorskite-group clay mineral. The smectite was evident due
to the increase in d-spacing of its 001 reflection upon ethylene glycol solvation.
The XRD data confirm that the only clay minerals present in the palygorskite
admixtures of soil are that added from the palygorskite drilling mud. This is crucial
in the determination of the suitability of the commercially available palygorskite
drilling mud as an additive to clay soils for use in clay liner systems.
The natural clay soil from Highlands Ranch consists of, in order of decreasing
63


abundance: quartz, potassium feldspar, plagioclase feldspar, calcite and a trace of
anhydrite. The two-micron clay fraction consists predominantly of calcium-smectite
and kaolinite, with traces of a highly illitic, illite/smectite and chlorite.
The South Dakota zeolite (SDA) consists predominantly of the zeolite mineral
clinoptilolite. Dilutant minerals in the SDA sample are, in order of decreasing
abundance: quartz, calcite, plagioclase feldspar, potassium feldpsar, and opal (?)
(Desborough, 1994).
6.3 Geotechnical Index Properties
A summary of the geotechnical index properties for the sand fill material and
the palygorskite admixtures is given in Table 6.1. The geotechnical index properties
of the HRC and the admixed HRC with ten percent SDA are given in Table 6.2.
64


Table 6.1 Geotechnical index properties of the Sand Fill material and the three
Palygorskite admixtures. Paly = Palygorskite. NP = Non-plastic.
Sand Fill + 10% Paly +20% Paly +30% Paly
clay size < .002mm 0 % 6 % 11 % 17 %
silt size .002-.074mm 1.1 % 8 % 12 % 15 %
sand size > .074mm 98.1 % 84 % 77 % 68 %
Liquid Limit NP 63 % 96.5 % 137 %
Plastic Limit NP 37.8 % 56.3 % 78.5 %
Plasticity Index NP 25.2 % 40.2 % 58.5 %
Specific Gravity 2.67 2.64 2.61 2.59
Optimum moisture content (OMC) 12 % 14.25 % 16 %
Porosity @ OMC 29.6 % 33.6 % 37.4 %
Porosity @2% wet of OMC 30.5 % 33.9 % 38.2 %
Max. dry density, g/cm3 1.86 1.73 1.62
Activity 4.2 3.7 3.44
USCS Classification SP SM SM SM
65


Table 6.2 Geotechnical index properties for the Highlands Ranch Clay and the
Highlands Ranch Clay admixed with ten weight percent South Dakota zeolite. HRC
= Highlands Ranch Clay, SDA = South Dakota zeolite.
HRC HRC + 10% SDA
clay size < .002mm 27 % 23 %
silt size .002-.074mm 58 % 55 %
sand size > .074mm 15 % 22 %
Liquid limit 53.3 % 53.6 %
Plastic limit 18 % 20 %
Plasticity index 35.3 % 33.6 %
Specific gravity 2.67 2.61
Optimum moisture content 22.7 % 22.75 %
Porosity @ OMC 39.7 % 38.7 %
Porosity @2% wet of OMC 40.9 % 40.0 %
Maximum dry density, g/cm3 1.611 1.600
Activity 1.31 1.46
USCS Classification CH CH
66


6.4 Hydraulic Conductivity of Volume Control Specimens
The results of the volume controlled hydraulic conductivity tests for the 10%
and 20% palygorskite admixtures are found on the compatibility test graphs in
Appendix A. The results for the 30% palygorskite admixture under volume control
conditions are shown in Figures 6.1, 6.2, and 6.3. Figure 6.1 shows the hydraulic
conductivity and effective stress versus time. Figure 6.2 shows the hydraulic
gradient and flow rate versus time. Figure 6.3 shows the hydraulic gradient versus
flow rate, which should plot as a straight line through the origin if Darcys Law is
valid under these testing conditions.
Figure 6.1 Plot of hydraulic conductivity and effective
stress versus time for 30% palygorskite admixture.
67


15
0.0003
10-
I s
£ -5
-10
-15

i*V*k*VV*
Hy*oUk yodlant
yr>+**~**t
Flow rote
12. 100 2 <*"< gw 12. SO 2
tefb* -H-H ifflw ^


-0.0002
-0.0001
-IE-04
0.0002
0 500 1000 1500 2000 2500 3000 3500
tlma (min)
-0.0003
Figure 6.2 Plot of hydraulic gradient and flow rate for
30% palygorskite admixture.
Figure 6.3 Plot of hydraulic gradient versus flow rate
for 30% palygorskite admixture.
68
flow rato (cubic cm/eoc)


6.5 Compatibility Tests
6.5.1 Scope of compatibility tests
The compatibility tests consisted of determination of the baseline hydraulic
conductivity of the soil admixtures with demineralized water, a reversal of flow to
check for the validity of Darcys Law, followed by determination of the hydraulic
conductivity with acetone, methanol, and industrial waste leachate. The properties
of the these fluids are given in Table 6.3.
Table 6.3 Properties of permeant fluids
viscosity @20 C (cp) density @ 20 C g/cm3 unit weight @ 20 C g cm'2 sec'2 dielectric constant Melting point 0 C
demineralized water .95 .9931 974 80 0
acetone .3265 .78 765 33.6 -95
methanol .5686 .78 765 33.6 -98
industrial waste leachate .9996 .9979 979 - -
6.5.2 Compatibility test data for palygorskite admixtures
The data for each test were plotted on three graphs. The first graph is a plot
of hydraulic conductivity and effective stress versus time and pore volumes of flow.
The second graph is a plot of flow rate and hydraulic gradient versus time and pore
volumes of flow. The third graph is a plot of intrinsic permeability versus time and
pore volumes of flow. Tests were performed at times varying from about eight hours
to about sixty hours. Typical plots for one test the 10 % palygorskite admixture
permeated with acetone are shown in Figures 6.4, 6.5, and 6.6. Plots for the
remaining tests can be found in Appendix A. Tables 6.3 and 6.4 summarize the
following parameters: the hydraulic conductivity with demineralized water (k, the
69


baseline hydraulic conductivity), the hydraulic conductivity with the permeant liquid
(kf), a ratio of the final hydraulic conductivity divided by the hydraulic conductivity
to demineralized water (kf/k,,), the initial and final intrinsic permeabilities ( Kc and
Kf, respectively), a ratio of the final intrinsic permeability divided by the initial
intrinsic permeability with demineralized water (K/KJ, the initial and final effective
stresses at the base of the specimen ( effective stress divided by the initial effective stress (oV hydraulic gradients (i0, if), the ratio of the final hydraulic gradient divided by the
initial hydraulic gradient (if/i0), and the flow rate (Q) for the palygorskite admixtures
of soil.
Figure 6.4 Hydraulic conductivity and effective stress
for the 10% palygorskite admixture permeated with
acetone.
70


0.004
Mu
Figure 6.5 Hydraulic gradient and flow rate for 10%
palygorskite admixture permeated with acetone.
Figure 6.6 Intrinsic permeability for the 10%
palygorskite admixture permeated with acetone.
71
flow rota (alle em/aae)


Table 6.4 Summary of compatibility test data for 10% palygorskite admixed soil.
IWL = Industrial waste leachate. Abbreviations defined in section 6.5.2.
K kf Kf
Constant stress
acetone 6E-5 7E-6 .12 5.5E-8 3.5E-9 .06
methanol 5.5E-5 6E-5 1.09 5.4E-8 4.4E-8 .81
IWL 1.5E-4 1.5E-4 1 1.5E-7 1.5E-7 1
Constant volume
acetone 2.6E-5 2E-6 .08 2.6E-8 8.5E-10 .03
methanol 4.5E-5 8E-5 1.78 4.4E-8 5.9E-8 1.34
IWL 1.4E-4 9.5E-5 .68 1.4E-7 9.7E-8 .69
Table 6.4 continued.
ff'o ff'A'o i0 h Q
Constant stress
acetone 5 5 1 2 1.5 .75 2.3E-3
methanol 5 5 1 2 2 1 2.3E-3
IWL 5 5 1 .9 .9 1 2.3E-3
Constant volume
acetone 2 20 10 2 23 11.5 9.23E-4
methanol 6 6 1 2.6 1.2 .46 2.3E-3
IWL 5.1 5.7 1.12 1 1.3 1.3 2.3E-3
72


Table 6.5 Summary of compatibility test data for 20% palygorskite admixed
soils. IWL = Industrial waste leachate. Abbreviations defined in section 6.5.2.
K Kf K/K,
Constant stress
acetone 2.4E-6 1.3E-6 .54 2.4E-9 5.5E-10 .23
methanol 2E-6 IE-6 .5 2E-9 7.4E-10 .37
IWL 6.5E-6 6E-6 .92 6.3E-9 6.1E-9 .97
Constant volume
acetone 1.2E-6 IE-7 .08 1.2E-9 4.5E-11 .04
methanol 2E-6 5E-7 .25 2E-9 3.7E-10 .19
IWL 1.5E-6 1.5E-6 1 9.75E-9 IE-8 1.02
Table 6.5 continued.
a'. ff'A'o /io Q
Constant stress
acetone 5 5 1 5 8 1.6 2.3E-4
methanol 5 5 1 6 12 2 2.3E-4
IWL 5 5 1 2 2.2 1.1 2.3E-4
Constant volume
acetone 5 28 5.6 3 50 16.6 9.23E-5
methanol 5 14 2.8 5.5 20 3.64 2.3E-4
IWL 5 5.5 1.1 1.3 1.3 1 2.3E-4
73


6.5.3 Compatibility test data for natural soil/zeolite admixture
The results of the compatibility test of the natural soil/zeolite admixture are
shown on Figures 6.7, 6.8, and 6.9. Figure 6.7 shows the hydraulic conductivity
and effective stress versus time and pore volumes of flow. Figure 6.8 shows
breakthrough curves for the elements Na, Ca, K, and Mg. Figure 6.9 shows
breakthrough curves for the elements Fe, As, Pb, Cu, and Zn. The time of the test
was about twelve days. Table 6.6 summarizes all chemical data for the HRC +
10%SDA permeated with industrial waste leachate. Table 6.7 shows net gains and
losses for specific cations analyzed for.
Figure 6.7 Hydraulic conductivity and effective stress
versus time and pore volumes of flow for the HRC +
10% SDA admixture.
74


2
(kr (mfei)
(Thousorxte)
Figure 6.8 Breakthrough curves of Na, Ca, Mg, and K
for the HRC + 10% SDA.
11m# (mfn)
(Thouwndi)
Figure 6.9 Breakthrough curves for Fe, Zn, Pb, Cu, and
As for the HRC + 10% SDA.
75


Tabic 6.6. Compilation of Chemical Analyses. HRC = Highlands Ranch Clay, SDA = South Dakota Zeolite (CRR), BDL = Below
Detection Limit. Values in ppm for solids, and mg/L for liquids._____________________________________________________________
Sample Ca Na K Mg Fe Cu As Pb Zn pH
Initial Concentrations in Soil
HRC 19000 7800 26000 8600 34300 BDL BDL BDL 79 8.4
SDA 14000 25000 33000 5000 12700 BDL BDL BDL 35 9.9
HRC + 1096SDA 18000 9600 24000 8000 31400 BDL BDL BDL 87 9.1
HRC 2micron 24000 4500 21000 15000 49000 BDL BDL BDL 153 8.6
Final Concentrations in Soil
HRC top 18000 7800 25000 8600 34800 BDL BDL BDL 84 8.1
HRC middle 19000 8300 25000 8300 31500 BDL BDL BDL 87 8.6
HRC bottom 19000 7600 24000 8700 34000 BDL BDL BDL 97 8.1
Fluid Analyses
Demin.water 1.45 4.7 0.5 0.45 BDL BDL BDL BDL 0.21 7.1
In.Leachate 807 475 45.9 242 BDL 0.02 0.0025 BDL BDL 7.1
Effluent tt\ 21.4 163 9.7 9.3 0.7 0.02 0.002 BDL 1.06 7.0
Effluent #2 74.8 424 26.4 30.8 8.6 BDL 0.002 BDL 0.82 7.1
Effluent #3 212 706 29.8 84 15.3 0.04 0.003 BDL 0.78 7.1
Effluent #4 390 862 32.8 152 25.6 BDL 0.004 BDL 0.64 7.0
Effluent US 527 818 23.8 218 38.2 BDL 0.006 BDL 0.45 7.0
Effluent H6 593 776 28.4 247 42 BDL 0.007 BDL 0.15 6.9


Table 6.7. Net gains (+) and losses (-) of elements for the top, middle, and bottom segments of
the soil specimen. All values in parts per million. [BDL = Below Detection Limits]__________
Ca Na K Mg Fe Cu Pb As Zn
HRC top -1000 0 -1000 0 +500 BDL BDL BDL +5
HRC+1096SDA middle + 1000 -1300 + 1000 +300 + 100 BDL BDL BDL 0
HRC bottom 0 -200 -2000 + 100 -300 BDL BDL BDL + 18
6.6 Scanning Electron Microscopy Analyses
The scanning electron microscopy analyses concentrated on the microfabric
of the test soils. The microfabric consists of the regular aggregations of particles and
the very small pores between them. Typical fabric units are up to a few tens of
micrometers across (Mitchell, 1993). Figure 6.10 shows a typical photomicrograph
of a water saturated and freeze dried palygorskite admixed soil specimen. The
remainder of the SEM photomicrographs of the palygorskite admixed soil specimens
permeated with the different fluids can be found in Appendix B.
77


Figure 6.10 Photomicrographs of palygorskite admixed soil after permeation with
demineralized water, frozen in liquid nitrogen, and freeze dried. (A) Blow up of
quartz grain from (B) on right side of photo, and palygorskite on left side of photo.
Twenty micron scale bar, 780X magnification. (B) Fine grained intergranular
palygorskite. 200 micron scale bar, 78X magnification.
78


7. Discussion of Test Results
7.1 Introduction
This chapter discusses the results obtained by the mineralogical analyses, the
geotechnical index tests, hydraulic conductivity of volume controlled specimens,
compatibility tests, and the SEM analyses.
7.2 Mineralogical Analyses
7.2.1 Palygorskite admixtures
The mineralogy of the sand fill (SF) material was clay free. The addition of
the palygorskite drilling mud created a soil consisting of granitic minerals,
palygorskite, and other minor constituents (sepiolite, smectite). Although it is highly
unlikely to find such a soil in nature, this admixture allows the study of the effects
of demineralized water, acetone, methanol, and an industrial waste leachate on the
palygorskite soils hydraulic conductivity, effective stress (taken at the base of the
soil specimen), and microfabric. The palygorskite gives this admixed soil its
geotechnical engineering characteristics.
7.2.2 Natural soil/zeolite admixture
The Highlands Ranch clay is defined as a fat clay, i.e. enriched in smectite.
The major clay constituent is calcium-smectite (typical of the Front Range area), with
kaolinite, mixed layer illite/smectite, and chlorite also being present. This is an
expansive soil. The addition of the South Dakota zeolite (a natural clinoptilolite rich
rock), did not add additional expansive clay to the soil. However, the cation
exchange capacity of the admixed soil was increased.
79


7.3 Geotechnical Index Tests
7.3.1 Palygorskite admixtures
As can be seen in Table 6.1, the geotechnical index properties of the three
palygorskite admixtures are behaving as one would expect for an increasing amount
of the percentage of palygorskite present. The clay and silt size fractions increase
and the sand size fraction decreases for increasing amounts of palygorskite. Also,
the liquid limit, plastic limit, and plasticity index all increase with increasing amount
of palygorskite. Figure A-l in Appendix A shows a graph of Atterberg limits versus
water content for the three admixtures. It is apparent that the liquid limit increases
faster than the plastic limit and plasticity index for increases in amount of
palygorskite. It appears that the sorptive properties of the palygorskite are such that
the activity decreases with increasing clay content. All of the admixtures were
classified as silty sands (SM), and the sand fill material was classified as a poorly
graded sand (SP).
7.3.2 Natural soil/zeolite admixture
As can be seen in Table 6.2, the geotechnical index properties of the
Highlands Ranch Clay were not significantly altered by the addition of ten weight
percent SDA (< No. 4 sieve, 4.75mm). The optimum moisture contents (OMC) for
both soils were about 23 percent. The maximum dry density at OMC for the HRC
+ 10% SDA was within one percent of the HRC alone. The porosities of both soils
at OMC were essentially the same, about 39 percent. The specific gravity of the
HRC + 10 % SDA mix was lower due to the lower specific gravity of the SDA.
The percentage of material smaller than the no. 200 sieve size indicates that the soil
would be adequate for use in clay liners (EPA/600/R-93/182).
According to the Environmental Protection Agency, soils with a high plasticity
index (>30% to 40%) tend to form hard clods when dry and sticky clods when wet
(EPA/600/R-93/182). Also, highly plastic soils tend to shrink and swell when dried
80


or wetted. Both soils studied were classified as fat clays with sand (CH). The HRC
soil is a highly plastic soil and its amendment with SDA resulted in a 2% decrease
in the plasticity index. This behavior will help render this material more workable
in the field and more suitable for use in clay liner systems.
7.4 Hydraulic Conductivity of Volume Control Specimens
The constant volume tests were performed using demineralized water with the
three soil admixtures of palygorskite. The 10% and 20% mixes were tested prior to
leaching with a second permeant fluid (these graphs can be found in Appendix A),
and the 30% mix was tested with demineralized water only.
The 30% admixture makes an excellent example to qualitatively understand
what occurs during a flow pump test (refer to Figures 6.1 and 6.2). The test was
performed using a flow rate of 9.23 x 10'5 cm3/sec (gear 12, 100%, on the flow
pump, with a one inch diameter bore "Olsen" actuator) in an upward direction. The
hydraulic conductivity decreases as the hydraulic gradient (dependent on the pressure
difference) between the top and bottom of the specimen approaches a constant value.
This represents steady state flow, and the hydraulic conductivity has then reached
equilibrium. The initial effective stress, 5 psi, decreases slightly due to the increase
in pressure at the base of the specimen. At approximately 1350 minutes, the
direction of flow through the sample was reversed (downflow), so the hydraulic
conductivity is now the negative of the upflow value. The effective stress has
increased to about 5.3 psi, due to the elevated pore pressure at the top of the sample
(now the influent side) and the decrease in pore pressure at the base of the sample.
Subsequent reversals in flow direction give similar results. Cutting the flow rate in
half decreases the hydraulic gradient in half, but results in the same value of
hydraulic conductivity. It should be noted that a plot of hydraulic gradient versus
flow rate should result in a straight line passing through the origin for Darcys Law
to be valid. Figure 6.3 shows a least squares regression line through the data points
81


for the 30% admixed sample. The plot is a straight line and goes through the origin,
proving that the hydraulic gradient is directly proportional to the flow rate and
Darcys Law is indeed valid for this test. The 30% admixture equilibrated at a
hydraulic conductivity of about 6 x 10'7 cm/sec and at a hydraulic gradient of about
7.5.
The 10% and 20% samples were tested for Darcys Law validity prior to
permeation with a second fluid. All samples passed this test at hydraulic gradients
ranging from about 0.9 to 6. The samples were not tested under reverse flow
conditions after the second permeant fluid was introduced. The 10% and 20%
samples equilibrated at average values of hydraulic conductivity of about 7 x 10'5
cm/sec and 1.6 x 10-6 cm/sec, respectively. The three admixtures initially differed
by about one order of magnitude when permeated with demineralized water.
It should be noted that incomplete saturation of the soil specimen gives
erroneous results under constant volume conditions. Consider the 10% admixture
permeated with demineralized water followed by acetone. The initial effective stress
was applied at a value of 5 psi. After allowing the specimen to equilibrate, a check
for leaks indicated that the system was not stable, i.e., the effective stress was
decreasing. This was probably due to air in the system going into solution. After
allowing to equilibrate for about one week, the effective stress decreased to 2 psi.
Upon checking for leaks in the system, the effective stress was now stable and the
test began. This sample had been saturated in the vacuum desiccator for a period of
three days. Subsequent samples were left in the vacuum desiccator for five days and
no problems occurred thereafter.
7.5 Compatibility Tests
7.5.1 Palygorskite Admixtures
The 10% and 20% admixtures were permeated with demineralized water
followed by acetone, methanol, or industrial waste leachate. These compatibility tests
82


were performed under volume control, and stress control conditions. It was
discovered that a true "volume control" condition may not be realized during
compatibility tests due to the interaction of the permeant chemical and water. Stress
controlled tests allow volume change of the specimen, and the sample behavior is
quite different. The qualitative principles discussed in section 7.4 are employed in
the following discussions. Graphs for all tests can be found in Appendix A. Large
aberrations in an otherwise equilibrated response curve represents the interruption of
flow to refill the actuator with permeant. There is one case in which this is not true,
and will be discussed in the methanol section below. Both acetone and methanol
would discolor the teflon tape surrounding the soil samples. It was observed that
even after several pore volumes of acetone or methanol, the teflon tape was only
partially discolored. This phenomena is interpreted as the occurrence of channel flow
within the soil sample. The industrial waste leachate did not discolor the teflon tape,
so there is no indicator for channel flow in this case.
7.5.1.1 Acetone
The constant volume hydraulic conductivities of the 10% and 20%
palygorskite admixtures permeated with demineralized water followed by acetone may
have been affected by acetone/water interactions. Air was observed in the effluent
line within the flexible wall permeameter (triaxial cell) after the acetone leaching was
begun. The air was not there during initial permeation with demineralized water.
This did not occur in the stress control specimens, nor in the methanol or industrial
waste leachate tests. Upon mixing 20ml of deaired/demineralized water with 20 ml
of acetone in a graduated cylinder, it was found that 38.5ml of solution was obtained.
The mixing of fluids resulted in a small volume change of the soil/fluid system.
However, the results obtained for the hydraulic conductivity, effective stress, and
hydraulic gradients all showed similar trends as in the other specimens.
The hydraulic conductivity decreased by an order of magnitude, from about
83


2.6 x 10"5 cm/sec to about 2 x 10-6 cm/sec for the 10% admixture, and about 1.2 x
lO-6 cm/sec to about 1 x 107 cm/sec for the 20% admixture. In both cases the
intrinsic permeability decreased by more than an order of magnitude. The effective
stress and the hydraulic gradients both increased. The increase in hydraulic gradient
is a result of a decrease in pressure at the base, and a resulting increase in differential
pressure across the sample. As acetone leaching begins, the double layer thickness
around the clay particles is reduced at the influent end of the sample. This is due to
the lower dielectric constant of the acetone. Since the differential pressure has
increased, the hydraulic conductivity decreases. Also, the effective stress rises as the
base pressure decreases.
The constant stress tests with acetone showed no changes in effective stress.
Since the chamber pressure and the base pressure are controlled, this is expected.
The hydraulic conductivity for the 10% admixture decreased from about 6 x 105
cm/sec to about 7 x 10^ cm/sec, approximately one order of magnitude. The
hydraulic conductivity for the 20% admixture decreased from about 2.4 x 10-6 cm/sec
to about 1.3 x 10"* cm/sec, approximately by a factor of one half. The hydraulic
gradient increased by a factor of 1.6 for the 20% admixture but decreased slightly
for the 10% admixture. This slight difference may be due to channel flow, or
viscosity effects, since the change in intrinsic permeability was also more for the 10%
admixture than the 20% admixture.
7.5.1.2 Methanol
The hydraulic gradient during the constant volume test for the 10% admixture
exhibited a genuine reversal in direction in which the base pressure was elevated.
This is also evident in the hydraulic conductivity measurement. This was not an
interruption in the test to refill the actuator. Instead, this phenomenon is attributed
to a temporary fabric adjustment of the soil sample. Upon further leaching, the soil
fabric reequilibrated, and the trend of a decrease in base pressure resumed. The
84


hydraulic conductivity of the sample increased slightly from about 4.5 x 10'5 cm/sec
to about 8 x 10'5 cm/sec. The effective stress showed a continual increase due to a
decrease in base pressure, with a marked change in slope concurrent with the fabric
adjustments. The intrinsic permeability increased slightly.
The volume controlled test for the 20% admixture gave a decrease in
hydraulic conductivity from about 2 x 10"6 cm/sec to about 5 x 10'7 cm/sec. Similar
trends of increasing effective stress and increasing hydraulic gradient were also
observed. At approximately two pore volumes of total flow, the hydraulic gradient
began to decrease, and the hydraulic conductivity began to increase. This is
attributed to fabric changes. The intrinsic permeability decreased by almost an order
of magnitiude.
The stress controlled test on the 10% admixture did not have much effect on
the sample. The hydraulic conductivity increased slightly, from about 5.5 x 10'5
cm/sec to about 6 x 10 s cm/sec. The hydraulic gradient increased slightly when
initially permeated with methanol, but reequilibrated at the same value as the
demineralized water, about 2. The intrinsic permeability dropped slightly due to
viscosity effects. The hydraulic gradient of the 20% admixture increased from 6 to
about 12. The hydraulic conductivity was reduced by one half, from about 2 x 10^
cm/sec to about 1 x 10"6 cm/sec. The intrinsic permeability dropped more due to
viscosity effects. In both cases the effective stress remained at about 5 psi, as
expected.
7.5.1.3 Industrial waste leachate
The properties of the industrial waste leachate were very similar to that of
water. However, it did have some organic components (oils). The constant volume
tests for the 20% mix showed no appreciable changes in hydraulic conductivity (lx
10 s cm/sec), intrinsic permeability (1 x 10'8 cm2), or hydraulic gradient (1). The
effective stress increased by a very small amount. The hydraulic gradient of the 10%
85