Prediction of hydraulic conductivity of clay liners

MISSING IMAGE

Material Information

Title:
Prediction of hydraulic conductivity of clay liners a field and laboratory study
Physical Description:
xvi, 300 leaves : ill. ; 29 cm.
Language:
English
Creator:
Al-musawe, Sadik Jaffer
Publication Date:

Subjects

Subjects / Keywords:
Soil permeability   ( lcsh )
Sanitary landfills   ( lcsh )
Civil Engineering thesis Ph. D
Dissertations, Academic -- Civil Engineering -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 186-193).
Statement of Responsibility:
by Sadik Jaffer Al-Musawe.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 025013993
oclc - 24160932
System ID:
AA00013542:00001


This item is only available as the following downloads:


Full Text












PREDICTION OF HYDRAULIC CONDUCTIVITY OF CLAY LINERS:
A FIELD AND LABORATORY STUDY










By

SADIK JAFFER AL-MUSAWE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1990













To All Al-musawe Family Members:
Now and forever with all my love.














ACKNOWLEDGMENTS

First and foremost I would like to thank sincerely and

whole heartedly my advisor and supervisory committee

chairman, Dr. Paul Y. Thompson, for his much needed help,

guidance, and continuous supervision in every aspect of my

Ph.D. degree program and this research project. Dr. Thompson

was directly instrumental in my leaving Georgia Institute of

Technology and joining the University of Florida. At a

time when everything looked dark, he took me under his wing

and gave all and every support I needed to carry on. Without

him I could not have finish my Ph.D. at this University. He

is the one who selected this research project out of a number

of proposed ones. Dr. Thompson instantly put me in contact

with MFM Industries who ended up financing most of the

expenses of the research project. He was there whenever I

needed him during the course of this research. I do not have

proper words to express my gratitude to him, but I will say

that I shall always be his student, and he will always be my

professor.

Special thanks and acknowledgments go to the cochairman

of my supervisory committee, Dr. David Bloomquist, whose

willingness to help went above and beyond the call of duty.


iii







I am deeply grateful to him for his participation in most of

the discussions concerning the experimental works. He

supplied me with a quick and instantaneous solution to every

problem I faced throughout the Ph.D. program and the research

project. Dr. Bloomquist was there for me whenever I needed

him. I always have and will consider him as friend.

I would like to sincerely thank the members of my

supervisory committee: Chairman of Department of Geology, Dr.

Anthony F. Randazzo, Professor Wally H. Zimpfer, and Dr.

Fazil T. Najafi for their invaluable comments during frequent

discussions about various aspects of this research project.

I am deeply grateful for their encouragement and moral

support during the whole of my Ph.D. program. I shall

never forget their friendship.

Sincere thanks and appreciation go to MFM (Mid Florida

Mining) Industries located in Ocala, Florida, for their

sponsorship of this research. MFM Industries have supplied

me with all the materials that I needed for testing and paid

all the expenses that I incurred in the course of this

research. Special thanks and acknowledgments go to the

former president of MFM Industries, Mr. Allen Edgar, and Mr.

Allen Stewart, P.E, Project Manager with MFM Environmental,

for their continuous support in each and every aspect of the

research project. Their frequent comments and inputs were

invaluable. Without them this research would not have been

possible.







Many thanks and appreciation go to Messrs. James B.

Abbott, P.E (Assistant Public Works Director) and Allen

Ellison (Landfill Operations Supervisor) of Waste Management

Department, Alachua County; Miss Claire E. Bartlett, Director

of Solid Waste Department, Lake County; and Mr. Earl Holmes

of ERC, Inc., in Orlando for their invaluable support for the

field work. Without them all field work would not have been

possible. They also supplied me with all the field

documentation about the S.W. Alachua and Astatula Landfills.















TABLE OF CONTENTS


ACKNOWLEDGEMENTS .......................... ..... ... iii

LIST OF TABLES ........................................ viii

LIST OF FIGURES........................................ .. ix

LIST OF SYMBOLS ...................................... xiii

ABSTRACT ............................................... xv

CHAPTERS

1 LITERATURE REVIEW, BASIC CONCEPTS, AND PURPOSE
AND SCOPE OF THIS STUDY.................... ...... 1

Definition of the Problem........................ 1
Clay Liner and Landfill Technology............... 3
Background Information of Previous Work Related
to this Study ....................... ........... 4
Purpose and Scope of this Research Project....... 43

2 BULK SAMPLING, PROPERTIES, AND SAMPLE
PREPARATION ....................... ............... 78

Bulk Sampling. ................................ 78
Properties of the Project Clay.................. 79
Sample Preparation ............................... 81

3 LABORATORY TESTS, RESULTS, AND DISCUSSION........ 95

Laboratory Hydraulic Conductivity Tests........... 95
Soil Suction and Saturation vs. Density vs.
Moisture Content................................ 97
Hydraulic Conductivity vs. Sample Thickness....... 106
Hydraulic Conductivity vs. Number of Layers...... 111
Hydraulic Conductivity vs. Hydraulic Gradient.... 116







Conductivity vs. Unit Weight vs. Time............ 118
Moisture Content Distribution After Conductivity
Tests ................... ............... ..... 121
Laboratory Desiccation Tests ..................... 122

4 FIELD WORK, RESULTS, AND DISCUSSION............... 140

Field Infiltration Tests ............ ......... .... 140
Southwest Alachua Landfill-Top Cover.............. 147
Astatula Ash Residue Monofill-Bottom Liner....... 151

5 CONCLUSIONS AND RECOMMENDATIONS................. 179

Conclusions ...................................... 179
Recommendations................................... 184

REFERENCES.............. ................................. 186

APPENDICES

A PHYSICAL AND INDEX PROPERTIES OF THE RESEARCH
CLAY.................................. ............ 194

B MINERAL AND CHEMICAL PROPERTIES OF THE
RESEARCH CLAY.................................... 199

C PROJECT NO. 1: SOUTHWEST ALACHUA LANDFILL
TOP COVER ..................................... .. 218

D PROJECT NO. 2: ASTATULA ASH RESIDUE MONOFILL
(HAZARDOUS SOLID WASTE) ............................... 251

BIOGRAPHICAL SKETCH ............ ...................... 299


vii












LIST OF TABLES


Table Page

1 Methods of Measuring Suction.................... 47

2 Saturated Salt Solution versus Relative
Humidity......................................... 48

3 Various Parameters for Three Permeameters ...... 40

4 Comparison of Range of Index and Physical
Properties of the Project Clay.................. 92

5 Average Temperature vs. Depth Along Soil Sample. 125

6 Comparison of Conductivity Values Obtained by
Different Methods (S.W. Alachua Landfill-Top
Cover)........................................... 164

7 Comparison of Conductivity Values Obtained by
Different Methods (Astatula Field Test Strips).. 165

8 Comparison of Conductivity Values Obtained by
Different Methods (Astatula Western and Eastern
Evaporation Basins) ............................. 166


viii












LIST OF FIGURES


Figure

1 Examples of Natural Liners........................

2 Types of Compacted Liners.........................

3 Typical Landfill Section and Components............

4 Hydrological Cycle as Applied to Landfill System..

5 Zones of Laminar and Turbulent Flow................

6 One-Dimensional Schematic of Consolidation Cell
Permeameter................................... ..

7 Schematic of Flexible Wall Permeameter............

8 Schematic of Rigid Wall Permeameter...............

9 Schematic of Mariotte Tube.........................

10 Schematic of Single and Double Ring
Infiltrometers....................................

11 Schematic of a Sealed-Double Ring Infiltrometer...

12 Soil Suction versus Water Content................

13 Soil Suction versus Conductivity..................

14 Scales for Reporting Suction Values................

15 Filter Paper Calibration Curves....................

16 Conductivity vs. Dry Unit Weight vs. Molding
Water Content for Two Different Clays..............

17 Summary of Laboratory and Field Infiltration
Tests ................. ..... ...................

18 Conductivity vs. Confining Pressure................


Page

50

51

52

53

54


55

56

57

58


59

60

61

62

63

64


65


66

67







19 Conductivity vs. Degree of Saturation vs. Aging... 68

20 Conductivity vs. Sample Diameter.................. 69

21 Conductivity vs. Aging............................ 70

22 Conductivity vs. Sample Height................... 71

23 Conductivity vs. Plasticity Index................. 72

24 Conductivity vs. Pore Volume...................... 73

25 Schematic of Single Ring Infiltrometer and
Suction Head....................................... 74

26 Suction vs. Water Content......................... 75

27 Distribution of Soil Saturation after Field
Infiltration Tests ............................... 76

28 Field Conductivity vs. Time...................... 77

29 Cross Section of the Laboratory Rigid Wall
Permeameter............................... ..... 93

30 Cross Section of the Steel Sleeves Used in Field
Infiltration Test and Undisturbed Sampling........ 94

31 Degree of Saturation vs. Dry Unit Weight vs.
Moisture Content .................................. 126

32 Suction vs. Filter Paper Water Content............ 127

33 Soil Suction vs. Dry Unit Weight vs. Moisture
Content .......................................... 128

34 Hydraulic Conductivity, Dry Unit Weight,
Saturation, and Porosity vs. Sample Thickness..... 129

35 Hydraulic Conductivity, Dry Unit Weight,
Saturation, and Porosity vs. Number of Layers
for 1.5" Sample .................................... 130

36 Hydraulic Conductivity, Dry Unit Weight,
Saturation, and Porosity vs. Number of Layers
for 4.6" Sample ................. .. ..... ......... 131







37 Hydraulic Conductivity, Dry Unit Weight,
Saturation, and Porosity vs. Number of Layers
for 12" Sample.................................... 132

38 Hydraulic Conductivity vs. Hydraulic Gradient
for 4.6" One Layer Sample ......................... 133

39 Hydraulic Conductivity vs. Hydraulic Gradient
for 4.6" Three Layer Sample....................... 134

40 Hydraulic Conductivity vs. Elapsed Time............ 135

41 Moisture Content vs. Depth of 1.5" Sample.......... 136

42 Moisture Content vs. Depth for 12" Sample......... 137

43 Hydraulic Conductivity vs. Elapsed Time for
Desiccated Sample ................................. 138

44 Moisture Content vs. Depth for Desiccated Sample
Before and After Hydraulic Conductivity Test...... 139

45 Field Infiltration Test Setup...................... 167

46 Location and Vicinity Map of S.W. Alachua
Landfill........................................... 168

47 Field Infiltration Test Locations and Cross
Section (S.W. Alachua Landfill-Top Cover)......... 169

48 Various Scales of Reporting Hydraulic
Conductivity Values .............................. 170

49 Location and Vicinity Map of Astatula Ash
Residue Monofill Landfill......................... 171

50 General Location of Test Strips, Landfill, and
Evaporation Basins ................................ 172

51 Test Strips Showing Dimensions and Locations of
All Performed Field Tests........................ 173

52 Schematic of Typical Soil Block Showing All
Dimensions ........................................ 174

53 Average Dry Unit Weight vs. Depth of Soil Block... 175

54 Average Moisture Content vs. Depth of Soil Block.. 176







55 Typical Desiccation Crack Study Location and
Cross Section .................................. 177

56 Hydraulic Conductivity vs. Hydraulic Gradient on
Field Obtained Sample (Astatula Western
Evaporation Basin) ................ ..... ............ 178


xii













LIST OF SYMBOLS


A Cross Sectional Area of Soil Sample

Ac Percent Activity of Soil

Ad Discharge Area and Equal to A

Ar Percent Area Ratio

As Seepage Area

a Cross Sectional Area of the Small Standpipe

D Diameter of Soil Sample

e Void Ratio of Soil

Gs Specific Gravity of Soil Solid

Ho Hydraulic Head Difference Applied to Soil Sample

Hs Suction Head Within Soil Sample

i Hydraulic Gradient

K Steady State/Saturated Hydraulic Conductivity
(Permeability)


Ki Transient Hydraulic Conductivity/Coefficient of
Infiltration

L Length of Soil Sample

LL Percent Liquid Limit

n Percent Porosity

ne Percent Effective Porosity


xiii







PL Percent Plastic Limit

PI Percent Plasticity Index

Q Quantity of Water Discharged

R Drainage Impedance

S Percent Degree of Saturated

T Time

V Velocity of Discharge

Vs Velocity of Seepage Discharge

w Percent Moisture Content

Yd Dry Unit Weight

yw Wet/Moist/In-situ Unit Weight


xiv












Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PREDICTION OF HYDRAULIC CONDUCTIVITY OF CLAY LINERS:
A FIELD AND LABORATORY STUDY

By

Sadik Jaffer Al-musawe

December 1990

Chairman: Dr. Paul Y. Thompson
Cochairman: Dr. David Bloomquist
Major Department: Civil Engineering

Low hydraulic conductivity clay soils are used in

landfills to impede the movement of leachate down to the

natural groundwater table. Hence, hydraulic conductivity is

an important soil property in the design and assessment of

liner thicknesses and integrity. Because of the sensitivity

of hydraulic conductivity to many factors, there are no

standard laboratory or field testing methods, and therefore,

there exists wide variation in predicted values.

A local natural clay soil, "Terra-Seal Natura Premix@,"

was used for this study because this soil is used in the

construction of a number of landfills in Florida. A series

of rigid-wall permeameter tests were performed for a

quantitative prediction of hydraulic conductivity variation

as a function of sample thickness, number of layers,







hydraulic gradient, porosity, degree of saturation, dry unit

weight, and time. The variation of moisture content versus

depth for the fully saturated samples was found to vary

significantly. Measurements were also made of the variation

of partially saturated soil suction with dry unit weight and

moisture content.

A number of field infiltration tests were conducted at

two existing landfill projects, and coefficients of hydraulic

infiltration were measured. Using these values and the

amount of suction obtained in the laboratory, saturated

hydraulic conductivities were predicted. These predictions

agreed very closely with those obtained in the laboratory by

the author and others.

Desiccation cracks, depth versus dry unit weight, and

moisture content variations were studied in the field. Three

test strips, constructed using one, two, and three layers,

were subjected to equal compactive energies. Variations of

dry unit weight and moisture content with depth were found to

be the least for the one layer strip. The surfaces of all

test strips were cracks, of which the depth and width varied.

Covering the test strips with Visqueen did not prevent

cracking but did minimize it. In the laboratory, the effect

of these cracks on the hydraulic conductivity diminishes

after 16 hours of testing. Field hydraulic conductivity can

be predicted accurately and efficiently by the method

developed in this study.


xvi












CHAPTER 1

LITERATURE REVIEW, BASIC CONCEPTS, AND PURPOSE AND
SCOPE OF THIS STUDY


Definition of the Problem

Hydraulic conductivity of soil has become the most

important property in geotechnical and geo-environmental

engineering, agronomy, agriculture, and in all fields that

involve seepage and drainage of water and industrial liquids

through soil. Yet it is the most varied, least known, least

studied, and most difficult soil property to determine. In

one of the geotechnical and geo-environmental engineering

areas where a reliable and accurate estimate of the hydraulic

conductivity is most needed is in the determination of the

clay liner thicknesses. Clay liner is a soil layer of

certain thickness consisting of sandy silty clay with low

hydraulic conductivity. Clay-lined facilities have been used

extensively for the containment and disposal of hazardous and

nonhazardous solid and liquid waste. Occasionally slowly

permeable natural clay-rich deposits were relied upon to

retard the movement of leachate and liquids from landfills or

surface impoundments. Presently, in most cases, remolded

layers of soils with laboratory hydraulic conductivities of







1 10-7 cm/s or less have been used with the intention of

retaining leachate and liquids.

There is an increasing body of data which indicates that

hydraulic conductivity of in situ (recompacted) clays may be

greater than those measured on samples in the laboratory

(Daniel 1987, Mitchell 1976, Schmid 1966, Sowers 1979).

Although there is no set standard for laboratory hydraulic

conductivity tests on clays, all existing methods yield

comparable values. Major errors in the laboratory values are

due to the large sample disturbances, relatively small

dimensions of the tested samples, and the very large applied

hydraulic gradient.

On the other hand, proposed field methods are

complicated, difficult to run, time consuming, require

lengthy analysis, require highly technical personnel, very

sensitive to minor errors in the setup, and do not resemble a

laboratory setup (Chen et al. 1986, University of Texas,

College of Engineering 1990, Gorden et al. 1989, Hamilton et

al. 1979, Mitchell 1976, Olsen and Daniel 1979, Peirce et

al. 1987(b), Schmid 1966, Stewart and Nolan 1987, Wit 1966).

The major errors in field values of hydraulic conductivity

are mainly caused by soil suction (capillary pressure) which

is due to incomplete saturation of the soil and the ability

of the permeant liquid (water) to travel in both vertical and

horizontal directions (Daniel 1984, Stewart and Nolan 1987).








Clay Liner and Landfill Technology

A clay liner (sometimes referred to as soil or earthen

liner) may be manmade (compacted), or a naturally occurring

deposit (not disturbed). Natural clay liners are formed by

aquitards or aquiclude. Wastes may be buried wholly within a

natural clay liner (Fig. 1A ), partially within a natural

liner (Fig. 1B), or as in Fig. 1A and Fig. 1B but not within

a natural liner (Fig. lC).

Manmade liners consist of a horizontal liner, an

inclined liner, or a cover over a landfill (Fig. 2). The

soil in these liners can be either naturally occurring soils

or manmade soils by the mixing of natural soils with one or

more different materials. In either case, the soils must meet

set specifications concerning fineness content, clay content,

plasticity index, liquid limit, and moisture content. The

soil then is placed in horizontal layers with suitable

thicknesses and compacted to achieve a certain dry unit

weight.

A typical section of a landfill containment system,

including typical dimensions of various components, is shown

in Fig. 3. The clay liner impedes or controls outward

seepage of contaminant-laden fluids from the structure. The

leachate collection and removal system conveys fluids off the

clay liner to collection sumps and where the liquid is

removed. The final cover impedes or eliminates infiltration

of meteoric water into the refuse, thereby controlling







leachate generation. The entire concept of waste containment

is basically the successful interruption of the natural

hydrological cycle, as depicted in Fig. 4.



Background Information of Previous Work
Related to this Study

This investigation deals with the hydraulic conductivity

of naturally occurring soils as clay liners. Therefore, only

similar previous work will be dealt with in this literature

review. Note, however, that the concepts are the same in

either case. Although the principle of hydraulic

conductivity was recognized in 1911, its application to

landfill liners became extensive in the last 10 years when

landfill technology started to surface. Early work included

studies of simple prediction of the transient time of a

wetting front and the seepage rate after achieving saturation

(Green and Ampt 1911). This study is still frequently used

and commonly referred to as the Green-Ampt model.



Hydraulic Conductivity of Saturated Clay Soils

Hydraulic conductivity is the speed with which water

flows through soil media under unit hydraulic gradient. The

laws by which this flow takes place are very well understood

in sand and coarser grained soils but are still under debate

for clay soils. Flow can be classified as one-, two-, or

three-dimensional. One-dimensional flow is flow in which all

the fluid parameters, such as pressure, velocity,








temperature, etc., are constant in any cross section

perpendicular to the direction of flow. These parameters can

vary from section to section along the direction of flow but

are generally assumed to be constant. This in turn means

that the soil media is assumed to be homogeneous. In two-

dimensional flow, the fluid parameters are the same in

parallel planes, whereas in three-dimensional flow, the fluid

parameters vary in three coordinate directions. For the

purpose of analysis, in all the literature reviewed and in

all geotechnical engineering applications, flow problems are

assumed to be at most two-dimensional.

Flow can also be described as laminar (zone I, Fig. 5),

where the fluid flows in parallel layers without mixing, or

turbulent (zone III, Fig. 5), where random velocity

fluctuations result in mixing of fluid and internal energy

dissipation. There can also be intermediate or transition

states between laminar and turbulent flow. These states are

shown in Fig. 5. The flow in most soils is considered

laminar when the particle size is less than 0.05 cm and

uniform size, with a low seepage velocity, and a hydraulic

gradient (i) of one (Holtz and Kovacs 1981, Mitchell 1976,

Lambe and Whitman 1979, Sing 1967, Taylor 1948). In case of

clays, the flow is laminar when the particle size is 0.0002

cm or less, particles are not of uniform dimension, the

hydraulic gradient is always much greater than one, and the







seepage velocity is very high. D'Arcy (1856) showed

experimentally that for clean sands in zone I,



V = K i (1)



(Darcy's Law) where

K = hydraulic conductivity, saturated hydraulic

conductivity, Darcy coefficient of permeability, or

permeability (cm/s),

V = Q/A*T = discharge velocity (cm/s),

i = Ho/L = hydraulic gradient (cm/cm).



Therefore, equation 1 becomes



K = (Q L)/(A T Ho) (2)



where

Q = quantity of discharge (cm3),

A = cross sectional area of soil (cm2),

T = time (s),

Ho = hydraulic head difference applied to soil (cm), and

L = length of flow path in soil (cm).



Another concept in fluid mechanics is the law of

conservation of mass, and for incompressible steady state

flow; this law reduces to the equation of continuity:









q = Q/T = Vi AI = V2 A2 = constant


where

q = rate of discharge (cm3/s),

VI, V2 = velocities of flow at section 1 and 2,

A1, A2 = cross sectional areas of soil at section 1 and 2



The velocities of flow outside the soil media (V or Vd) are

not the same as the seepage velocity (Vs) of flow inside the

soil media, or


V = Vd = n Vs,


(4)


where n = percent porosity, and since the water can only seep

through the connected pores,


V = Vd = ne Vs,


(5)


where ne = percent effective porosity



Using the above notations and combining equations 1, 2, and

3,


q = V A = { [Q / (T A ne )]} *

A = (K Ho A)/L,


(6)


(3)







and by simplifying,



K = (Q L)/(A T Ho ne) (7)



Currently only equations 1 and 2 are used to obtain

hydraulic conductivity for all types of soil, including

clays.



Prediction of Hydraulic Conductivity of Saturated Clays

Techniques for measuring hydraulic conductivity in

coarse-grained soils (falling head or constant head tests in

the laboratory, and pump tests from wells in the field) are

very well established and standardized. Techniques for

measuring hydraulic conductivity in fine-grained (clays)

soils, however, are not very well known or standardized. The

reason for this is that past practice frequently has been to

assume that clays are effectively "impervious," and

therefore, attempts to measure their hydraulic conductivity

were not undertaken. But with the progression of landfill

use and technology, the need for measuring hydraulic

conductivity has become very important and vital in order to

monitor the water and leachate movements over many hundreds

of years, thereby providing parameters necessary to protect

the integrity of the groundwater below the landfill.








Empirical methods

Many empirical formulas exist for predicting the

saturated hydraulic conductivity of soils with a particle

size greater than 0.0002 cm (sand). This is mainly because

the sand has particles that are uniformly distributed and

spherical in shape which result in pores that are relatively

uniform in size and distribution. Clays, on the other hand,

have particles that are flaky, less than 0.0002 cm in size,

have large electrically charged surface areas, and many types

of pores with different sizes and distributions. However,

there are two known empirical equations that can be used with

large inaccuracy to predict the hydraulic conductivity (k) in

clayey soils. These are



1. Kozeny-Carman equation (Mitchell 1976, Sing 1967, Taylor

1948):



k = K (I/.1) = [1/(Po t2 So2)] [n3/(l-n)2] (8)



where

k = intrinsic permeability or permeability (terminology

used in hydraulic and fluid mechanics engineering) in

cm2,

p = viscosity of water in g s/cm2,

yw = unit weight of water in g/cm3,









Po = pore shape factor (2.5 for sand), and

t = tortuosity factor (20.5 for sand).



2. Loudon's Formula (Sing 1967):



Logio (K Ss2) = a + b n (9)



where

Ss = specific surface of soil particle in cm2/cm3,

a = constant = 1.365 at 10oC for sand, and

b = constant = 5.15 at 10oC for sand.



Laboratory methods

One-dimensional consolidation cell permeameter. A

typical one-dimensional consolidation cell permeameter

consists of a 4 to 10 cm diameter by 1.9 to 10 cm high

consolidation ring mounted in a cell as shown in Fig. 6. A

reservoir of water surrounding the consolidation ring

maintains atmospheric pressure at the effluent end (top) of

the specimen. The hydraulic pressure at the base of the

sample is controlled using the system described by Olsen and

Daniel (1979). The hydraulic conductivity can be calculated

by


K = (C cc j,)/(1 + e)


(10)








where

C = coefficients of consolidation in cm2/s,

cc = compressibility in 1/(g/cm2), and

e = void ratio.



Flexible-wall permeameter. Hydraulic conductivity tests

with flexible-wall (triaxial type) devices are performed

typically using the cell shown in Fig. 7. Interchangeable

base pedestals permit testing of specimens with a diameter

between 2 and 15.2 cm and with a wide range of heights

ranging from 2 cm to 20 cm. Double drainage lines to both

the top and bottom of the test specimen are used to flush air

bubbles out of the lines. The spare drainage lines are also

used in conjunction with an electrical pressure transducer to

measure pore-pressure response during back pressuring of the

soil, and to measure the pressure drop across the specimen.

There are two standard test types that use the flexible-wall

permeameter.

1. Constant head: In this test, the hydraulic gradient

(i = Ho/L) is maintain constant and the volume of discharge

(Q) is measured during a time (T). The hydraulic

conductivity is calculated using equation 2. For fine-

grained soils the constant head is typically applied using a

Mariotte bottle similar to that shown in Fig. 9. Such

equipment is designed to apply only small heads (a few feet

of water) so it is most useful with rather pervious soils or








in a case where prolonged testing times can be tolerated.

The main advantages of constant head tests are the simplicity

of interpretation of data and the fact that use of constant

head minimizes confusion due to changing volume of air

bubbles when the soil is not saturated.

2. Falling head: This is a more common test for fine-

grained soils in which the time (T) for the hydraulic head to

drop from one level (H1o) to a lower level (Ho2) in a

volumetric tube (typically a pipet or a buret with cross

sectional area (a)) due to flow through a soil sample of

cross sectional area (A), and length (L), is measured. The

hydraulic conductivity is calculated using



K = [(a L)/(A T)] In Hol/Ho2 (11)



The advantages of using this procedure are that small flows

are easily measured using the pipet or buret. The

observation time may still be long, in which case corrections

for water losses due to evaporation or leakage may be added.

The testing time may be reduced by increasing the flow rate

by superimposing an air pressure (Ap) on top of the water in

the pipet, thus increasing the heads by a certain amount

equal to Ap/y .

Rigid-wall permeameter. The Rigid-wall (compaction-

wall) permeameter consists of a 10 cm diameter compaction

mold with variable heights. The mold is clamped between two








acrylic end plates and sealed with either gaskets or O-rings,

as indicated in Fig. 8. The soil is either allowed or not

allowed to swell. The influent water is usually stored in a

separate device which contains an air-water interface inside

a glass pipet. The pressure acting on the water is

controlled with an air pressure regulator. Flow quantities

are measured by reading the position of the air-water

interface inside the pipet. The effluent water is collected

in a reservoir that is open to atmospheric pressure. The

drainage line leading to the permeameter is saturated with

water, but no back pressure is applied nor is the effluent

line de-aired. Both falling and constant head test methods

can be used with the rigid-wall permeameter as described

above together. The same equations are used to calculate the

hydraulic conductivity (i.e., equation 11).



Field methods

The soils in landfill liners are not saturated soils

and, therefore, cannot be considered saturated soils.

Consequently, the existing field methods of determining

hydraulic conductivity for saturated soils cannot be used.

However, if the soils are saturated, then the hydraulic

conductivity can be measured in the field by drilling a hole

in the ground, measuring the rate of flow of water into or

out of the hole and using an appropriate formula to calculate

the conductivity (Harr 1962, Lambe and Whitman 1979, Olsen








and Daniel 1979, Schmid 1966). Tests may be performed at a

constant head by establishing a high head of water in the

borehole and pumping at a rate sufficient to maintain this

head. Also, tests may be made with a variable head, that is,

with the head set at a nonequilibrium value initially and

then measured as a function of time with no further pumping.

Other field test methods are used and sometimes

erroneously called "field hydraulic conductivity tests."

These are actually field infiltration tests (ASTM 1989,

Bagchi et al. 1985, Bond and Collis-George 1981, University

of Texas, College of Engineering, 1990, Daniel 1984, Gorden

et al. 1989, Hamilton et al. 1979, Kraatz 1977, Stewart and

Nolan 1987). This is because the soil below the testing

apparatus cannot be completely saturated. Two relatively

simple test set-ups, single and double ring infiltrometer,

are shown in Fig. 10. A more complex setup is shown in Fig.

11 (University of Florida, Department of Geology and Civil

Engineering, 1990). In all three test arrangements the

function of the outer ring is to prevent lateral flow of

water during the tests. All suggested tests are complicated,

very sensitive, time consuming, not rugged, expensive,

require lengthy analysis, and require a highly technical

person to perform them.








Prediction of Hydraulic Conductivity in Partially
Saturated Clays

A practical science for prediction of moisture migration

in partially saturated soils has not been fully developed for

unsaturated soils for two main reasons. First, there has

been a lack of an appropriate science with a theoretical

base. Second, there has been a lack of practical technology

to render engineering practice economically viable. There is

a need for further experimental studies and case histories to

substantiate the available concepts and theories (Fredlund

1979). This summary includes a brief review of the concepts

of moisture flow in partially saturated soils, including

analysis techniques for application to geotechnical problems.



Basic concepts and definitions

Water in soil is continuously under the influence of one

or more forces that determine its energy status or potential.

There are four types of potential gradients that cause flow

of water through soil--hydraulic, electric, chemical, and

thermal. However, under most circumstances the hydraulic and

chemical gradients do exist. Hydraulic potential includes

the gravitational and matrix components. Chemical potential

is often referred to as osmotic potential. The total

potential is the sum of the component potentials, or


Total = Oh + (e + Oc + (t


(12)







where

0 stands for potential energy.

(h = hydraulic potential

Oe = electric potential

Oc = chemical potential

t = thermal potential



The potential is expressible physically in at least three

ways (Cedergren 1977, Harr 1962, Mitchell 1976):

1. Energy per unit mass. This is a fundamental

expression of potential, using units of ergs per gram or

joules per kilogram.

2. Energy per unit volume. This yields the dimensions

of a pressure (e.g., kilopascals, atmospheres, or pounds per

square inch). This expression is convenient for the osmotic

and pressure potentials.

3. Energy per unit weight (hydraulic head). This is the

height of a liquid column corresponding to the given

potential. This expression of potential is certainly

simpler, and often more convenient, than the preceding

expressions. Hence, it is common to characterize the state

of soil water potential in terms of water head in

centimeters, meters, or feet.

Consideration of the potential is important because of

its relation to the movement of water in soils. The

gravitational component of potential is due to the continuous








downward action of the Earth's gravitational field. The

higher the elevation, the greater the potential. Matric

potential is due to capillary action, which in turn depends

on the adhesion between soil and water and cohesion between

water molecules. If free water is adsorbed by soil without a

change in elevation, its potential energy is decreased, the

extent of decrease being a function of how tightly the water

is attracted to the soil. Matric heads are also referred to

as suction heads and are always negative in sign. Matric

potential varies directly with the soil water content; that

is, as the water content is increased toward saturation, the

matric potential increases toward its maximum value, which is

zero at full saturation, as shown in Fig. 12 (Hillel 1971).

Osmotic forces represent the attraction between

dissolved ions and water. The higher the concentration of

ions, the greater the osmotic forces. Like matric forces,

osmotic forces reduce the potential of water, which causes

the osmotic potential to be negative in sign.

The rate of water flow through soils depends on two

factors: (1) the driving force (potential gradient), which is

normally taken as the change in water potential per unit of

distance, and (2) the conductivity, or the ability of the

soil to transmit water. The conductivity, as used here, is

analogous to the hydraulic conductivity for saturated flow.

The coefficient is multiplied by the gradient to obtain fluid

velocity. The higher the water content, the higher the








conductivity. The water content affects the ability of the

soil to transmit water for several reasons, principally by

influencing the total cross sectional area of pores through

which water flows, the amount of friction encountered, which

is maximum where water moves in thin films close to soil

particle surfaces, and the length of flow path through the

pore. Figure 13 (Hillel 1971) shows the typical

relationships between conductivity and suction for two

partially saturated soils, sand and clay.

Infiltration is the entry of surface applied water into,

and its movement through, soil. Infiltration is normally

assumed to occur in response to the combined influence of

matric and gravitational forces. The advance of water is

along a boundary known as the wetting front. During wetting,

at least a thin saturated zone is maintained at the surface

where water first enters the soil. Since the pores in this

zone are water filled, they exhibit a maximum and constant

conductivity equal to the hydraulic conductivity. The

magnitude of saturated conductivity is very important in

determining how fast water can infiltrate and move through

the soil. Evidence of this fact is that sands, which have a

high hydraulic conductivity when saturated, have a relatively

high infiltration rate during wetting.

An important characteristic of soil wetting is that it

slows with time. There have been several reasons cited for

this decrease in velocity of the wetted front. Colloids in






19

the soil may swell and reduce the pore size, or fine material

from the surface may be washed into the soil, plugging up the

pores. The continuous sheet of water above the soil and in

the upper layer of soil makes it difficult for the air in the

soil to escape and to make room for further water to enter.

Potential gradient across the wetted front zone decreases as

the potential difference is dissipated over a widening wetted

front region (Hillel 1971).



Prediction of moisture flow in partially saturated soils

The success of a field hydraulic conductivity prediction

depends quite heavily on the prediction of the depth and

extent of the wetted zone, because water is the main factor

in saturating the soil, thus allowing the saturated hydraulic

conductivity to be measured. The prediction of moisture

movement in partially saturated soils is very complicated

because of the following potential variabilities associated

with the soil, water, and driving forces.

1. Soil type, gradation, structure, and dry unit weight.

2. Amount and type of dissolved salts.

3. Temperature changes in space and time.

4. Moisture changes in space and time.

5. Soil suction and conductivity changes with moisture

content, temperature and dissolved salts.

6. Nonlinearity of the conductivity versus soil suction


curve.








7. Hysteretic nature of the conductivity versus soil

suction relationship.

8. Difficulty in obtaining accurate measurements of soil

suction and conductivity.

9. Volume change upon inundation.

10. Sources of moisture differ in their character by way

of amount of available water, rate of supply, and location

within the soil profile.

11. Soil anisotropy and inhomogeneity.

12. Thickness of soil profile.

13. Water properties change according to temperature,

dissolved salts, and capillary attraction.

14. Soil fluids including adsorbed water, free liquid

water, water vapor, and air.

15. In situ stress conditions and mechanisms are not

easily defined.

16. Boundary conditions for analysis are related

to environmental conditions which are difficult to

predict.

Since the beginning of the twentieth century, the

problem of partially saturated flow has been studied by

physicists, soil scientists, hydrologists, petroleum

engineers, and geotechnical engineers. The following is a

brief review of some of the more known studies.

Buckingham (1907) developed the following equation as

the general fluid flow law.









Q = 1 S (13)



where

Q = the mass of water per square centimeter,

S = i = Y/Dx = gradient of capillary attraction, and

X = ki = capillary conductivity = infiltration coefficient.



He noticed that both the capillary conductivity and the soil

suction pressure change with water content. Green and Ampt

(1911) studied the motion of a wetting front through the soil

and developed the following equation:



dV/dt = A (dl/dt) n (14)



where

V = volume of liquid water,

1 = depth of water infiltration, and

n = porosity.



The combination of this equation with Poiseuille's law of

flow in capillary tubes was used to develop the Green-Ampt

wetting front motion equation.

Richards (1931) used the general equation of motion of

viscous fluid, the Navier Stokes equation:







22

dv/dt = F V (P/Pw) + (A/Pw) (V V v/3 + V V v) (15)



where

dv/dt = acceleration,

F = external or body forces = V *,

V P/Pw = force due to pressure gradient, and

(g/Pw) (V V v/3 + V V v) = expression of viscous

retarding forces.


V = del operator = a/ax + a/ay + a/az


Richard used Darcy's law (1856) to describe the fluid

flow and the continuity equation to develop the following

equation:



V q = yd (ae/at) (16)



where

V q = divergence of the flow,

Yd = dry unit weight,

O = volumetric moisture content, and

a9/at = rate of change of moisture content.



Richards then related the soil suction changes to the

moisture variations.









oW/o = Cc = capillary capacity (17)



Combining equations 17 and 18 with Darcy's law, and extending

to three dimensions, the following flow expression was

obtained:



K V2 Y + (aKx/ax) (Y/a/x) + (aKy//y) (a~/ay) +

g (aKz/az) (Di.az) =- yd A (~iy/at) (18)



where

K = hydraulic conductivity,

y = total potential, and

-Yd A h/a8t) = rate of volume change of fluid.



Philip and de Vries (1957) combined the equations of

liquid flow and vapor flow into the following equation:



ae/1t = V (DT V T) + V (Do VO) + aK/az (19)



where

DT = DTliq + DTvap = thermal moisture diffusivity and

Do = Deliq + DOvap = isothermal moisture diffusivity.



Blight (1971) suggested that Fick's law represented gas

transport better than did Darcy's law. The diffusivity in

Fick's law (D) which relates mass flux (am/at) and pressure









gradient (aP/az) is a constant. On the other hand, the

conductivity relating velocity and pressure gradient varies

with the pressure gradient. Fick's law can be stated as



m/t = D (aP/az) (20)



Philip (1969) stated Darcy's law as



v = K(8) V# (21)



where

v = vector flow velocity,

K(O) = conductivity, a function of 0,

0 = total potential = y(O) + Z,

y(0) = capillary pressure potential, a function of 0, and

e = volumetric water content.



He combined the continuity equation



a80/t = V v (22)



with Darcy's law to write the following diffusion equation:


a9/at = V (K V ) + aK/lz


(23)









Defining the diffusivity D = K (dV/a8), Philip rewrote

equation (20) as follows:

ae/at = V (D VO) + (aK/aO) (Oa/az) (24)



The diffusivity (D) is analogous to the coefficient of

consolidation Cv in the consolidation equation.

Bear (1979) separated partially saturated flow into

three ranges:

1. Pendular saturation at very low saturation levels

leads to almost no flow or pressure transfer.

2. Equilibrium water saturation or the funicular

saturation at which both the soil air and the soil water are

continuous.

3. Insular saturation, high saturation levels at which

the air phase is no longer continuous.

Bear defined the piezometric head in both the saturated

and the partially saturated zones as total potential,

including both a gravity term and a pore water pressure term,

as



= z + AV (25)



where

V = P,/'y for saturated soil,

V = Pc/yw for partially saturated soil,








z = Elevation head (potential),

Pw = Pore water pressure,

Yw = Unit weight of soil, and

P, = capillary pressure



Mitchell (1976) discussed the validity of the Kozeny-Carmen

equation for partially saturated soil (Kozeny 1927, Carmen

1956):



k = K (P/Yp) = [(Cs Vs2)/So2] [e3/(l+e)] s3 (26)



where

k = permeability,

K = hydraulic conductivity,

g = viscosity of the permeant,

yp = unit weight of the permeant,

Cs = pore shape factor,

Vs = volume of solid,

e = void ratio,

s = degree of saturation, and

So = specific surface per unit volume of particles.



Although this equation works well for the description of

conductivity in uniformly graded sands and some silts,

serious discrepancies are found in clays. The major factor

responsible for failure of the equation in clays is that








the fabrics of such materials do not contain uniform pore

sizes. Particles are grouped in clusters or aggregates that

result in large intercluster pores and small intracluster

pores.



Measurement of matric suction

Matric suction determination is useful in analyzing

fluid flow through partially saturated soils. Measurements

of suction can be made by several techniques as shown in

Table 1.

Soil suction potential is often measured as a negative

water head. The absolute value of the logarithm to base ten

of suction heads in centimeters is defined as the "pF" value,

a common expression of soil suction. One atmosphere of

suction is approximately equal to a "pF," value of 3, a suction

head equal to 103 centimeters of water. The logarithmic unit

PF is preferred because most of the soil behavior is linearly

related to suction in PF units. Qualitatively, a PF value of

about 2 corresponds to a very wet condition, 3.5 PF

corresponds to the plastic limit, and a value of

approximately 6 PF is the driest condition for soil.

The following is a summary of the most used techniques

of measuring soil suction (Mitchell 1976, McKeen 1988, Kohnke

1968):

1. Piezometers. Water in the piezometer communicates

with the soil through a porous stone or filter. Pressures








are determined from the water level in a standpipe, by a

manometer, by a pressure gauge, or by an electronic pressure

transducer. A piezometer used to measure pressures less than

atmospheric is usually termed a tensiometer. Piezometers are

often used to measure positive pore water pressures.

2. Gypsum block. The electrical resistance across a

gypsum block is measured. The water held by the gypsum block

determines the resistance, and the suction in the surrounding

soil controls the amount of moisture in the gypsum block.

The gypsum block technique is used for measurements of pore

pressures less than atmospheric (Kohnke 1968).

3. Pressure-membrane devices. An exposed soil sample is

placed in a membrane or a ceramic plate in a sealed chamber.

Air pressure in the chamber is used to push water from the

pores of the soil through the membrane. The relationship

between soil water content and applied pressure is used to

establish the relationship between soil suction and water

content. The applied pressure at a given water content is

taken as the soil suction for that same water content.

4. Consolidation tests. The consolidation stress

applied to a sample is taken as the soil water suction when

the sample is in "equilibrium" with respect to fluid flow.

If the consolidation pressure were instantaneously removed,

then a negative water pressure of the same magnitude would be

needed to prevent water movement.








5. Vapor pressure methods. The relationship between

relative humidity and water content is used to establish the

relationship between soil water content and soil suction.

The soil is allowed to come to equilibrium with an atmosphere

of known relative humidity in a sealed constant-temperature

room or container. The relative humidity may be controlled

by a solution having a concentration of 3.3% of sulfuric acid

(H2S04) in water, whose aqueous vapor pressure corresponds to

98% relative humidity, or PF 4.5. Figure 14 (McKeen 1988)

shows various scales for reporting suction values. The

disadvantages of using a dilute solution for this purpose is

that its concentration may change during the determination

because water is given off or received from the soil sample.

Therefore, the concentration of the H2SO4 has to be checked

and adjusted. More recently, saturated salt solutions have

been used for establishing more stable vapor pressure levels

for determining the relationship between soil suction and

soil water content in the dry range.

The saturated salt solutions have the advantage that

their vapor pressure remains the same as long as the

solutions are in equilibrium with the solid phase, provided

that the temperature remains constant. Change of soil water

content does not alter the vapor pressure of such a solution

as long as part of the solid phase of the salt is remaining.

Table 2 shows five examples of saturated salt solutions used








to obtain water vapor tensions at a temperature equal to 25oC

(Kohnke 1968).

The United States Geological Survey (McQueen and Miller

1968) developed a filter paper method for measurement of

suction on field gathered samples which were returned to the

laboratory for evaluation. The method employs a filter water

content versus relative humidity curve, which has been

calibrated using salt solution. The filter paper is placed

with the soil sample in a temperature controlled closed

container for at least a seven-day period for the purpose of

reaching equilibrium. The water content of the filter paper

and the soil are measured, and the suction is inferred using

the calibration curves as shown in Fig. 15 (McQeen and Miller

1968). The advantage of the filter paper method is that it

is theoretically applicable over a very wide range of suction

values.

6. Freezing-point-depression method. From saturation to

a total tension of about 2 or 3 atmospheres, the freezing

point of water changes very little. From a tension of 3 to

about 25 atmospheres, there is a pronounced change of the

freezing point. Beyond this level, there is so little water

in the soil that it becomes practically impossible to

determine its freezing point. Therefore, the best range to

determine total tension by the freezing-point-depression

method is from PF 3.5 to 4.4.








7. Centrifuge. The centrifuge can be used to determine

the amount of soil moisture retained against particular

centrifuge forces. Briggs and McLane (1907, 1910) have

developed a technique in which a wet sample of soil is

subjected to a centrifugal force 1000 times the force of

gravity for 40 minutes. The resultant water content is

called the moisture equivalent (similar to "field capacity").

In this centrifuge test, the results are only used to provide

qualitative data for comparisons of suction between various

soil types (Kohnke 1968).

8. Thermocouple psychrometer. A psychrometer is defined

as two similar thermometers with the bulb of one being kept

wet so that the loss of heat that results from evaporation

causes it to register a lower temperature than the dry

thermometer; the difference between the two temperature

readings represents a measure of the dryness of the

atmosphere and is called the wet bulb depression. From this

information, the relative humidity can be computed. For more

details and discussion, refer to McKeen (1988).



Factors Affecting the Prediction of Saturated Hydraulic
Conductivity of Clay Liners

Several investigators have addressed the influence of

various factors on the measurement of the saturated hydraulic

conductivity of compacted clays both in the laboratory and

in-situ (Acar et al. 1987, Bagchi et al. 1985, Berystorm

1985, Bogardi et al. 1989, Boynton and Daniel 1985, Carpenter









and Stephenson 1986, Daniel 1984, Elzeftawy and Cartwright

1979, Gorden et al. 1989, Korfiatis et al. 1987, Mitchell and

Younger 1966, Mitchell 1976, Oakley 1987, Olsen et al. 1979,

Peirce et al. 1987(a), Schmid 1966, Siva et al. 1979, Stewart

and Nolan 1987, Taylor 1948, Wit 1966). Therefore, the

factors affecting the prediction of saturated hydraulic

conductivity will be separated into laboratory and field

factors, and each will be briefly reviewed.



Laboratory factors

Several investigators have studied the many factors that

affect the measurement of the saturated hydraulic

conductivity of compacted clays in the laboratory. Broadly

speaking, the factors influencing hydraulic conductivity can

be classified into three categories.

1. Testing apparatus factors. These factors are

associated with testing variables such as type of

permeameter, confining pressure, direction of flow, and

hydraulic gradient. The three most common types of

permeameters are the consolidation cell, rigid wall, and

flexible wall. These permeameters were discussed previously.

a. Type of permeameter. Boynton and Daniel (1985)

have outlined qualitatively the difference in some parameters

when using the three type of permeameters. This outline is

shown in Table 3. Figure 16 (Boynton and Daniel 1985) shows

the results of two types of clays tested using the three








different permeameters. Based on these results, it is

concluded that the type of permeameter did not have a large

effect on the measured hydraulic conductivity; the

differences in the values of conductivity were substantially

less than one order of magnitude; and no one type of

permeameter consistently yielded higher or lower values than

the other types. However, Stewart and Nolan (1987) have

found that the conductivity measured from the rigid wall

permeameter is consistently lower than the other types as it

is shown in Fig. 17 (Stewart 1987).

b. Confining pressure. This factor affects the

hydraulic conductivity measured by the flexible wall

permeameter only since the other types do not apply an all-

around confining pressure to the tested sample. This is done

in order to prevent side wall leakage and facilitate sample

saturation. Figure 18 (Boynton and Daniel 1985) shows that

as the confining pressure increases, the conductivity

decreases. Korfiatis et al. (1987) have shown that the

conductivity value decreased twice as much as that reported

by Boynton and Daniel for the same increase in confining

pressure.

c. Direction of flow. In all laboratory

conductivity tests the flow is restricted to the vertical

direction. This is because it is easier and better simulates

the flow in the field. Also for compacted soils, the lateral

flow is the same as the vertical flow.








d. Hydraulic gradient. Mitchell and Younger (1966)

have shown that for clays, tested in flexible wall

permeameter, at low hydraulic gradient, the hydraulic

conductivity tends to be very low and the flow deviates from

equation 1. They found that this phenomenon exists due to

dislodging and washing down of fine particles in samples with

low initial compaction density. Mitchell and Younger also

showed that samples tested under increasing hydraulic

gradient have lower hydraulic conductivity than a decreasing

one. Olsen and Daniel (1979) has reported some studies which

showed that as hydraulic gradient increases so did the

predicted conductivity by 5 to 84 times.

2. Permeant factors. These factors are associated with

the type and properties of the permeant. When hydraulic

conductivity is mentioned, it is understood that water

conductivity is referred to. There are two main water

properties that can affect the speed of water flow through

soils.

a. Viscosity and density. The relationship between

viscosity and density of water with the conductivity is given

in the well-known Kozeny-Carman equation 26, and it can be

rewritten as


K = k (Yp/)


(27)








where

K = Hydraylic conductivities,

k = Permeability,

yp = Unit weight of permeant (water), and

g = Viscosity of permeant.



Equation 27 suggests that the conductivity varies directly

with the density and inversely with the viscosity of

percolating water (or any other fluid). The density and the

viscosity terms are usually taken as constant and equal to

one for water at laboratory temperature.

b. Normal and deaired water. Hydraulic

conductivity was thought to be less when using normal (tap)

water because a greater number of flow channels could become

blocked by evolved air bubbles than when using deaired water.

The opposite was found (Stewart and Nolan 1987).

3. Soil factors. These factors associated with physical

and chemical characteristics of the soil. Furthermore, these

factors affect the measured conductivity differently for

different soils. Soil properties by far have the largest

influence on the predicted conductivity.

a. Molding water content and degree of saturation.

Darcy's law and other relations for predicting the

conductivity have been developed or experimentally

established on soils with 100% saturation. Conductivity is

greatly affected if air, even in small amounts, remains in








the pores of soil. Conductivity drops to very low values at

degree of saturation less than 75% (Sing 1967). Figure 19

(Mitchell 1976) shows that as the degree of saturation

increases so does the conductivity for compacted clays tested

in flexible wall permeameter. Most of the time, it is easier

to obtain and more accurate to relate the conductivity to the

molding water content instead of degree of saturation. Both

Fig. 16 and Fig. 17 show a plot of conductivity versus

molding water content, and it can be seen that as the molding

water content increases, the conductivity decreases up to a

maximum (optimum) value. Beyond this optimum value a further

increase in the molding water content will result in an

increase in the conductivity. This can be explained by the

fact that at lower molding water content (or lower degree of

saturation) the water flows through the soil under both the

hydraulic head and suction head. As the soil becomes

saturated, most of the air will be driven out of the soil,

the suction head will be minimal, the water will flow under

the hydraulic head only, and will result in the lowest

conductivity value. Beyond the lowest conductivity an

increase in water content will result in a change of soil

fabric from a semidispersed to a fully dispersed structure

which possess higher conductivity.

b. Dry unit weight of soil. The relationship

between the conductivity and the dry unit weight of soil is

shown in Fig. 16. At low molding water content and dry unit








weight the fabric structure of the soil is mainly flocculated

(possesses a high degree of porosity or void ratio), and the

conductivity is highest. As the molding water content

increases, the dry unit weight increases, the degree of

porosity or void ratio decreases, the soil structure changes

gradually from fully flocculated to semiflocculated, and this

results in a decrease in the conductivity value. At and

around the optimum molding water content, the dry unit weight

is maximum, the soil structure tends to be semidispersed, and

the conductivity is lowest. At a molding water content

greater than this region, additional water tends to force

soil particles apart, changing the soil structure to near

fully dispersive. This will lead to a high degree of

porosity or void ratio and, therefore, a lower dry unit

weight and higher conductivity value.

c. Sample diameter. Boynton and Daniel (1985) have

studied the effect of sample diameter of fire clays tested in

flexible wall permeameter and obtained the plot shown in Fig.

20. He concluded that the measured conductivity was

essentially independent of sample diameter and that the

conductivity of the largest sample used was one third to

twice the value measured on the smallest samples. However,

the larger the sample diameter, the more likely the sample

will contain more hydraulic defects and the closer the sample

will be in resembling the field conditions.









d. Adsorbed water. The adsorbed water surrounding

the fine-grained soil particles is not free to move, and,

hence, it causes an obstruction to the flow of free water by

reducing the effective pore space available for the passage

of water. It is difficult to define the pore space occupied

by adsorbed water in a soil. According to a crude

approximation after Casagrande, 0.1 may be taken as the

voids ratio occupied by adsorbed water, and the conductivity

may roughly be assumed to be proportional to the square of

the net void ratio of (e 0.1)2. Adsorbed water has a marked

influence on the conductivity of clays. In a laboratory, it

is normal to use a high gradient for testing clays, but in

actual field problems, the hydraulic gradient is much less.

There is a hydraulic gradient (threshold gradient) for clays

at which the conductivity is essentially zero. Lambe and

Whitman (1979) reported that this gradient for some clays is

equal to 20 to 30. Mitchell (1976) suggested that the value

of threshold gradient could be higher for montmorillonite

clays and reported a maximum value of 900.

e. Mini-aging. Figure 19 (Mitchell 1976) shows the

conductivity of clay samples aged for 21 days and tested in

flexible permeameter. Aged samples did not display

consistently higher or lower conductivity than the unaged

samples. The same conclusion is reached by Boynton and

Daniel (1985) after testing different clays in exactly the

same way as it is shown in Fig. 21. Olson and Daniel (1979)








have suggested that prolonged conductivity tests (and

probably aging) may result in a substantial reduction in

conductivity due to clogging of the flow channels by organic

matter that grows in the soil during the test (and may be

during aging too).

f. Direction of flow. Lambe and Whitman (1979)

suggested that compacted clays are flocculated dry of

optimum, resulting in a lower degree of hydraulic anisotropy,

and dispersed wet of optimum, resulting in higher degree of

hydraulic anisotropy. Olsen and Daniel (1979) suggested that

clods of clay are hard when the molding water content is dry

of optimum, resulting in large interclod void space, and soft

when they are wet of optimum, resulting in minimal interclod

void space. In this case, the only source of anisotropy

would be the flattening of clods during compaction. Boynton

and Daniel (1985) have used flexible wall permeameter to test

compacted clays that were sampled in horizontal and vertical

directions. He concluded that soil fabric has no discernable

effect on hydraulic anisotropy.

g. Desiccation. Literally no data were found on

desiccation cracking in compacted clays and its influence on

hydraulic conductivity. Boynton and Daniel (1985) prepared

2.5-inch thick compacted clay slabs and found a 1 millimeter

wide crack appeared after 4 hours, and the crack penetrated

the slab after 8 hours. The cracked clays were then sampled

and tested in a flexible wall permeameter under different








confining pressures. The result is shown in Fig. 18. It was

concluded that desiccation cracks can penetrate compacted

clay to a depth of several inches in just a few hours.

Furthermore, the cracks tend to close when moistened and the

hydraulic conductivity is not affected by a large amount.

h. Sample height. Sample height was studied by

Korfiatis et al. (1987). In this study, he tested compacted

clays in a flexible wall permeameter and followed an orthodox

procedure. He tested a sample 3 inches thick and 2.5 inches

in diameter. Then, the same sample was divided into two

halves and tested, and the same two halves were divided into

four equal pieces and also tested. He concluded that the

hydraulic conductivity increases with increasing sample

height as shown in Fig. 22.

i. Amount and type of clays. Little data exist on

the effect of the amount and type of clays on the measured

hydraulic conductivity. Mitchell (1976) tested compacted

clays in flexible wall permeameter and found that increasing

amounts of clay from 5% to 15% led to a decrease in

conductivity by four times. Daniel (1987) has tested

compacted clays with different plasticity indices in a

flexible wall permeameter and found generally that as the

plasticity index increases, the measured hydraulic

conductivity decreases. This is shown in Fig. 23.

j. Termination criteria. This factor deals with

the amount of outflow of water from the tested sample








necessary to assume that a steady state value of hydraulic

conductivity has been reached. This amount is usually

expressed in terms of the total volume of pores. Peirce and

Witter (1986) has studied this factor on compacted clays in a

flexible wall permeameter and concluded that about one-half

of the pore volume is necessary to reach a steady state

conductivity. This is shown in Fig. 24.



Field factors

The factors that affect the field measurement of

hydraulic conductivity of clays are many and are very

difficult to quantify and measure, each of which tends to

have large influence on the measured conductivity. Olsen and

Daniel (1979) stated, "Field testing for measurement of

conductivity in unsaturated soils is at such a elementary

stage of development that field measurements cannot be

recommended" (p. 55). Field conductivity testing is still at

a rudimentary stage and still not performed even in large

landfill projects. Some of the suggested methods for field

infiltration tests are shown in Figs. 10 and 11. However, in

addition to the permeant and soil factors mentioned above,

there are other factors to be considered. These are

1. Homogeneity and isotropy. These factors affect the

field conductivity more than the laboratory conductivity.

This is because in the field the volume, thickness,

placement, and compaction of the clays are much greater and








different than those in the laboratory. Due to the

relatively large volume of soils handled in the field, soils

might have different amounts and types of clays even if the

supply source is the same. This will result in different

in-situ densities upon compaction, and different areas might

experience different amounts and types of compaction. This

will lead to inhomogeneity and anisotropy of the compacted

clays.

2. Discontinuities. Field discontinuities in the

compacted clays exist as desiccation cracks due to exposure

to temperature, areas of low densities due to low compaction,

pockets of high sand content and low clay content, zones of

contaminated clays with the in-situ sandy soils, and areas

with large interclod void space. All these discontinuities

will result in an increase of hydraulic conductivity of the

compacted clays.

3. Suction and saturation. In the field, both the

compacted clays and the sandy subgrade below it are partially

saturated soils and, consequently, both possess a certain

amount of suction. This suction is very difficult to

measure, will increase the hydraulic gradient, and leads to

an increase in the infiltration of water through the clays.

This is shown in Fig. 25. Many investigators have measured

the suction of the clays as a function of the clay moisture

content (Daniel et al. 1979, Elzeftawy and Cartwright 1979,

Hamilton et al. 1979, Gorden et al. 1989, McKeen 1988, Olsen








and Daniel 1979, Pachepsky and Scherbakov 1984). Figure 26

(Daniel et al. 1979) represents the typical result of such an

investigation, and it shows that as the moisture content

increases from 7% to 20%, the suction decreases from 43 to

1.5 atmospheres (632.1 to 22.1 psi), respectively. Stewart

and Nolan (1987) showed that the distribution of soil

saturation after performing the field infiltration tests is

not uniform as can be seen in Fig. 27. The figure also shows

that the moisture migrated laterally in all the tests by a

considerable amount. Stewart also measured the field

hydraulic conductivity with time and found it to vary by one-

half to one order of magnitude, as can be seen in Fig. 28.

4. Clay thickness. The thickness of the clay liner in

the field ranges from 8 inches (top cover) up to 5 feet,

while the thickness of the clay sample tested in the

laboratory is no greater than 3 inches. The only available

data on this factor are shown in Fig. 22 (Korfiatis et al.

1987). This research was performed on a compacted clay

sample with a thickness of 3 inches and, therefore, cannot be

compared to the field thickness.



Purpose and Scope of this Research Project

The purpose of this research project is to develop a new

and rugged methodology of predicting field hydraulic

conductivity for compacted natural Floridian clays and to

study a number of field and laboratory factors that are









affecting the prediction of hydraulic conductivity. Some of

these factors were expressed by the local industry in

Florida, and the others were deduced based on the

deficiencies observed from the review of previous work

related to hydraulic conductivity. Furthermore, this study

was designed such that the predicted hydraulic conductivity

values are rugged and insensitive, to some degree, to

possible mathematical manipulations. The scope of work for

this research involved the following:



Bulk Sampling. Properties. and Sample Preparation

A number of bulk soil samples were obtained from the Mid-

Florida Mining Corporation's (MFM) clay mine in Ocala,

Florida. This clay was, and still is, used in the

construction of a number of landfills. It is marketed under

the tradename of "Terra-Seal Natural Premix@." One

homogenous soil sample was obtained from these samples, and

all subsequent laboratory tests were performed using this

homogeneous sample. A number of laboratory tests were

performed to obtain the index and physical properties of the

soil. These properties were compared with those established

previously by a local professional testing laboratory.

Samples for laboratory conductivity tests were prepared using

a 4-inch inside diameter, with variable length, cast acrylic

plastic tubing, and in accordance with ASTM D698A and D1557A

(ASTM 1989). Undisturbed field samples were obtained using








three different steel sleeve sampling apparatuses designed

using some of Hvorslev's (1962) recommendations. These

apparatuses were also used to perform field infiltration

tests. Undisturbed field samples were also obtained using

block sampling techniques.



Laboratory Work

A large number of compacted soil samples were tested in

a rigid wall permeameter, and a number of relationships and

the influence of various factors on the soil hydraulic

conductivity were established. The most important of these

relationships is the soil conductivity versus soil suction,

versus dry unit weight, versus molding water content. Other

factors studied are the effect of sample height, number of

layers in the sample, hydraulic gradient, time, drying time,

and field sampling. The distribution of moisture content

versus depth of a number of samples after conductivity and

after drying were also established.



Field Work

Field work was performed at two landfill projects

located in Florida. The clay used in the construction of

these two projects is from the same source and is the same

Terra-Seal Natural Premix used in this study. Two field

infiltration tests were performed on the top cover at the

Southwest Alachua Landfill located in Archer, Florida. This









landfill was constructed during 1986. Three 10- by 9-foot by

9-inch-thick test strips were constructed, using three

different layerings. These test strips were constructed

prior to the construction of the second project, Astatula

Ash-Residu Monofill landfill located in Astatula (40 miles

south of Ocala), Florida. These test strips were used to

study the method of construction, desiccation cracks, density

and moisture content distribution, and to perform five field

infiltration tests. Three additional infiltration tests were

performed on the actual landfill after it was constructed.



Prediction and Comparison of Hydraulic Conductivity

The results of the suction tests and the field

infiltration test results were used to predict the saturated

hydraulic conductivity of the field compacted clays. The

predicted values agreed very closely with those obtained in

the laboratory by the author and two independent professional

testing laboratories. The relationship between laboratory

conductivity and the various factors studied were obtained

and quantified.








Methods of Measuring Suction (McKeen 1988)


Technique Range (pF) Remarks


Suction Plate

Pressure Plate

Pressure Membrane

Osmotic Cell

Centrifuge

Vacuum Desiccator

Sorption Balance

Thermocouple
Psychrometer

Filter Paper Method

Heat Dissipation
in a Ceramic


1.0-3.0

1.0-3.0

0.0-6.2

2.0-4.1

3.7-4.1

5.0-7.0

5.0-7.0


2.5-4.8

0.1-7.0


0.0-4.2


Matric

Matric

Matric

Total

Matric

Total

Total


Total

Total


Matric


TABLE 1.








TABLE 2.


Saturated Salt Solution Versus Relative Humidity
(Kohnke 1968)


Salt Relative Humidity pF
at 25oC


CaSO4 97.8 4.49

NH4H2P04 93.0 5.00

NH4Cl 79.3 5.51

Mg (NO3) 2 52.0 5.96

KC2H302 19.9 6.36










TABLE 3.


Various Parameters for Three Permeameters (Boynton and
Daniel 1985)


Test Type of Permeameter
parameter Compaction mold Consolidation cell Flexible wall
(1) (2) (3) (4)


Side-wall
leakage


Void ratio (e)






Degree of
saturation





Voids formed
during trimming







Portion of sample
tested


Leakage is
possible


Relatively high e
because applied
vertical stress
is zero


Specimen may be
unsaturated





Impossible; soil
is tested in the
compaction mold
and is not
trimmed




All of the
compacted
specimen is
tested, including
the relatively
dense lower
portion and the
relatively loose
upper portion;
the dense lower
portion may lead
to measurement of
relatively low k


Applied vertical
stress makes
leakage unlikely

Relatively low e
because a
vertical stress
is applied


Specimen may be
unsaturated





Voids may have
formed, but
application of a
vertical stress
should help in
closing any voids


Only the central
portion of the
specimen is
tested; the upper
and lower third
of the specimen
are trimmed away


Leakage is
unlikely


Relatively low e
because an all-
round confining
pressure is
applied

Application of
back-pressure is
likely to cause
essentially full
saturation

Voids are not
relevant; the
flexible membrane
tracks the
irregular surface
of the soil
specimen

One centimeter of
soil is trimmed
off both ends of
the compacted
sample









































. I *.*.'. I .
... C .of .. .. .. ...

Noturoa Liner
~c) \\ \ \ ~ .,'..,`


Fig. 1. Examples of Natural Liners (Daniel 1987).



























Compacted


Sidewall
Liner
(Horizontal Compacted
Lifts) Bottom Liner



Leachate Collection Zone
Primary Liner.
Leak Detection Zone
Secondary Liner


Compacted
Sidewall Liner
(Lifts Parallel
to Slope)


Types of Compacted Liners (Daniel 1987).


Fig. 2.















































DRAINAGE BLANKET

LEACHATE COLLECTION PIPE

COMPACTED CLAY LINE


Fig. 3.


Typical Landfill Section and Components
(Oakely 1987).

























































Fig. 4.


Hydrological Cycle as Applied to Landfill System
(Oakely 1987).



























III. Turbulent


Velocity, v


Zones of Laminar and Turbulent Flow (Taylor 1948).


I. Laminar


II. Transition


Fig. 5.
































vent
\


load cap porous
stone






:-*-:::---: \ri
0 001 001, 00, l -
;~..,A 16S~~;2~ 9~E^P
'( ^^^ ^ % %^ ^ ^


Fig. 6. One-Dimensional Schematic of Consolidation Cell
Permeameter.


nple
ng


----- -- ------


inflow










vent
port

(9


top
plate


acrylic
tube






porous
stone



bottom -
cap


bottom
plate


fill and
drain port


drains


Fig. 7. Schematic of Flexible Wall Permeameter.


















top --
plate


clamping -
rod





bottom --
plate


pporous
stone



rigid
wall

p porous
stone


port


Schematic of Rigid Wall Permeameter.


Fig. 8.







MARIOTTE TUBE





| O
0
0
0
0
0
0
















0

0
0,,



05\
)r.l
\\C\
mI\\





O
% %%j


Schematic of Mariotte Tube.


Fig. 9.







































Open Single-Ring Infiltrometer


<2r 1:Q :~c V'. I.% ':"' 'Q'
Ac'%rX.; U
4fi;A-+ ___'' r


t
-..r.'~


Pk~ ~3ia;7'C'
V Y


.............. ....... .. ......... .'.....*...'...............

















Fig. 10. Schematic of Single and Double Ring Infiltrometers.


i
...; : "
I 't~"?i' '
r~ ~\


























Inner ring flush
tensiometer port

CIa \n \o


grout


Inlet
port
tubing

flexible bag


outer ring


--------------r rr rr rr rr r- rr
----l((-------- 1-~~
tes pad ...... .................
........oe ............................ 0 ............................. 0 ....................................
.. 0eo o oo o o ............ ........... ........... ........... ........... o 0 ........................o ..
.. .. 0o .. .. .. 00 .. .. .. .. 0 .......................... ............... ........
..oo ... .. ..o o ..o .. eo... .... ...... ..... .... o..o o 0..... ...... .. ... .....


Fig. 11.


Schematic of a Sealed-Double Ring Infiltrometer
(University of Texas, College of Engineering,
1990).








































Water content


Fig. 12. Soil Suction versus Water Content (Hillel 1971).































Ks'






Ks2











C
a
U
0
0



0
h>
*o
>*


Suction


Soil Suction versus Conductivity (Hillel 1971).


Fig. 13.











Ib/in2


SBars
01',"-


Scales for Reporting Suction Values (McKeen 1988).


g/cm2
cm H 0
10'-=





106--


Pascals
x 10'
or kPa
10'-3


pF
7-





6





5





4





3





2












0-
0


tons
ft2


kg/cm2
10'-=












100





10-





-1.0 :





0.1





0.01-






0.001-


10"-:






10'-=





100-


10-
-












0.1-





0.01--





0.001


10--.
1 0.






-1000]-





100-





10-





1-


10'--





10'









1









1.0





0.1


-1.


0.001


Fig. 14.













































FILTER PAPER WATER CONTENT (wf)


Fig. 15.


Filter Paper Calibration Curves
1968).


(McQeen and Miller






































: I C M
o Cm wseleimM Ce
S* Pa tte w.l Cell


25 30 35 40
Water Content m)


Molding Water Content (


Fig. 16. Conductivity vs. Dry Unit Weight vs. Molding Water
Content for Two Different Clays (Boynton and Daniel
1985).


I 10


72- -i
15 20
Molding


a-


i ZlO-8
I dO





























-0
10














-7
10







10-


MOLDING WATER CONTENT, %

NOTE: SOLID SYMBOL INDICATES UNDISTURBED SAMPLE


Fig. -17.


Summary of Laboratory and Field Infiltration Tests
(Stewart and Nolan 1987).


0 RIGID WALL


* FLEX WALL

V OEDOMETER

A INFILTR

4


0


~--J---

















(kPa)
50


-8


10


Effective


4 8 12


Confining


Pressure


100


16

(psi)


Fig. 18. Conductivity vs. Confining Pressure
Daniel 1985).


(Boynton and
















100-


90


80


70


E 60
!













20






0-
0.5










Fig. 19.


0.6 0.7 0.8 0.9
(Degree of Saturation)3


Conductivity vs. Degree of Saturation vs. Aging
(Mitchell 1976).





















-9
4 x10


3 x109


2 x10-9
2xlO0


E
0






C
0

=

0



z
"0
13

Xrs


0 I 2


3 4 5 6 7


Sample Diameter (In)


Fig. 20.


Conductivity vs. Sample Diameter
Daniel 1985).


(Boynton and


I x 10 9
























-9
4x10



3xi0"9


-9
2x10


-9
Ix 10


0


I I__I I 111 I I111


*


U
O

E








C
U




u

a


w
X


I0


A I I I I I I


100 300


Storage


Time


(days)


Aging (Boynton and Daniel 1985).


I I 111LI


I I I "I I


' "I


Fig. 21. Conductivity vs.





















-6




1 5X10-
S ---- _

-7




11
S -8- --


5X10
> ~--- -4 ----
/







oTable 1
ATable 2 (A)

*Table 2 (B)
1X10o 8 i
01234 5 6 78 910
SAMPLE HEIGHT (cm)


Fig. 22. Conductivity vs. Sample Height (Korfiatis et al.
1987).




















-6
- 10 6---------
. 10
oU


I 10-r -a
;, Upper Bound



U

io-9
10

= I l I I I I
S-10
10 20 30 40 50 60

Plasticity Index ()


Conductivity vs. Plasticity Index (Daniel 1987).


Fig. 23.











































000
2-


n.


r I

Beg In
Chemical
Test


SPore Volumes


Fig. 24. Conductivity vs. Pore Volume (Peirce and Witter
1986).


Standard
Waer






























.-. ,*', t ^ ro '., ,
*-.... .,. S? .^?, ,-; ...
'. ., i.< i.. ., "', ^ .
'* ***^ *^ ; '.'


. ; _.Nl~f' 'r~.
. YV ; Sii
:. ~x
.:n. rr~
:


.1 6 0 1 F-
ff ff f7 f ... ..*.. ..' .* .
...,..............
:%%%% % %% %%~ ~~%% %% % % % %%%% %%
\ % \\\\% % % \ % % % % *% %% %^ -. %. %' ..% %% % %% .
\~*\s*\^A/^l^ ^^''' \\c\ ;',\',\ \\\^,\''l'' \ C~\*\\ *\ \ \\ '4',\ '\
t#IItP*lIf
*^ \\',\\\\\\\\\\\\\\*\**\UI\~\\~\\\\^^\ \\^\\\\\\\\\\\



1 l J. 4% p 4IlI I d ld*FC l 10- III l ll II~

''%. ..% % % % % % % % % % %% %% ~% %. %%% % % % % %% % % % % %%
\'^\\\\^,\\ \~\\\\\\\\~\\\\\\~\\\\\\?^\\',\\\\\\^\\
'\\5\\\^^^\%^^,\**\\\\\\ \\\ \\'>>^ e'^V/^^^^^^^^^^V^^^^%A


% % %. .% % .% % % % % % % % % % X % % % % % % % % %% % % %


. .V 0 Il 0 h 0 0 0 0 11 IhI 11111j
* ( / / I) C / 1 ) I C ( / 1 I ^ ^ r C ( ( ( ( ^ ^ Z ^ U ^ r I I 1 I I I I I V I I I I I 1 I1
% ^ % % % % % % % % % % % % % % % % ^ %
. .
% % % % % % % % %% % % % % %% % % % % %% %


i


H+D+Hs


Hs=?


Fig. 25. Schematic of Single Ring Infiltrometer and Suction
Head.









































0 5 10 15 20 25
COMPACTION WATER CONTENT (5)


Fig. 26.


Suction vs. Water Content (Daniel et al. 1979).
















0 93-100 [] < 93 BENTONITE PUTTY


16




2.5




13




20




30


2.3



K X 10' CM/SEC


B4



B5




B6


SCALE
0 100 MM
0 100 MM


Fig. 27. Distribution of Soil Saturation after Field
Infiltration Tests (Stewart and Nolan 1987).


1l00



























AA,

s006 o38OBM'& &0
& &

E. .A

a.# *U@^^ **
4 a I g


jL~i "-es


ELAPSED TIME, DAYS


Fig. 28.


Field Conductivity vs. Time (Stewart and Nolan
1987).


ttLw'


10-r


10-1


ww












CHAPTER 2
BULK SAMPLING, PROPERTIES, AND SAMPLE PREPARATION


Bulk Sampling

A total of five bulk samples of the Terra-Seal Natural

Premix clays were obtained from different parts of an

existing stockpile at MFM surface clay mine (Lowell mine) in

Ocala, Florida. The stockpile was very large and was made by

excavating the natural clay deposits, mixing it, and then

stockpiling it. This operation was performed in order to

disrupt any existing soil stratification. The stockpile soil

consisted of firm to stiff yellowish and reddish brown

mottled light gray and green silty clay with a trace of fine

to medium sand and a trace of fine to coarse gravel-sized

limestone nodules. The stockpile also contains small to

medium boulder-sized clay lumps. Approximately 1000 pounds

of the clay was obtained and brought back to the University

of Florida laboratory in Gainesville. In the laboratory, all

of the sampled clay was placed in a large tray, mixed

thoroughly, and every effort was made to insure that the

nominal size of all clay lumps was not larger than 2 inches.

This operation was necessary to obtain an average and

homogeneous clay sample. This sample was then placed in a








tightly sealed large container, stored in a controlled

environment, and was used as the project clay in all

subsequent laboratory tests.



Properties of the Project Clay

The properties of the project clay consisted of index

and physical, mineral, and chemical properties. Some of

these properties were measured directly and some were

collected from previous work performed on the same clay.



Index and Physical Properties

Index and physical properties of the clay were obtained

in the laboratory by the author. These properties consisted

of natural moisture content (as received moisture content),

percent passing the No. 200 sieve (percent fine which

represent silt and clay), Atterberg Limits, and specific

gravity of the solid particles. All of these tests were

performed in accordance with ASTM (1989) standard methods of

testing. A summary of the results of these tests is shown in

Table 4. Detailed results of these tests are shown in

Appendix A. Table 4 also shows a summary of results of

similar tests performed by two different professional testing

laboratories on the same clay used at two different projects.

Details of these test results are shown in Appendices C and D

for South West Alachua and Astatula landfill projects,

respectively. The table shows that the clay properties








obtained by the author are very close to the others. This

suggests that the project clay can be safely assumed to be

the same as that used in the two projects.



Mineral and Chemical Properties

Type and amount of minerals present in the clay were

studied by Dr. James Eades and Dr. E. C. Pirkle of the

Department of Geology at the University of Florida during

1988. These properties were established by a combination of

hydrometer and X-ray analysis performed on a number of clay

samples. They found that the clay contains 19% to 78% of

fine sand, 2% to 18% silt, and 13% to 73% clay. Furthermore,

the clay was found to be mainly montmorillonite with a trace

of sepiolite, attapulgite, illite-waverlite, and kaolinite-

weathered. Details of the mineralogical studies are included

in Appendix B.

Chemical analysis of the clays was performed by Post

Buckly, Schuh, and Jernigan, Inc. They perform the analysis

on clay samples obtained from that used in the construction

of the Astatula landfill project. Total metal tests and

Toxicity Characteristics Leaching Procedure were performed on

the sampled clays. As part of the total metal testing

procedures, the clays were tested for arsenic, barium,

cadmium, chromium, lead, mercury, selenium, silver, and

sodium. It was concluded, based on these tests, that the

clays meet the EPA (Environmental Protection Agency) and FDER








(Federal Department of Environmental Regulations) standards

and that the clays are not hazardous to the existing

surficial aquifer water quality. Detailed results of the

chemical analysis are included in Appendix B.



Sample Preparation

Different sets of procedures were followed in the

preparation of clay samples for laboratory and field tests.

However, in each case, the preparation procedures were

conducted using the recommendations suggested by well-known

documented standard and nonstandard procedures.



Laboratory Samples

All samples prepared in the laboratory for conductivity

tests were in accordance with ASTM (1989) standard

procedures. Two particular ASTM test procedures were

followed in the preparation of compacted clay samples for

hydraulic conductivity tests. These were ASTM D-698-method A

(standard or Proctor method of compaction for fine grained

soils) and ASTM D-1557-method A (modified method of

compaction for fine grained soils). Method D698A specifies

that the clay should be placed in a standard mold in three

equal layers, each layer subjected to 25 blows of a 5.5 pound

rammer falling from 12 inches above the surface of the clay.

While method D15 57A specifies that the clay is to be placed

in the same standard mold in five equal layers, each layer








compacted by 25 blows of a 10-pound rammer falling from 18

inches above the surface of the clay. This means that

resulting samples are subjected to higher compaction energy

and, therefore, possess higher unit weight for the same

molding water content than those prepared by method D698A.

By the measurement of sample volume, wet weight, and moisture

content, the dry unit weight and the molding water content of

each sample was obtained. Detailed procedures of these

methods can be found in the ASTM (1989) handbook.

The mold for the laboratory samples was made of cast

acrylic plastic tubing with inside diameter of 4 inches,

outside diameter of 4.5 inches, and with variable lengths.

The inside diameter of the tubing is the same as that for the

standard mold.



Samples used for suction measurements

Samples for the suction tests were prepared by following

ASTM procedures in which the clays were air dried, passed

through the No. 4 sieve, mixed with appropriate amount of tap

water, cured for 48 hours, and then two samples with the same

moisture content were compacted in accordance with D698A and

D1557A. A rubber ring was placed on the surface of the

prepared sample, a Fisher Scientific standard filter paper

No. 09-790A was placed on top of the ring, and the top of

sample was air-tight sealed for at least 10 days. The type

of filter paper was the same as that used by McKeen (1988).








After the test termination, the moist filter paper was placed

in a preweight sealable plastic bag and its weight recorded.

Then, the moist filter paper was placed in a 1100C constant

temperature oven for 24 hours and then in a fresh preweight

sealable plastic and weight.



Samples used to study the effect of desiccation

Samples for desiccation study tests were prepared as in

those for suction tests except the dried clay was mixed with

about 24% moisture content (wet). This is because compacted

wet soil dries more and, hence, desiccates more than

compacted dry soil. Two identical 18-inch-thick compacted

samples were prepared in 12 equal layers compacted in

accordance with D698A. Six thermocouples were placed in one

of the samples at 1.27, 3.81, 7.62, 15.24, 26.67, and 41.91

cm from the top. This was to monitor the temperature profile

with time. The two samples were placed in an ultraviolet

chamber with a constant temperature of 38oC. Daily readings

of the temperatures of the six thermocouples were taken for

16 days. At the end of this period, moisture content profile

tests were performed on the sample with thermocouples. In

addition, conductivity tests were performed on the other

sample. Moisture content profile tests were performed after

the completion of the conductivity tests. Figure 29 shows a

cross section of the adopted laboratory hydraulic

conductivity setup.








Samples used to study the effect of soil thickness

Samples for the soil thickness study were prepared using

the average homogeneous Terra-Seal Natural Premix soil that

was discussed in the Bulk Sampling section. Four samples

with thicknesses of 1.5, 4.5, 12, and 18 inches were prepared

in accordance with D698A. All samples were placed in 1.5-

inch layers, applying the same amount of compaction energy

per layer. Then conductivity tests under a constant

hydraulic gradient of 70 were performed on each sample. When

the hydraulic conductivity reached a stabilized value, within

5% to 10% of the previous reading, the value was recorded and

the test was terminated.



Samples used to study the effect of number of layers

Samples used to study the effect of number of layers on

the predicted conductivity were prepared using the average

homogeneous clays described in Bulk Sampling section. Samples

with a total thickness of 1.5 and 4.6 inches were prepared in

one and three layers; samples with total thickness of 12

inches were prepared in two, four, and eight layers. The

total applied compaction energy per unit volume was the same

for all samples and for ASTM D698A. Then, conductivity tests

under a constant hydraulic gradient of 70 were performed on

each sample. The hydraulic conductivity value was recorded

when it reached within 5% to 10% of the previous reading.