CHARACTERIZATION OF HYDRAULIC PROPERTIES OF BATCH
ORGANOMODIFIED INITIALLY DISPERSED AQUIFER TYPE SOILS FOR
SORPTIVE ZONE APPLICATIONS
By

Gholamreza Rakhshandehroo

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department Of Civil and Environmental Engineering

1997

 

 

 

21;

CC

((

ABSTRACT
CHARACTERIZATION OF HYDRAULIC PROPERTIES OF BATCH
ORGANOMODIFIBD INITIALLY DISPERSED AQUIFER TYPE SOILS FOR
SORPTIVE ZONE APPLICATIONS
By

Gholamreza Rakhshandehroo

The hydraulic feasibility of organomodified soils for potential use in sorptive zone
applications was studied by measuring hydraulic conductivity and porosity of treated and
untreated soils of l) a natural sandy loam and 2) a reconstituted soil with different clay
contents as a function of efiective stress. The reconstituted soil samples with 6, 12, 18,
and 24% fine material were employed. Organomodification of all soil samples was
conducted in a batch process with hexadecyltrimethyl ammonium (HDTMA) to different
levels of their cation exchange capacities. Treated and untreated samples of the natural
soil were permeated with water and a pure organic liquid (Soltrol). Those of the
reconstituted soil were only permeated with water. Upon initial packing all untreated soils
(natural and reconstituted) were more porous than treated ones. In natural soils permeated
with water, untreated soil retained its higher porosity at early loads and equaled the
treated one at higher loads. Its conductivity was higher than the treated soil at early loads
and lower at high loads. In the natural soils permeated with Soltrol, untreated soil
retained its higher porosity and exhibited a higher conductivity, compared to treated soil,
throughout the loading sequence. In reconstituted soils, with application of the first load

(0.25 tst) all untreated soils became equal or less porous than the treated soils and

 

Ct

 

01

Cl

Ir

exhibited lower conductivities. Higher loads decreased the difference in porosity between
heated and unheated soils while increasing the difference between their conductivities.
At the highest load (8 tst), heated soils were 5 to 129 times more conductive than
unheated ones depending on their clay contents and heahnent level. Observed differences
in conductivities were explained in terms of the role of heated and unheated clays in
conholling initial efiective pore size and its change during consolidation. Creation of
organomodified sorptive zones by batch process is Shown to be hydraulically feasible as
evidenced by similar or higher conductivities of heated soils compared to unheated ones.
In addition, an increase in the sorptive capacity of the zone could potentially be achieved,
without experiencing a loss in conductivity, by increasing the clay content during

modification.

To the one who Showed me the path

ACKNOWLEDGMENTS

First and foremost I would like to thank my major advisor Dr. Roger B. Wallace
for his support, encouragement, and fatherly patience with me throughout my PhD work.
His intelligent diagnosis and invaluable guidance were values without which I could not
complete this task. Furthermore, I would like to thank the members of my guidance
committee, Dr. Thomas Voice, Dr. Steve Boyd, and Dr. Thomas Wolff for their
assistance to complete this dissertation.

In addition, I wish to thank my ofiice mates and hiends who made my study
easier and my life more enjoyable. My especial thanks to one of my dearest friends I have
ever had, W. James Gellner for his help with the preparation of this manuscript and
sharing his bright comments with me.

This research was supported by the National Institute for Environmental Health
Sciences Grant E80491] and the Michigan State University Institute for Environmental
Toxicology.

Finally, I am grateful to my family. My father and mother have provided me with
the motivation to continue my study. My wife was the lovely mother of our three

beautiful children and patiently shared the challenges of a student life with me.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES viii
LIST OF FIGURES ix
INTRODUCTION 1
Background 1
Front Micro-Scale Processes to Macro-Scale Hydraulic Properties 4
Clay Content And The Soil Structure 8
Treatment Level And The Soil Structure 10
References 12
CHAPTER 1
HYDRAULIC CONDUCTIVITY OF AN ORGAN OMODIFIED SOIL TO
WATER AND SOLTROL 15
Abstract 15
Introduction 16
Materials And Methods 20
”teary 23
Results And Discussion 23
Hydraulic properties of the permeants ................................................................................. 24
Clay structure ....................................................................................................................... 24
Hydraulic conductivity ......................................................................................................... 29
Under no load ......................................... - 29
Water saturated soils under load ...... - ........................... 32
Solhol saturated soils under load '37
Implications For Field Application 41
Referenm 43
CHAPTER 2

HYDRAULIC CHARACTERISTICS OF ORGAN OMODIFIED SOILS

vi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FOR USE IN SORPTIVE ZONE APPLICATIONS 46
Abstract 46
Introduction 47
Materials And Methods 50
Results And Discussion 53

Treatment and clay structure ............................................................................................... 53
Porosity ................................................................................................................................. 53
Conductivity .......................................................................................................................... 59
Implications For Field Application 68
References 71

CHAPTER 3

HYDRAULIC FEASIBILITY OF IN-SI TU MODIFIED SORPT IVE ZONES;

HYDRAULIC CONDUCTIVITY OF SOILS ORGANOMODIFIED TO

DIFFERENT LEVELS IN A BATCH SYSTEM 74
Abstract 74
Introduction 75
Materials And Methods 78
Results And Discussion 81

Treatment level and clay structure ....................................................................................... 81
Porosity. - ..... .................................................................................. 84
Conductivity .......................................................................................................................... 91
Implications For Field Application 96
Conclusions 99
References 100

APPENDIX

DETAILED MATERIALS AND METHODS 102
Soil Processing 102
Turbidity Tech 105
Consolidometer Assemblage 107
Vacuum Saturation 111
Blank Conductivity Measurements 115
Soil Packing 121
Loading And Conductivity Measurements 126
Consolidometer Disassembly 130
Calculations And Data Reduction 131

 

vii

Table l -

Table 2 -

Table 3 -

Table 4 -

Table 5 -

Table 6 -

Table 7 -

LIST OF TABLES

Hydraulic properties - specific weighty and viscosity u - of the two
permeants equilibrated with heated and unheated soils at 18 °C .................... 25

Relative turbidity of the supernatant for unheated and heated soil
samples in the two permeants scaled to the turbidity of unheated soil ............ 28

Average values of hydraulic properties of unheated and heated
soils at zero load ............................................................................................... 30

Average porosity, 11 ($8,) of soils at zero load, the first load,
and porosity decrease due to application of the first load ............................... 55

Average hydraulic conductivity, K (:tSK) of soils at zero load,
the first load, and its drop due to application of the first load ......................... 61

Average porosity, n (3:8,) of soils at zero load, the first load,
and porosity decrease due to application of the first load ................................ 85

Average hydraulic conductivity, K (:hSK) of soils at zero load,
the first load, and its drop due to application of the first load ......................... 93

Table 1A- Hydraulic conductivity of stones used in water and Solhol experiments ..... 121

viii

Figure] -

Figure 2

Figure 3

Figure 4

Figure 5

Figure6 -
Figure7 -
Figure8 -

Figure9 -

Figure 10-

Figurell-

Figure 12 -

Figure 13 -

Figure 14-
Figure 15 -

Figure 16-

Figure 17-

LIST OF FIGURES

 

Turbidity of soil suspension as affected by HDTMA heahnent level ......... 26
Hydraulic conductivity of heated and unheated soils permeated with

water ........ -_ ...................... 33
Effective pore radii of heated and unheated soils permeated with water ..... 34
Porosity of heated and unheated soils permeated with water ...................... 36

Hydraulic conductivity of heated and unheated soils permeated with
Solhol ........................................................................................................... 38

Effective pore radii of heated and unheated soils permeated with Solhol...39

Porosity of heated and unheated soils permeated with Solhol .' ................... 4O
Turbidity of soils with different fine contents .............................................. 54
Porosity of soils with 6%, 12%, 18%, and 24% added fines at different

loads .............................................................................................................. 58
Average porosity of a) unheated and b) heated soils with 6, 12, 18, and

24% added fines ........................................................................................... 60
Hydraulic conductivity of soils with different fines contents ....................... 64
Hydraulic conductivity of a) unheated, and b) heated soils with

different fines contents .................................................................................. 66
Comparison of average hydraulic conductivities (K); a) ratio of heated

and unheated soils, and b) all samples .......................................................... 69
Turbidity of soils with difi‘erent fine contents .............................................. 82
Porosity of soils with 6%, 12%, 18%, and 24% fines ................................... 88
Porosity of a) low-heated and unheated, and b) mid- and high-heated
samples (unheated and mid-heated data from the previous study) ............. 90
Hydraulic conductivity of soils with 6%, 12%, 18%, and 24% fines .......... 95

ix

Figure 18-

Figure 1A-

Figure 2A-

Hydraulic conductivity of a) low-heated and unheated, and b) mid- and
high-heated samples (unheated and mid-heated data fiom the previous

study) ............................................................................................................ 97
Components of the consolidometer ........................................................... 108
Schematic of the vacuum saturating set up ................................................ 112

A.

US

TE

et

19

al.

 

 

INTRODUCTION

BACKGROUND

The study of clay mineral-organic reactions was initiated in the early 19305
(Evans and Pancoski, 1989). In 1934, Smith reacted organic compounds with clays and
presented evidence that the reaction was ion exchange in nature (Grim, 1968). Jordan
(1949) was one of the early investigators studying clay-organic interactions who found
that organophilic bentonites swell and disperse in organic fluids. Since that time clay-
organic interactions have been studied by a number of authors and proven to be effective
in hansforming a naturally hydrophilic clay into an organophilic clay (Grim, 1968;
Mortland, 1970; Raussell-Colom and Serratosa, 1987; Boyd et al., 1988). However, the
use of organically modified clays in hazardous waste management applications is a
relatively new technology (Evans et al., 1989).

The enhanced sorption characteristics of modified clays could be managed to
intercept and retard a migrating contaminant plume and thus, serve as a useful tool for
many different remedial or containment applications such as containment barriers (Alther
et al., 1990; Boyd et al., 1991); in-situ sorptive zones (Boyd et al., 1991; Burris and
Antworth, 1992), for stabilization of hazardous wastes (Alther et al., 1990; Boyd et al. ,
1991), for water and wastewater heahnent (Srinivasan and Fogler, 1989, 1990; Alther et
al., 1990), and for an immediate response to terrestrial chemical spills to limit the extent

1

2
of contamination (Boyd et al., 1991). Each application requires a blockage or a bulk

transport of fluid through a heated soil mahix. Therefore, hydraulic conductivity plays an
essential role in the study of the feasibility of organomodified technology in all
applications. In fact, it is probably the most important parameter for assessing their
hydraulic feasibility (Evans, 1991).

The dependence of hydraulic conductivity K, [LfI‘], on the media’s permeability
1:, [L2], and the fluid’s specific weighty and viscosity u is well known (Freeze and Cherry

1979);

=7
“‘4

Based on this equation, the changes in hydraulic conductivity of a soil matrix can be
athibuted to changes in the permeant properties (of specific weight and viscosity) and
changes in the soil mahix geomehy (i.e. permeability).

While no generally accepted expression for permeability exists, two approaches
have been taken in the literature to relate permeability of a soil mahix to l) the physical
properties of soil grains or 2) the pore space properties. The first approach bases its
derivation primarily on the effect of soil grains on resisting the flow through a porous
media. Thus, in this approach, factors such as grain sizes and their dishibution, shape of
the grains, and grains packing factor are considered the major parameters influencing the
flow. The typical form of the expression for permeability in this approach can be
formulated as;

k= Cd2

3
where C is an empirical coefficient and d is a length characteristic of the soil grains

(Leonard, 1962; Freeze and Cherry, 1979; Holtz and Kovacs, 1981). This approach has
been mainly used for granular soils and is not valid when a significant amount of clay
exists in the soil mahix (Leonard, 1962).

The second approach considers the model of “bundle of capillary tubes” for the
soil mahix. In this approach, fundamental equations for flow through tubes with different
cross sectional areas are used and the porous media is simulated to a pack of tubes with
different diameters. Therefore, total porosity of the soil, some measure of the micro-scale
pore radius, and wetted perimeter of pores are the primary variables which are taken into
account in this approach. Hagen-Poiseuille and Kozeny—Carman equations, which relate
permeability to porosity and a characteristic radius of the pores in the soil mahix are
examples of this approach (Leonard, 1962; Corey, 1986). The mathematical relationships
derived in this approach can be simplified as;

k = nrez
in which n is the total porosity and re is an effective pore radius for the soil mahix
(Allred and Brown, 1994). This approach was adopted in our study because of the
presence of clays in the soil as well as the focus of study being the changes in hydraulic
conductivity due to organomodification which primarily affects the clay shucture.

In general, organomodification has the potential to alter hydraulic conductivity of
a soil matrix by changing the size and orientation of clay particles, geomehy of the pore
space in the mahix, and also through its effect on the fluid-clay interaction. In the
following subsection of this chapter, these microscale processes will be discussed when

the soil is permeated with water and/or an organic permeant. The effect of clay content,

4
and heahnent of the soil to different levels in this regard will be presented in the second

and third subsections, respectively. In each subsection, relevant studies that looked at the
changes in hydraulic properties of soils due to organomodification will also be presented.
FROM MICRO-SCALE PRocesses r0 MACRO-SCALE HYDRAULIC PROPERTIES

Theng (1974) reported that the replacement of natural cations on clays by organic
cations influences the shucture of clays. He found that organomodification induces a
face-to-face aggregation in clay particles, a process called flocculation. More recently, Xu
and Boyd (1994) studied the effect of initial cation type on flocculation process when
hexadecylhimethyl ammonium (HDTMA) is used to replace the cation. They showed that
initially dispersed Na-saturated clays flocculated when all Sodium cations were replaced
by HDTMA. They rationalized this behavior by measuring the elechophoretic mobility of
clays and showing that the replacement neuhalized the net negative elechostatic charge
on the clay particles.

Pure organic liquids are also known to produce flocculation in clays, however, by
a totally different mechanism (Brown, 1988; Broderick and Daniel, 1990). Bowders
(1985) mentions that the pertinent mechanisms are not well understood however, the
process can be summarized in shrinkage of diffuse double layer due to low dielechic
constant of organic liquids (compared to water) and aggregation or flocculation of clays.
Bowders and Daniel (1987) mention that, in addition to flocculation of clays, mixing
soils with pure organic compounds also deshoys the cohesive, plastic nature of the soil
for the same reasons.

In a somewhat macroscopic perspective, clay flocculation can be viewed as

affecting particle size dishibution, which in turn, affects consolidation behavior of the

5
soil (Holtz and Kovacs, 1981). Fernandez and Quigley (1991) packed clay soils in a dry

state and saturated them with water and two organic liquids (ethanol and dioxane). They
applied effective shesses to all samples, and measured their consolidation. Their data
indicate that the dry packed clay soil has less consolidation when saturated with an
organic liquid compared to water. They rationalized that flocculation of dry clays in
organic liquids (due to the collapse of diffuse double layer by pure organic liquids)
shengthens the soil peds and makes the soil less compressible. The effect of clay
flocculation, due to organomodification of the soil, on consolidation behavior of the soil
mahix has not been reported in the literature.

Clay flocculation also aflects pore sizes and their dishibution which influences
conductivity of the soil mahix (Bowders, 1985). In fact, clay flocculation is among the
interactions that increase the hydraulic conductivity of clay soils significantly
(Shackelford, 1994). Fernandez and Quigley (1991) reported increases of more than 10 to
IOOO-fold in conductivity of dry packed clay soils permeated with ethanol or dioxane
compared to water. This increase in conductivity was sustained throughout an effective
shess sequence of 0 to 160 kPa due to shonger flocculated soil peds which greatly
resisted macropore closure under effective shesses (Fernandez and Quigley, 1991).
Significant increase in conductivity of soils containing organobentonite, in replacements
for unheated bentonite, has been reported by Smith and Jaffe (1994). The comparison
between clay flocculation induced by the two totally different mechanisms (i.e. organic
permeant and organomodification) and its effect on consolidation and hydraulic

conductivity of the mahix is not explored in the literature.

6

Organomodified clays, which are flocculated in aqueous solutions, will disperse if
placed in pure organic liquids (Jordan, 1949). The mechanism of dispersion is established
by drawing the organic liquids (and not water) in between organoclay platelets. Wolfe et
al. (1985) used X-ray diffraction studies of organomodified monhnorillonite to determine
that d-spacing of organobentonite increases when hexane is used as the wetting fluid
relative to water.

Clay dispersion due to a different mechanism (high water content at the time of
packing a clay soil) and its effect on hydraulic conductivity of a wet packed clay soil is
discussed in geotechnical literature (Boynton and Daniel, 1985). According to Lambe’s
(1958) “particles orientation theory”, the arrangement of individual particles, which is
influenced by molding water content, conhols hydraulic conductivity (Benson and
Daniel, 1990). Lambe suggested that the soil particles are oriented in a random pattern
(with larger voids and higher conductivities) when soil is compacted dry of optimum
water content, and in a dispersed pattern (with smaller voids and lower conductivities)
when compacted wet of optimum water content. Although the mechanism of dispersion
of clays is different in this example (compared to dispersion of organoclays in organic
liquids), however, it is insightful as to the potentials of clay dispersion influence on
hydraulic conductivity of clay soils.

Another fundamental change in the clay surface property due to
organomodification is that by replacing natural inorganic cations of clays with organic
ones, their natural hydrophilic character changes to organophilic (Evans et al. 1989).
Hydrophilic nature of the clays is due to hydration tendency of naturally occurring metal

exchange cations on them. However, once these inorganic exchange cations are replaced

7
by organic ones, the clay surface property changes from hydrophilic to hydrophobic (or

organophilic) due to the affinity of organoclays for organic liquids (Jaynes and Boyd,
1990).

One implication of surface property change in organoclays is that they do not
draw water in between their platelets. This causes a major change in hydraulic behavior
of clays, especially swelling ones such as monhnorillonite, in water. Results of free swell
volume tests with many different organically modified swelling clays Showed that they
swelled less in water compared to organic liquids (Evans and Pancoski, 1989; Alther et
al., 1988; Evans et al., 1990; Evans and Alther, 1991). Therefore, it is implied that clays
(specially swelling ones) with low hydraulic conductivities will have higher
conductivities when organically modified.

Smith and Jaffe (1994) studied the incorporation of organobentonites into landfill
liners to retard the transport of organic contaminants through conventional liners. They
mixed 88% (of total mixture mass) Ottawa sand, none (or 4%) organobentonite, 12% (or
8%) unheated bentonite, and water uniformly and compacted it to simulate sand and
bentonite liners. Their results indicate that, due to partial replacement of organobentonite
for unheated bentonite, hydraulic conductivity increased by 10-fold. Similarly, Smith et
al. (1992) noted that the complete substitution of organobentonite for unheated bentonite,
in the same composition of soil as above, results in a hydraulic conductivity increase of
four orders of magnitude. Inability of organobentonite to intercalate any significant
amount of water was identified by these authors as the likely cause of the large

differences observed in their conductivities.

8
In conventional study of the flow through porous media, it is a fundamental

assumption that at boundaries where the fluid is in contact with solid, the fluid velocity
relative to the boundary is zero (Corey, 1986). This is referred to as “no-Slip” boundary
condition at solid surfaces. In the case of gas flow in porous media, at ordinary low
pressures, no-slip condition is not satisfied. This phenomenon has been called “gas
slippage” (Corey, 1986). As mentioned earlier, organomodification of clays changes their
surface property fi'om hydrophilic to hydrophobic. In the context of water flOw through
porous media, it could appear that a hydrophobic surface has a different potential to resist
flow compared to a hydrophilic surface. In exheme, it may change the boundary
condition at the clay water interface from “no-slip” to somewhat “yes-slip”! Furthermore,
in a water saturated soil mahix, a hydrophobic surface might be capable of
accommodating easier rearrangement of other particles under an external load, compared
to a natural hydrophilic clay surface. To my knowledge, possibility of such differences
has not been explored or investigated in previous researches in this area, nor was it
studied in the present work. However, the lumped effect of all such microscale changes
due to organomodification on consolidation and conductivity of soils was investigated
and is reported in the present work.
Cu 1’ CONTENTANO THE SOIL STRUCTURE

The shucture of natural soils can be described as being composed of clay particle
associations (microfabric) within a macroshucture of granular portion of the soil
(Mitchell, 1976). In terms of porosity, the total pore volume of a soil mahix can be
divided into micropores (within the microfabric) and macropores (in between larger

aggregates) (Yang and Barbour, 1992). While the role of clay content and its shucture

9
(flocculated or dispersed) is critical to hydraulic properties of a soil mahix, there are only

a few articles addressing organomodified clay content and its influence on the soil
shucture and its conductivity. However, in geotechnical literature, this subject has been
studied extensively using natural clays combined with sand and gravel at different
contents, for different engineering applications. Study of the influence of clay content on
hydraulic properties of the soils used in those applications can provide an insight as to the
sensible range of clay content and its potential effects on organomodified soil shucture.

On the upper limit, researchers were concerned about how much of gravel or sand
may be added to a clay deposit while keeping its very low hydraulic conductivity. Holtz
and Kovacs (1981) indicated that when the sand content is below 50%, the sand grains
are essentially floating in a clay mahix and have little effect on its engineering behavior.
Shakoor and Cook (1990) performed hydraulic conductivity tests on silty clay mixed with
different amounts of 13- to l9-mm diameter gravel particles and reported a large increase
in conductivity for gravel contents greater than 50%. Shelly and Daniel (1993) conducted
similar experiments on two different clay soils mixed with 10- to 19-mm diameter gravel
and found the critical gravel content of 60% at which conductivity sharply increased.
Both researchers indicate that at gravel contents of above 50 or 60%, amount of clays are
not enough to totally fill the voids created by gravel particles. Therefore, it can be
concluded that hydraulic behavior of the soil mahix is more sensitive to clay content if
the clay content is kept below 50%.

On the lower limit, amount of bentonite used in clay sand liners, to achieve a low
hydraulic conductivity of 10'7 cm/s, typically varies between 6 and 15% (Alther, 1987).

Daniel (1987) has noted however, that the benefits derived from increasing the percent

lO

bentonite in a sand bentonite mixture diminish as the percent bentonite increases above
10-12%. Increasing the percent bentonite above this percent he concludes, causes little
additional reduction in hydraulic conductivity. This is consistent with Kenney et al.
(1992) who reported that in a sand bentonite mixture, depending on the amount of
bentonite (varied fi'om 4 to 22% in their experiments), clays resided in the pore space
created by sand grains or in between them which made the clays part of the load bearing
mechanism of the soil mahix. They concluded that by the contents of 12%, bentonite was
completely filling the pore space created by the sand portion, and hence resulting in a low
conductivity of the mixture. It is concluded, in a very general sense, that a minimum of 4
to 6% of any type of clay is needed before a detectable change in hydraulic conductivity
of the soil mahix can be expected. In the experiments conducted in Chapters 2 and 3 of
this dissertation clay contents was varied fiom 6 to 24%.
TREA TMENT LEVEL AND THE SOIL STRUCTURE

Few studies on the effect of organomodification level on the clay shucture Show
that clay particles can be flocculated or dispersed depending on the modification level
(Xu and Boyd, 1994, 1995). Xu and Boyd (1995) description of the pertaining
microscopic processes may be summarized as follows; starting with a Na saturated
dispersed clay soil, HDTMA added in amounts far less than the cation exchange capacity
(CEC), replaces the Na cations on clays via ion exchange process. Their quantity is not
enough to totally neuhalize the original net negative charge on the clays and therefore,
not significant flocculation occurs. As HDTMA substitution approaches the CEC, the
clay surfaces will become elechostatically neuhal and the clays flocculate. The

replacement of Na by HDTMA is still mainly cation exchange process. HDTMA’S will

11
continue to adsorb to the exterior clay surfaces beyond the CEC, however, via

hydrophobic bonding which is the athaction of hydrophobic tails of adsorbed HDTMA’S.
This will create a net positive elechostatic charge on clays and cause their re-dispersion.
The effect of clay dispersion (though by other mechanisms) on soil shucture and
its influence on hydraulic behavior of the soil mahix was discussed earlier. Another
adverse effect of hydrophobically bonded HDTMA is that it desorb in water easier than
the cation exchanged one due to its weak bonding (Xu and Boyd, 1994). The effect of this
desorption on the clay shucture is not explored in the literature. However, based on Xu
and Boyd’s (1995) rationale on the cause of the dispersion (i.e. net positive charge on the
clays due to hydrophobically bonded excess HDTMA), one can conclude that after
desorption of excess HDTMA, clay particles will retain their flocculated shucture which
they had before hydrophobic adsorption of the excess HDTMA took place.
Redistribution of heahnent profile in an in-situ modified soil column, possibly
due to the desorption of excess HDTMA, its hansport down gradient, and its re-
adsorption to un-neuhalized clay surfaces, is reported by Burris and Antworth (1992).
They Showed that the initial heahnent level varied from ~0.5% organic carbon (CC) to
zero in the first 10 cm of their 25-cm long columns. This profile spread down gradient
(with a maximum of ~0.3% OC which dropped to zero in the first 20 cm of the column)
due to 400 pore volume flushing of the column with HDTMA-free water. While their
data does not distinguish whether the redishibution occurred due to desorption and later
re-adsorption of hydrophobically bonded HDTMA or due to partial migration (and later
enhapment) of the dispersed clay particles, it gives rise to concerns on potential adverse

effects of heahnent to levels far above CEC on conductivity of the soil, in an in-situ

12
injection of organic surfactants. No specific study of the effects of heahnent level on

hydraulic conductivity of organomodified soils has been reported in the literature.
REFERENCES

Allred, B., Brown, G. 0., 1994. Surfactant-Induced Reductions in Soil Hydraulic
Conductivity. Ground Water Management and Remediation, 174-184.

Alther G. R., 1987. The Qualifications of Bentonite as a Soil Sealant. Eng. Geology,
23:177-191.

Alther, G. R., Evans, J. C., and Pancoski, S. E., 1988. Organically Modified Clays for
Stabilization of Organic Hazardous Wastes. Proceeding of Superfund ‘88., pp.
440-445.

Alther, G. R., Evans, J. C., and Pancoski, S., E., 1990. No Feet of Clay. Civil Eng., New
York., 60(8), 60-61.

Benson, C. B., and Daniel, D. E., 1990. Influence of Clods on Hydraulic Conductivity of
Compacted Clay. J. Geotech. Eng. ll6(8)1231-1248.

Bowders, J. J ., Jr., 1985. The Influence of Various Concenhations of Organic Liquids on
the Hydraulic Conductivity of Compacted Clay. Geotechnical Eng. Dissertation
GT85-2, Geotech. Eng. Center, Civil Eng. Dept, U Texas, Austin, TX, 219 pp.

Bowders, J. J ., and Daniel, D. E., 1987. Hydraulic Conductivity of Compacted Clay to
Dilute Organic Chemicals. J. Geotech. Eng., ll3( l2)1432-l448.

Boyd, 8. A., Mortland, M. M., and Chiou, C. T., 1988. Sorption Characteristics of
Organic Compounds on Hexadecylhimethylammonium-smectite. Soil Sci. Soc.
Am. J. 52:652-657.

Boyd, S. A., Jaynes, W. F ., and Ross, B. S., 1991. Immobilization of Organic
Contaminants by Organo-clays: Application to Soil Restoration and Hazardous
Waste Containment. In R A. Baker (ed.) Organic Substances and Sediments in
Water, Vol. 1, Publishers, Chelsea, MI., pp 181-200.

Boynton, S. S., and Daniel, D. E., 1985. Hydraulic Conductivity Tests on Compacted
Clay. J. Geotech. Eng., 111(4)465-478.

Broderick, G. P. and Daniel, D. E., 1990. Stabilizing Compacted Clay Against Chemical
Attack. Am. Soc. Civ. Eng., Geotech. Eng., 116:1549-1567.

Brown, K. W., 1988. Review and Evaluation of the Influence of Chemicals on the
Conductivity of Soil Clays. Project summary, EPA/600/S2-88/016.

Burris, D. R. and Antworth, C. P., 1992. In Situ Modification of an Aquifer Material by a
Cationic Surfactant to Enhance Retardation of Organic Contaminants. J. Cont.
Hydrol., 10:325-337.

Corey, A. T., 1986. Mechanics Oflmmiscible Fluids in Porous Media, Water Resources
Publication, Littleton, CO., 255 pp.

Daniel, D. E., 1987. Earthen Liners for Land Disposal Facilities. Proc. Geotech. Practice
for Waste Disposal, ASCE, New York, N. Y.

Evans, J. C., Pancoski, S. E., and Alther, G. R., 1989. Organic Waste Treahnent with
Organically Modified Clays. Res. Dev. [Rep.] EPA/600/9-89/072, Int. Conf. New
Front. Hazard Waste Manage, 3rd, pp. 48-58.

13

Evans, J. C. and Pancoski, S. E., 1989. Organically Modified Clays. Preprint:
Transportation Research Board, 68th Annual Meeting, pp. 160-168.

Evans, J. C., Sambasivam, Y., and Zarlinski, S., 1990. Attenuating Materials in
Composite Liners. Waste Containment Systems: Conshuction, Regulation, and
Performance, R. Bonaparte, ed., Am. Soc. Civ. Eng., New York, N.Y., pp. 246-
263.

Evans, J. C., 1991 . Geotechnics of Hazardous Waste Conhol Systems. Chapter 20,
Foundation Engineering Handbook, 2nd ed., Ed. H. Y. Fang, Von Noshand
Reinhold Co., New York, N. Y.

Evans, J. C. and Alther, G. 1991. Hazardous Waste Stabilization Using Organically
Modified Clays. Proc., Geotech. Eng. Congress, Vol. II, ASCE, New York., NY,
pp. 1149-1162.

Fernandez, F. and Quigley, R. M., 1991. Conholling the Deshuctive Effects of Clay-
Organic Liquid Interactions by Application of Effective Shesses. J. Canadian
Geotech., 28:388-398.

Freeze, R. A. and Cherry, J. A., 1979. Groundwater, Prentice-Hall, Inc., Englewood
Cliffs, N. J. 604 pp.

Grim, R. E., 1968. Clay Mineralogy. 2nd ed., McGraw-Hill. New York.

Holtz, R. D. and Kovacs, W. D., 1981 . An Introduction to Geotechnical Engineering,
Prentice-Hall, Inc., Englewood Cliffs, NJ. 733 pp.

Jaynes, W. F. and Boyd, S. A., 1990. Trimethylphenylammonium Smectite as an
Effective Adsorbent of Water Soluble Aromatic Hydrocarbons. J. Am. Waste
Manag. Assoc. 40:1649-1653.

Jordan, J. W., 1949. Alteration of the Properties of Bentonite by Reaction with Amines.
Mineralogical Magazine and Journal of the Mineralogical Society, 28(205)598-
605.

Kenney T. C., Van Veen W. A., Swallow M. A., and Sungaila, M. A., 1992. Hydraulic
Conductivity of Compacted Bentonite-Sand Mixtures. Canadian Geotech. J.
29:364-3 74.

Lambe, T. W., 1958. the Shucture of Compacted Clay. J. Soil Mechanics and Found.
Division, Am. Soc. Civil Eng., Vol. 84, SM2, pp. 1654-1 -l654-34.

Leonard, G. A. Ed, 1962. Foundation Engineering. McGraw-Hill Book Company, Inc.,
New York, NY, 1136 pp.

Mitchell, J. K., 1976. Fundamentals of Soil Behavior, John Wiley & Sons, Inc., New
York, 422 pp.

Mortland, M. M., 1970. Clay-Organic Complexes and Interactions. Advances in
Agronomy, Academic Press, Inc., 22:75-117.

Raussell-Colom, J. A., and Serratosa, J. M., 1987. Reactions of Clays with Organic
Substances. in Chemishy of Clays and Clay Minerals, ed. Newman A. C. D., pp.
371-422, John Wiley & Sons, New York.

Shackelford, C. D. 1994. Waste-soil Interactions That Alter Hydraulic Conductivity.
Hydraulic Conductivity and Waste Contaminant Transport in Soil, ASTM STP
1142, David E. Daniel and Stephen J. Trautwein, Eds., Philadelphia, PA, pp. 111-
1 68.

l4

Shakoor, A. and Cook, B. D. 1990. The Effect of Stone Content, Size, and Shape on the
Engineering Properties of a Compacted Silty Clay. Bulletin of Association of
Eng. Geologists, XXVII(2), 245-253.

Shelly, T. L. and Daniel, D. E., 1993. Effect of Gravel on Hydraulic Conductivity of
Compacted Soil Liners, ASCE, J. Geotech. Eng., ll9(1)54-68.

Smith, J. A. and Jaffe, P. R., 1994. Benzene Transport Through Landfill Liners
Containing Organophilic Bentonite. Am. Soc. Civ. Eng., Environ. Eng. 120:1559-
1577.

Srinivasan, K. R and Fogler, H. S., 1989. Use of Modified Clays for the Removal and
Disposal of Chlorinated Dioxins and other Priority Pollutants from Indushial
Wastewaters. Chemosphere., l8(1-6)333-342.

Srinivasan, K. R. and Fogler, H. S., 1990. Use of Inorganic-Organoclays in the Removal
of Priority Pollutants from Industrial Wastewaters; Shuctural Aspects. Clays and
Clay Mineral, (38)277-286.

Theng, B. K. G. 1974. The Chemishy of Clay Organic Reactions. John Wiley & Sons,
New York.

Wolfe, T. A., Demirel, T., Daumann, E. R., 1985. Interaction of Aliphatic Amines with
Monhnorillonite to Enhance Adsorption of Organic Pollutants. Clays Clay Miner.,
33:301-311.

Xu, S. and Boyd, S. A., 1994. Cation Exchange Chemishy of Hexadecylhimethyl
ammonium in a Subsoil Containing Vermiculite. Soil Sci. Soc. Am. J ., 58:1382-
1391.

Xu, S. and Boyd, S. A., 1995. Cationic Surfactant Sorption to a Vermiculitic Subsoil via
Hydrophobic Bonding. Environ. Sci. Technol., 29:312-320.

Yang, N. and Barbour, S. L., 1992. The Impacts of Soil Shucture on the Hydraulic
Conductivity of Clays in Brine Environments. Canadian Geotech. J., 29(5)730-
739.

Chapter 1

HYDRAULIC CONDUCTIVITY OF AN ORGAN OMODIFIED SOIL

TO WATER AND SOLTROL

ABSTRACT

Hydraulic feasibility of organomodified sorptive zones was studied by
investigating the effects of batch organomodification (heahnent) on hydraulic
conductivity of a sandy loam under difl‘erent effective shesses. Porosity and conductivity
to water and Solhol 170 (an LNAPL) were measured as a function of effective shess; and
their correlation with clay flocculation or dispersion and its influence on pore size were
sought. Treahnent was accomplished in a well-mixed batch system by adding
hexadecylhimethyl ammonium (HDTMA) solution to the soil in an amount equivalent to
its cation exchange capacity (CEC). A fixed ring consolidometer, in which heated and
unheated soils were packed dry and loaded, was employed for conductivity
measurements. Under no load, heated soils were 9 and 20-fold less conductive than
unheated ones to water and Solhol, respectively. This occurred because heahnent
collapsed the initially dispersed structure of unheated clays which had “fluffed up” the
unheated soil mahix. Compared to unheated soil, heated soil saturated with water
exhibited smaller loss of conductivity due to increased load, such that it became 5-fold

more conductive than unheated soil at the highest efl’ective shess (8 tsf). Observed
15

16

differences in conductivity were explained in terms of changes in fluid properties and soil
mahix geomehy. Changes in clay shucture (flocculated or dispersed) due to the heahnent
and permeant were evaluated by measuring soil suspension turbidity. The role of clay
shucture in conholling the initial effective pore radius in soils and hence their
conductivities, as well as, in conholling changes in effective pore radius during
consolidation is discussed. It was found that ex-situ heahnent of soils with HDTMA for
the purpose of creating sorptive zones is hydraulically feasible under the experimental
conditions used in this study.
INTRODUCTION

Organomodified subsoils and aquifer materials have recently been advocated as
permeable sorptive zones for contaminant plume management (Boyd et al. , 1988, 1991;
Lee et al., 1989; Xu et al., 1996). In this application, dissolved organic contaminants
would be intercepted and immobilized when they flow through the zone. From a practical
stand point, knowledge of the degree to which organomodification alters hydraulic
conductivity is essential in assessing the hydraulic feasibility of these zones. Numerous
investigations have revealed much about the micro-scale changes in clay physical
shucture brought on by organomodification and hence the potential it possesses to alter a
natural soil’s conductivity to both aqueous solutions and pure organic permeants (Xu and
Boyd, 1994, 1995a, b, c). To date however, there has been very little investigation to
determine the degree of change in conductivity brought about by organomodification of
natural soils so that the hydraulic properties of sorptive zones remain unknown.

The limited sorptive capacity of some soils can be greatly improved by

organomodification. Natural deposits of low organic matter content soils, subsoils, and

l7

aquifer materials have little sorptive capacity for removing nonionic organic compounds
(N OCs) from water. Organomodification of soils is achieved by replacing their native
inorganic cations with large organic cations through ion exchange reactions (Lee et al.,
1989; Jaynes and Boyd, 1991). The organic cations that have been studied most

extensively are quaternary ammonium compounds (QACS) of the general form
[(CH3,)3NR]+ where R is an alkyl or an aromatic hydrocarbon (Boyd et al., 1991). When

R is a relatively large alkyl hydrocarbon, as is the case for hexadecylhimethyl ammonium

(HDTMA) where R=(CH2)15CH3, then the modified clays are rendered organophilic due

to the alkyl hydrocarbon tails anchored to the clay surface (Jaynes and Boyd, 1991). The
agglomeration of alkyl tails of sorbed QACS forms a highly effective sorptive phase for
removing NOC’s from water (Sheng et al., 1996a, 1996b).

Permeable sorptive zones could be created ex-situ by excavating soil, heating it
with cationic surfactants such as HDTMA, then returning the organomodified soil. The
hydraulic conductivity of organomodified soil is important because it provides basic
information required to determine the size of the sorptive zone needed to prevent
contaminated fluid from flowing around the zone. From a design standpoint it is desirable
to match the conductivity of the zone, as closely as possible, to that of the surrounding
media. This could minimize either rapid funneling through the zone or fluid bypass. For
sorptive zones placed downsheam of an LNAPL spill, the organomodified soil could be
exposed to both LNAPL and aqueous permeants. Therefore, knowledge of the
conductivity to both permeants is of importance. Because the aqueous solubilities of

common organic contaminants are usually low, the conductivity of soils to an aqueous

18

solution is not expected to be effected by the presence of a dissolved organic compound
(Abdul et al., 1989; Evans et al., 1989). Pure organic liquids, on the other hand, are
known to affect the physical shucture of both natural and organomodified clays and may
therefore alter the conductivities of soil containing these materials (Theng, 1974; Weiss,
1963; Brown, 1988; Broderick and Daniel, 1990).

The microscopic differences between organomodified clays and unheated clays
provide an indication of the possible affect of organomodification on hydraulic
conductivity. For example, QAC molecules, added in an amount close to the soil’s cation
exchange capacity (CEC), will replace the naturally occurring cations on clays causing
dispersed clays to flocculate in aqueous solutions (Xu and Boyd, 1994, 1995a, b). Pure
organic liquids are also known to affect the flocculation and dispersion of clays. They
flocculate natural clays (e. g. Na-bentonite) by shrinking their diffuse double layers
(Bowders, 1985). Alternatively, organic liquids may be intercalated by organomodified
clays (e. g. HDTMA-smectite) causing interlayer expansion and ultimately dispersion
(Jordan, 1949; Theng, 1974; Weiss, 1963; Wolfe, 1974). Although clay flocculation
typically increases the hydraulic conductivity of packed clay soils (Bowders, 1985;
Shackelford, 1994), there are many other factors that may influence hydraulic
conductivity of a batch heated organomodified soil, so that determination of the degree of
change in hydraulic conductivity produced by organomodification of soils that conduct
either water or organic liquids requires direct measurement of hydraulic conductivity.

Very few measurements of the hydraulic conductivity of organomodified soils to
aqueous permeants have been reported, and there are apparently no measurements with

regard to liquid organic permeants. Burris and Antworth (1992) employed columns and a

l9
sand box aquifer model to study the organomodification, using HDTMA, of an aquifer

material with the objective of enhancing retardation of organic contaminants. They
reported having observed no detectable increase in column back pressure, suggesting
qualitatively that hydraulic conductivity of their soil may not have been significantly
reduced by organomodification. More recently, Smith and Jaffe (1994) studied the use of
organomodified bentonites to retard the aqueous hansport of NOCs through conventional
sand/Na-bentonite liners. Their results indicate that, due to partial or complete
substitution of organobentonite for unheated Na-bcntonite, hydraulic conductivity to
water increased by 10 - 400 fold (Smith et al., 1992). They athibuted the increase in
conductivity to the change in bentonite from a swelling to an essentially non-swelling
clay due to organomodification. These authors apparently employed swelling clays (N a-
bentonite) to achieve very low hydraulic conductivities required for liner application, and
organoclays to retard NOC mobility. Their study focused on evaluating the suitability of
using organomodified clays in liner applications. Investigating sorptive zone application,
in the form of a permeable heahnent wall, necessitates measurements of conductivity of
different types of media at different effective shesses.

The objectives of this study were to evaluate the hydraulic feasibility of ex-situ
heated sorptive zones by (i) determining the change in conductivity of an aquifer type
soil due to batch organomodification using water and an LNAPL as permeants, (ii)
monitoring this change under different effective shesses and (iii) seeking its relation to
microscale processes of clay dispersion and flocculation. To evaluate changes in

conductivity that might be encountered under field conditions, unheated soils were used

20
essentially as they were obtained, and compared to heated soils that were subject to

mechanical mixing and chemical modification.
M4 TERIALS AND METHODS

Two soils were employed in the hydraulic conductivity experiments. They
differed only in that the heated soil was organomodified where as the unheated soil was
not. The unheated soil was a B-horizon Oshtemo soil that was obtained at a depth of
about 2 ft at the Kellogg Biological Station, Hickory comers, MI. The soil was air dried
and passed through a US standard #20 (0.84 mm) sieve. Exchangeable cations of Ca”,
Mg”, K‘, and Na+ were present at relative amounts of 72%, 22%, 4%, and 2%
respectively. The major clay minerals included the limited swelling clay (vermiculite) and
the non-swelling clays (illite, kaolinite, and hydroxy-aluminum interlayered vermiculite)
as determined by X-ray diffraction. Mechanical analysis showed 78% sand, 19% clay, 3%
Silt, and 0.4% organic matter content in the soil. Sieve analysis and hydrometer tests
(ASTM C136 and D422) were performed on the soil using seven sieves (US standard
sieve numbers 20, 30, 40, 60, 70, 100, and 200) and a gram-per-liter type hydrometer.
The overall grain size distribution had a uniformity coefficient (Du/D10) of 13 and a
curvature coefficient (Dwlelo‘Dw) of 4 so that the soil was considered well graded
(Holtz and Kovacs, 1981). The soil also met a conservative sand filter criteria (D,,SO.5
mm) for fine-grained clays (Sherard et al., 1984) and hence, minimal loss of fines was
expected. Specific gravity of solids was measured and found to be 2.67 (ASTM D854).

The heated soil was prepared by mixing HDTMA with the above soil in a batch

process. HDTMA heahnent level corresponded to the first plateau of the HDTMA

21
adsorption isotherm which was 46.1 mmole/kg. This was taken as a practical estimate of

the CEC. Treahnent was accomplished by adding an aqueous solution of HDTMA-Cl to a
20:1 suspension of HPLC-grade water and soil which was stirred at 120 rpm for 1 minute
followed by 20 rpm for 30 minutes. Treated soil was settled out of suspension and
washed once with HPLC-grade water. Subsequently, it was airdried, gently ground with a
mortar and pestle, sieved through a US standard #20 sieve, and kept in an open pan.

An aqueous solution of 1 mM NaCl, HPLC-grade water and Solhol 170 (Phillips
Peholeum Co.) were used as permeants. Solhol 170, a mixture ofClO-C15 isoalkanes, is
a light non-aqueous phase liquid. The kinematic viscosity and specific weight of water
and Solhol were measured before and after contact with heated and unheated soils
(ASTM D445-88 and D446-89a). For these measurements water and Solhol were
equilibrated with heated or unheated soil in a 5:1 ratio for 24 hrs. A Cannon-Fenske
Routine viscometer was used to measure viscosity. All measurements were conducted in
a constant temperature bath at room temperature.

Two sets of turbidity measurements were made. The first set was made to
establish how the HDTMA heahnent level affected soil suspension turbidity and thereby
provide a clear link to the understanding of microscopic clay behavior already developed
by Xu and Boyd. The procedure employed was consistent with that used by Xu and
Boyd (1995a) and in compliance with ASTM D1889. Soil samples were mixed with an
appropriate volume of an HDTMA aqueous solution in a 1:10 ratio in 25-ml Corex tubes
at HDTMA concenhations that ranged from 0.01 to 3.0 times the CEC. The tubes were
then shaken for 4 days in a rotating shaker and allowed to stand for 0.5 hr immediately

prior to measurement. Turbidity was determined by a Hach Model 2100A turbidimeter on

22
the soil suspension sampled from 1 cm below the liquid surface in each tube. The second

set of turbidity measurements was made to determine whether the flocculated shucture of
heated clays, suggested by the first set of measurements, was altered because soil
samples were air dried prior to packing in the consolidometer and to establish the
dispersed state of heated clays in the presence of Solhol. In these experiments, each soil
(heated air dried soil, and unheated air dried soil) was mixed with either water or Solhol,
and the turbidity of each supernatant was measured. The soil to liquid ratio and mixing
procedure employed above were followed here.

A fixed ring consolidometer, which allows falling head conductivity
measurements, and a load frame equipped with a precision displacement hansducer
(0.0001 inch) were employed to load the samples. Treated and unheated samples were
packed by pouring air dried soil into the consolidometer ring with filter paper (Whahnan
#1) on the top and bottom of the sample. Samples were then tamped slightly to flatten the
upper surface and compacted with a drop hammer to a nominal initial thickness of 2.5 cm
and vacuum saturated in a desiccator with deaired, HPLC-grade water or Solhol.
Saturated samples were subjected to an effective shess (loading) sequence from 0 to 8 tsf
using the conventional consolidation procedure (ASTM D243 5-90). Each load was
maintained for a duration greater than Taylor’s time for 90% consolidation before
porosity, conductivity, and temperature measurements were made.

Hydraulic conductivity of samples at different effective shesses was measured by
the falling head method (ASTM D5084-90) at room temperature. Head losses in the
tubes, connectors, porous stones, and filter papers were quantified by assembling the

consolidometer without soil and measuring its conductivity (Kufl) at different effective

23

shesses with water and Solhol. Hydraulic conductivity of the soil sample was calculated
accounting for the predetermined Kb“. Two or three conductivity measurements were
made at each load and adjusted to reflect a fluid temperature of 20 °C. At each load, the
sample thickness was measured and porosity was calculated from the measured thickness
of the consolidated sample.
THEORY

The dependence of hydraulic conductivity K on the media’s inhinsic permeability

k and the fluid’s specific weighty and viscosity u is well known:

K=k1
,u

Employing the analogy of Poiseuille flow through a bundle of capillary tubes (DeWiest,
1965; Corey, 1986; Allred and Brown, 1994) permits explicit expression of the

conhibution of porosity, n, to hydraulic conductivity:

K=nr2

hl‘

Here the porosity accounts for the volume of pore space available to conduct fluid while
r, the effective pore radius, is a macroscopic parameter that provides a lumped indication
of the relative size of the pores and the shape factor.
RESUL IS AND DISCUSSION

Because of their influence on hydraulic conductivity, the hydraulic properties of
the permeants and the clay shucture of the soils are discussed before presenting the

conductivity results.

24
Hydraulic properties of the permeants

Values of specific weighty and viscosity u were measured (Table 1) for the water
and Solhol that were employed as permeants. Prior to measurement, these permeants had
been equilibrated with either heated or unheated soil. The changes in viscosity and
specific weight of the permeants brought about by the addition of HDTMA to the soil
were less than 1% in the case of water and less than 2.5% in the case of Solhol (Table 1).
Consequently, the differences in the conductivity to either permeant between the heated
and unheated soils, which will be discussed later, were not the result of changes in the
hydraulic properties of the permeants. On the other hand, the values of y/u for water and
Solhol were significantly different so that the conductivities of heated and unheated soils
were expected to depend upon which permeant was employed.

Clay structure

Measurements of the soil-water suspension turbidity from soil samples that were
equilibrated with different amounts of HDTMA in a batch process are shown in Figure 1.
The turbidities plotted were normalized by the turbidity of a sample that contained no
HDTMA. The amount of HDTMA added is shown as multiples of the CEC (46.1
mmole/kg). Unheated sample (0 CEC), and to varying degrees samples heated to levels
above 1.2 CEC, were turbid. At heahnent levels between 0.03 CEC and 1.2 CEC, soil-
water suspensions had relative turbidities that were nearly zero.

On this basis the two soils selected for use in studying hydraulic conductivity, the
unheated (0 CBC) and the heated (1 CBC) soils, appeared to possess different clay

shuctures. The work of Xu and Boyd (1994, 19950) supports this determination, having

25

Table 1: Hydraulic properties - specific weight y and viscosity u - of the two
permeants equilibrated with heated and unheated soils at 18 °C

 

 

Unheated soil Treated soil
Specific weight (N I m’) ‘ 9767 9767
Water
Viscosity (Pa . s) 1.03 X10<1 1.03 X 10"
Specific weight (N I m’) 7587 7592
Solhol
Viscosity (Pa . s) 3.31 X 10.3 3.23 X 10'3

 

 

26

md

.62.. «sesame... (2.5: 3 afloat.» on «6.9.32.» =3 .0 3.235... a. 2:2".

owo 2 engage: 5.23%..» :3 2 emcee (5.5: B EsoE<

N

m.—.

 

v

md

 

Nd

 

\.\

V.
o

 

 

 

co.
0
Known uncles

 

 

 

 

 

 

 

md

27
established that the dispersed clay shucture of the unheated soil is associated with high

turbidity and the flocculated clay shucture of heated soil is associated with low turbidity.
Clay flocculation is among the interactions known to Significantly increase the hydraulic
conductivity of clay soils (Shackelford, 1994). Because the heated soil was air dried
following heahnent and then resaturated for conductivity experiments, we believed that it
was important to determine that the flocculated shucture of clays was retained despite air
drying and resaturation.

A second set of turbidity measurements (Table 2) was made to evaluate the affect
of air drying the heated soils prior to resaturating them and to provide an indication of the
differences in the clay shucture of the two soils when permeated by SolhoL Turbidities of
the soil suspension in both water and Solhol were measured. As in the previously
discussed set of turbidity measurements, the heated soil sample (1 CBC) was much less
turbid in water than was the unheated sample; it had only 4% of the turbidity of the
unheated sample. This indicated that the heated clays retained their flocculated shucture
although they were air dried prior to being packed in the consolidometer and then were
saturated with water. Therefore, in the conductivity experiments that are discussed later
where water was employed as the permeant the heated samples had a relatively
flocculated clay shucture and the unheated ones had a more dispersed clay shucture. On
the other hand, it was the unheated soil sample that was less turbid in Solhol than was the
heated sample; it had only 13% of the turbidity of the heated sample. Indicating that

unheated clays were more flocculated compared to heated ones when Solhol was the

permeant.

28

Table 2: Relaflve turbidity of the supernatant for unheated and
heated soil samples in the two permeants scaled to the turbidity of
unheated soil (i.e. 945 NTU)

Unheated soil Treated soil

 

Water 1 0.04

Solhol 0.003 0.03

 

29
Hydraulic conductivity

Under no load

Hydraulic conductivity to water of the unheated soil at zero load averaged 9-fold
higher than that of the heated soil (T able 3). As indicated earlier this difference in
conductivity was not influenced by differences in the hydraulic properties of the waters
that had contacted the heated and unheated soils All of the difference can be attributed to
differences in the inhinsic permeabilities (Table 3) of the unheated and heated soils.
Equation 2 shows that the additional pore volume provided by the 12% higher average
porosity of the water saturated unheated soil (Table 3), compared to the heated soil,
accounted for only 12% of the 9-fold higher conductivity of the unheated soil, and that
the higher conductivity of the unheated soil was mainly attributed to that soil’s larger
effective pore radius (Table 3). The higher conductivity, higher inhinsic permeability,
and larger effective pore size occurred in the unheated soil which had a dispersed clay
shucture; this will be discussed further below.

For the Solhol saturated soils, the hydraulic conductivity of the unheated soil was
on average 19-fold higher than the conductivity of the heated soil at zero load (T able 3).
The higher conductivity of unheated soil was again entirely attributed to differences in
inhinsic permeabilities (T able 3). The larger effective pore radius of the unheated soil
was the primary reason for its higher hydraulic conductivity. Larger effective pore radius
accounted for 87% of the 19-fold higher conductivity while porosity accounted for only
13% of this difference. The higher conductivity, higher inhinsic permeability, and larger

effective pore size all occurred in the unheated soil which was similar to what was

30

358.330 33.30. 095 ES. Be 33.9. 8205 corona. .
acestonxe 28.32 x5 EB. 2a 829. omega 3:80.. r

8983 85:82”. 8.988.“. 8588.”. 828cc 8283 2325. .50

3.8 «we. 88 «.8. 8.8 3.8. 8.8 3...... 6.8 «no... :8 3.3 :8 3.83..

33.8 #32 38mm. #33 .388 «883 33.8 «3.8.4 .58 8...:

22. aim

63... «8mg 2 $8 «new: ET. 8 «Home... 338 «33 £5. 3.38.52.

9.2.3...

$38 «38.... 638 «8a.... 838 «mom: 83.8 «News 5...... £533....“
2.3...

 

.3238 .228 .3828 5...... 89.8.. >5 .8523 .288 .8338 5...... 83.8.. to

 

 

=ou 6800.; =00 93005.5

dosage corona.— e5 .0 ..o....o ecu—EEO 23 26:0 335.3ch 5 3...; 2:. .eno. Eon
«a 232...: >50 :05 8 .3225 e. e=oo e838 use 0335:: ac 33.2.05 3.323: .0 3...: euEe>< “n 03¢...

31

observed in the water saturated soils previously discussed. Unlike the water saturated
soils where the higher conductivity was associated with the dispersed clay structure, the
Soltrol saturated soil that exhibited the higher conductivity, the untreated soil, had a
flocculated clay structure.

For each permeant, the untreated soil exhibited a higher conductivity than the
treated soil prior to loading (Table 3). This was due to the initially dispersed structure of
untreated clays which produced a larger effective pore radius in the untreated soil
compared to the treated soil. Our porosity data for the dry packed soils (Table 3) showed
that the untreated clays “flufl‘ed up” the untreated soil matrix relative to the treated soil
matrix when soil was first packed in the consolidometer. This “flufi‘ed up” condition was
maintained when the untreated soil was saturated, irrespective of which permeant was
employed. The untreated soil’s higher porosities (about 17%) were apparently due to the
more dispersed structure of the dry untreated clays which produced a greater ability for
the untreated clays to resist compaction during packing. For both permeants, the larger
effective pore radii (Table 3) were associated with the “fluffed up” untreated soil
matrices. As discussed earlier, this was the main contributor to the higher conductivity of
untreated soils relative to treated ones. In the dry packed treated soil the clay sheets were
stacked in flocs and this soil felt noticeably “greasier” both of which may have
contributed to its greater susceptibility to compaction during packing and the occurrence
of smaller effective pore radius (Table 3).

i For the untreated soil, the difference between its hydraulic conductivity to Soltrol

and to water depended in part on changes in the clay structure that were chemically

32

induced by the Solh'ol. This was apparent because the intrinsic permeability of the
untreated soil that was Solhol saturated was 5.7-fold larger than the inhinsic permeability
of the untreated water-saturated soil. This chemically induced effect was not so apparent
in the hydraulic conductivities because it was partially offset by the difl‘erence in the
hydraulic properties (y/u) of the two permeants. The greater intrinsic permeability of the
Solhol saturated soil occurred because saturation with Soltrol caused the dispersed clay
shucture of the dry packed soil to flocculate. This produced the larger effective pore
radius of the Soltrol saturated soil which was responsible for its larger intrinsic
permeability (Brown, 1988; Broderick and Daniel, 1990).

Water saturated soils under load

As a confining load was placed on the water saturated soils and then serially
increased, the untreated soil remained more conductive than the treated soil at low
efi‘ective stresses (Figure 2). At the lowest load (0.25 tsf) the untreated soil was 3.6-fold
more conductive than the treated soil; and the untreated soil exhibited a higher rate of
conductivity loss with increased load than did the treated soil. This resulted in a situation
where continued loading first eliminated the difference between untreated and treated soil
conductivities and then reversed the situation, so that at loads greater than 1.3 tsf,
conductivity of the treated soil was greater than that of the untreated soil. At the highest
load (8 tst) the treated soil was 5-fold more conductive than the untreated soil.
Throughout the loading sequence, most of the difference in conductivity between the
unheated and treated soils was due to the difference in their effective pore radii (Figure

3). At the lowest loads (0.25 tsf) the unheated soil had a larger efl‘ective pore radius than

33

1 E-02

Untreated

Hydraulic Conductivity, K (cmls)
F5
6
0)

E

 

1E-05
0.1 1 10

Load (tsf)

Figure 2: Hydraulic conductivity of treated and untreated soils
permeated with water

34

1 E-07

Untreated

Treated

Effective Pore Radius, r2 (cm’)
6‘:
b
on

.3
m
b
‘0

 

1E-10
0.1 1 10

Load (tsf)

Figure 3: Effectlve pore radius of treated and untreated soil
permeated with water

35

the heated soil had, where as the effective pore radius was larger in the treated soil at
high loads. Untreated soil porosities which were greater than or equal to heated soil
porosities (Figure 4) at all loads only made a minor conhibution (<l4%) to the higher
conductivities of the untreated soil at loads below 1.3 tsf.

Differences between hydraulic conductivity of the heated and untreated soils
during loading can be interpreted in terms of initial mahix sh'uctures and their different
responses to load. The greater hydraulic conductivity of the untreated soil at the lowest
load (0.25 tsf) was the result of its “flufi'ed up” mahix and the large effective pore radius
associated with this state; both of which were produced by the soil’s dispersed clay
shucture. Porosity was more sensitive to additional loading in this mahix than in the
heated soil mahix (Figure 4) at loads that were less than 2 tsf; while at loads of 2 tsf or
more, the porosities of the two soils were essentially identical. This is consistent with the
observation of Mitchell (1993) who indicates that at low loads a soil with a dispersed clay
structure is more compressible than the same soil with a flocculated clay shucture. The
significance of the dispersed clay shucture to changes in conductivity with increased
loading was even more apparent in the effective pore radius data. These data (Figure 3)
showed that there was a higher rate of loss of effective pore radius with increased loading
in the unheated soil and that this occurred throughout the entire range of loads; even
throughout the range of loads where the porosities of the two soils were indistinguishable.
Identical porosities of heated and unheated soils at loads greater than 2 tsf showed that
consolidation proceeded independent of whether the soil’s clay was dispersed or

flocculated at high loads; i.e. clays accommodated rearrangement of granular particles in

36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4o "
E 3\\ \\
C
B7 Treated I \
.5 \
e i\
O
O.
36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28

 

 

0.1 1 10
Load (tsf)

Figure 4: Porosity of treated and untreated soils permeated
with water

37

the same manner regardless of their shucture. If one imagines the clay in the space
between larger grains at this point, the dispersed untreated clays expectedly sub-divided
this space to a much greater degree than did the flocculated heated clays. Consequently,
untreated soil exhibited smaller effective pore radius and lower conductivity than the
treated soil at high loads (Figures 3 and 2).

Soltrol saturated soils under load

With application of the first load (0.25 tsf), the Solhol conductivity of the
untreated soil was 20—fold higher than that of the heated soil and its higher conductivity
was sustained throughout the loading sequence (Figure 5). Each soil’s conductivity was
relatively insensitive to increased loading at loads below 2 tsf, and high loads were
required to produce even small decreases in either of these two soil’s conductivities.
These results conhast greatly with what was observed when the soil was water saturated
(Figure 2).

Most of the difference in Solhol conductivity between the treated soil and the
untreated soil was due to the difference between their effective pore radii; compared to
the heated soil, the unheated soil maintained a much larger effective pore radius
throughout the loading sequence (Figure 6). Higher porosity (Figure 7) made only a
minor conhibution (<7%) to the higher conductivity of the unheated soil.

The unheated soil’s sustained higher conductivity throughout the loading
sequence can be explained by the difference in conductivity that existed prior to loading
and the relative incompressibility of both mahices. Compared to the treated soil, the

greater Solhol conductivity (Figure 5) of the unheated soil at the lowest load (0.25 tsf)

38

1 E-01

Untreated

.3
3}
N

Hydraulic Conductivity, K (cmls)
a
b
00

 

1E-O4
0.1 1 10

Load (tsf)

Figure 5: Hydraulic conductivity of treated and untreated soils
permeated with Soltrol

39

1 E-05

Untreated

A

E

(cm’)

r.2

Effective Pore Radius,

A
m
b
‘1

 

1E-08
0.1 1 10

Load (tsf)

Figure 6: Effective pore radius of treated and untreated soils
permeated with Soltrol

40

50

 

 

 

 

 

[ Untreated

 

 

 

 

 

 

 

 

 

 

Porosity, n (96)

 

 

aa-Tfl " 'm \

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

32

 

0.1 1 10
Load (tsf)

Figure 7: Porosity of treated and untreated soils permeated
with Soltrol

4|

was associated with the “fluffed up” character of this mahix that existed prior to loading
(Table 3). Flocculated clays were responsible for the relative incompressibility of both
mahices. The flocculation of dispersed clays that resulted from saturating the unheated
mahix with Solhol produced the unheated soil’s incompressibility (Brown and Anderson,
1983; Foreman and Daniel, 1986; Bowders, 1991; Fernandez and Quigley, 1991). The
Solhol-saturated heated soil’s relative incompressibility was achieved prior to saturation
because the heated clays were flocculated when they were packed (Table 3).

[mum n0Ns FOR FIELD APPLICA HON

In ex-situ application of sorptive zone technology, soils down-gradient from the
contaminated region could be excavated, organically heated in a batch reactor, and
replaced. Feasibility of using such zones requires that the contaminated fluid pass through
the zone rather than around it. This, in turn, relies on the hydraulic conductivity of the
heated soil relative to the surrounding unheated soil.

It was shown that batch organomodification of a natural soil using the cationic
surfactant HDTMA served to flocculate its clays. A comparison of packed heated and
unheated soils showed that in the absence of an external load, the heated soil was denser
with lower conductivity to water. The effect of increased effective shess (loading) on
treated and unheated soils was found to be different. Upon loading, heated soils lost less
hydraulic conductivity than unheated soils such that heated soils had higher relative
conductivity at high loads. Treated soils were less compressible than unheated soils such
that at high loads the difference in their porosities was eliminated. Saturation of packed

heated and unheated soils with Solhol and subsequent loading showed that both soils

42

were relatively load insensitive. Substantially higher conductivity and porosity of
unheated soil throughout the loading, compared to those of the heated one, was athibuted
to flocculation of unheated clays (and dispersion of the heated ones) in Solhol.

The data suggested that ex-situ organomodified sorptive zone technology is
hydraulically feasible as evidenced by the relatively small difference in conductivity of
heated and unheated soils. Differences in hydraulic conductivity of heated and unheated
soils to water remained less than one order of magnitude throughout the loading
sequence. The somewhat small loss of conductivity due to heahnent at shallow depths, as
suggested by the data, could be accommodated by expanding the lateral dimensions of the
sorptive zone to account for contaminant plume spreading that will result. However, the
relatively greater resistance of heated soil to further loss of conductivity to water will
offset this effect at higher depths where it becomes more conductive. Higher conductivity
would effectively draw in flow lines from surrounding soils, and therefore increase the
region of influence or capture area of the sorptive zone. Exposure of the organomodified
sorptive zone to an LNAPL may cause a reduction in conductivity. As mentioned earlier,
such a loss of conductivity could be accommodated by expanding the lateral dimensions
of the sorption zone somewhat to account for pure contaminant plume spreading that will
result. However, caution should be exercised in drawing conclusions based exclusively
on these data as to field behavior of the soil. Conductivity changes may not occur to the
same extent in the field for reasons such as subsoils having water as the wetting phase or

being consolidated prior to encountering an LNAPL.

43
REFERENCES

Abdul, A. S., Gibson, T. L., and Rai, D. N., 1989. Laboratory Studies of the Flow of
Some Organic Solvents and Their Aqueous Solutions Through Bentonite and
Kaolinite Clays. Ground Water, 28(4):524.S33.

Ahlfeld, B. P., Dahmani, A., and Ji, W., 1994. A Conceptual Model of Field Behavior of
Air Sparging and Its Implications for Application. Ground Water Management and
Remediation, 132-139.

Allred, B., Brown, G. 0., 1994. Surfactant-Induced Reductions in Soil Hydraulic

I Conductivity. Ground Water Management and Remediation, 174-184.

Bowders, J. J ., Jr., 1985. The Influence of Various Concenhations of Organic Liquids on
the Hydraulic Conductivity of Compacted Clay. Geotechnical Eng. Dissertation
GT85-2, Geotech. Eng. Center, Civil Eng. Dept, U Texas, Austin, TX, 219 pp.

Bowders, J. J, Jr., 1991. Permeability of Clays Under Organic Permeants. Am. Soc. Civ.
Eng., Geotech. Eng., 117:1278-1280.

Boyd, S. A., Lee, J. F., and Mortland, M. M., 1988. Attenuating Organic Contaminant
Mobility by Soil Modification. Nature (London), 333:345-347.

Boyd, S. A., Jaynes, W. F., and Ross, B. S., 1991. Immobilization of Organic
Contaminants by Organo-clays: Application to Soil Restoration and Hazardous
Waste Containment. In R. A. Baker (ed.) Organic Substances and Sediments in
Water, Vol. 1, Publishers, Chelsea, MI., pp 181-200.

Broderick, G. P. and Daniel, D. E., 1990. Stabilizing Compacted Clay Against Chemical
Attack. Am. Soc. Civ. Eng., Geotech. Eng., 116:1549-1567.

Brown, K. W., 1988. Review and Evaluation of the Influence of Chemicals on the
Conductivity of Soil Clays. Project summary, EPA/600/SZ-88/016.

Brown, K. W. and Anderson, D. C., 1983. Effects of Organic Solvents on the
Permeability of Clay Soils. EPA-600/83-016, Municipal Environ. Res. Lab.,
Cincinnati, OH, 153 pp.

Burris, D. R. and Antworth, C. P., 1992. In Situ Modification of an Aquifer Material by a
Cationic Surfactant to Enhance Retardation of Organic Contaminants. J. Cont.
Hydrol., 10:325-337.

Corey, A. T., 1986. Mechanics of Immiscible Fluids in Porous Media, Water Resources
Publication, Littleton, CO., 255 pp.

Evans, J. C., 1991. Geotechnics of Hazardous Waste Conhol Systems. Chapter 20,
Foundation Engineering Handbook, 2nd ed., Ed. H. Y. Fang, Von Noshand
Reinhold Co., New York, N. Y.

Evans, J. C., Pancoski, S. E., and Alther, G. R., 1989. Organic Waste Treahnent with
Organically Modified Clays. Res. Dev. [Rep.] EPA/600/9-89/072, Int. Conf. New
Front. Hazard Waste Manage, 3rd, pp. 48-58.

Evans, J. C., Sambasivam, Y., and Zarlinski, S., 1990. Attenuating Materials in
Composite Liners. Waste Containment Systems: Conshuction, Regulation, and
Performance, R. Bonaparte, ed., Am. Soc. Civ. Eng., New York, N.Y., pp. 246-263.

44

Fernandez, F. and Quigley, R. M., 1991. Conholling the Deshuctive Effects of Clay-
Organic Liquid Interactions by Application of Effective Shesses. J. Canadian
Geotech., 28:388-398.

Foreman, D. E. and Daniel, D. E., 1986. Permeation of Compacted Clay with Organic
Chemicals. Am. Soc. Civ. Eng., Geotech. Eng., 112:669-681.

Freeze, R. A. and Cherry, J. A., 1979. Groundwater, Prentice-Hall, Inc., Englewood
Clifis, N. J. 604 pp.

Holtz, R. D. and Kovacs, W. D., 1981. An Introduction to Geotechnical Engineering,
Prentice-Hall, Inc., Englewood Cliffs, NJ. 733 pp.

Jaynes, W. F. and Boyd, S. A., 1991. Clay Mineral Type and Organic Compound

. Sorption by Hexadecylhimethylammonium-exchanged Clays. Soil Sci. Soc. Am. J .,
55:43-48.

Jordan, J. W., 1949. Alteration of the Properties of Bentonite by Reaction with Amines.
Mineralogical Magazine and Journal of the Mineralogical Society, 28(205)598-605.

Lee, J. F., Crum, J. R., and Boyd, S. A., 1989. Enhanced Retention of Organic
Contaminants by Soils Exchanged with Organic Cations. Env. Sci. Tech. 23:1365-
1372.

Mitchell, J. K., 1993. Fundamentals of Soil Behavior, 2nd Edition, John Wiley & Sons,
Inc., New York, 437 pp.

Shackelford, C. D. 1994. Waste-soil Interactions That Alter Hydraulic Conductivity.
Hydraulic Conductivity and Waste Contaminant Transport in Soil, ASTM STP
1142, David E. Daniel and Stephen J. Trautwein, Eds., Philadelphia, PA, pp. 111-
168.

Sherard, J. L., Dunnigan, L. P., and Talbot, J. R., 1984. Filters for Silts and Clays, Am.
Soc. Civ. Eng., Geotech. Eng., 110:397-414.

Sheng G., Xu, S., and Boyd, S. A., 19960. Mechanisms Conholling Sorption of Neuhal
Organic Contaminants by Surfactant Derived and Natural Organic Matter. Environ.
Sci. Tech. 30:1553-1557.

Sheng, G., Xu, 8., and Boyd, S. A., 1996b. Cosorption of Organic Contaminants from
Water by Hexadecylhimethylammonium—Exchanged Clays. Water Res. 30: 1483-
1489.

Smith, J. A. and Jaffe, P. R., 1994. Benzene Transport Through Landfill Liners
Containing Organophilic Bentonite. Am. Soc. Civ. Eng., Environ. Eng. 120:1559-

1577.

Smith, J. A., Franklin, P. M., and Jaffe, P. R., 1992. Hydraulic Conductivity of Landfill
Liners Containing Benzylhiethylammonium—bentonite. Am. Soc. Civ. Eng. proc.,
Water Forum ‘92, Environ. Eng. Session, Baltimore, MD., pp. 186-191.

Theng, B. K. G. 1974. The Chemishy of Clay Organic Reactions. John Willey & Sons,

New York.

Weiss

Wolfe, T. A., Demirel, T., Daumann, E. R., 1985. Interaction of Aliphatic Amines with
Monhnorillonite to Enhance Adsorption of Organic Pollutants. Clays Clay Miner.,
33 :301-31 1.

45

Xu, S. and Boyd, S. A., 1994. Cation Exchange Chemistry of Hexadecylhirnethyl
ammonium in a Subsoil Containing Vermiculite. Soil Sci. Soc. Am. J ., 58:1382-
1391.

Xu, S. and Boyd, S. A., 1995a. Cationic Surfactant Sorption to a Vermiculitic Subsoil via
Hydrophobic Bonding. Environ. Sci. Technol., 29:312-320.

Xu, S. and Boyd, S. A., 1995b. Cationic Surfactant Adsorption by Swelling and Non-
swelling Layer Silicates. Langmuir, 11:2508-2514.

Xu, S. and Boyd, S. A., 1995c. An Alternative Model for Cationic Surfactant Adsorption
by Layer Silicates. Environ. Sci. Tech., 29:3022-3028.

Xu, 8., Sheng, G., and Boyd. S. A. 1996. Use of Organoclays in Pollution Abatement.
Adv. Agron., 59:25-61.

Chapter 2
HYDRAULIC CHARACTERISTICS OF ORGAN OMODIFIED

SOILS FOR USE IN SORPTIVE ZONE APPLICATIONS

ABSTRACT

Modification of clay soils to increase their sorption capacity has been
demonshated in the laboratory and suggested as a potential means for preventing
migration of contaminated groundwater. The hydraulic feasibility of organomodified soils
for use in sorptive zone applications was studied by measuring hydraulic conductivity and
porosity of heated and unheated samples of a sandy loam with different clay contents as a
function of effective shess. Samples with 6, 12, 18, and 24% fine material were prepared
by washing the soil on a standard US #200 sieve and recombining different amounts of
the fines with the sand portion. Samples were then batch heated with hexadecylhimethyl
ammonium (HDTMA) to 80% of their cation exchange capacities. Upon initial packing
all heated samples were denser than unheated ones but their conductivities did not
exhibit a consistent hend. However, under the first load (24 kPa) all heated samples
showed higher conductivities and similar or higher porosities than unheated ones. Higher
loads decreased the difference in porosity between heated and unheated samples while
increasing the difference between their conductivities. At loadings of 383 kPa and higher,

all heated and unheated samples had similar porosities. At the highest load (766 kPa),

46

47

heated samples were 5 to 124 times more conductive than unheated ones depending on
their clay contents. Observed differences in conductivities were explained in terms of the
role of heated and unheated clays in conholling initial effective pore size and its change
during consolidation. Creation of organomodified sorptive zones by batch soil processing
is shown to be hydraulically feasible as evidenced by a marked increase in the
conductivities of all heated samples compared to unheated ones. In addition, an increase
in the sorptive capacity of the zone could potentially be achieved, without experiencing a
loss in conductivity, by increasing the clay content during modification.
Imoovcnozv

The limited sorptive capacity of clays for organic contaminants can be greatly
improved by replacing the naturally occurring inorganic cations with large organic
cations through ion-exchange (Mortland et al., 1986; Boyd et al., 1988b, c, 1991; Jaynes
and Boyd 1990, 1991, Sheng et al., 1996a, b). The enhanced sorptive capability of the
resultant organoclays suggests their utility in clay barrier technology (Smith et al., 1992),
for example as components of clay liners in hazardous waste disposal reservoirs and
peholeum tank farms. Organic cations can also be used to convert soil clays to
organoclays thereby enhancing the sorptive capabilities of soils, subsoils, and aquifer
materials (Boyd et al., 1988a, 1991; Lee et al., 1989; Wagner et al., 1994).
Organomodified soil has been suggested for application in sorptive zones which could
intercept and immobilize organic contaminant plumes (Boyd et al., 1988a, 1991; Burris
and Antworth, 1992; Xu et al. , 1996). A sorptive zone, coupled with subsequent

biodegradation of contaminants within the zone, could provide a comprehensive aquifer

48

remediation technology (Burris and Antworth, 1990; Nye et a1. , 1994; Crocher et al. ,
1995). Ex-situ modification is one method of placing such a zone for shallow
contaminated aquifers. This could be accomplished by excavating soil down gradient
from the contaminated region, organically heating the soil in a batch reactor, and placing
it back in the excavation. This method would allow addition of clay or sand, if needed, to
meet the desired sorption or conductivity requirements. It would also allow conhol of the
modification agent and heahnent level across the sorptive zone so that it best suits the
type and level of contamination. In the simplest case, a constant heahnent level
throughout the modified zone can be created; something which is more difficult to
achieve in an in-situ modification method.

Critical to the success of such sorptive zones is that organomodification not
significantly reduce the hydraulic conductivity of the soil. Replacement of native
exchangeable cations (e.g. Na“) on clays with large quaternary ammonium cations such as
hexadecylhimethylammonium (HDTMA) results in flocculation of clay particles.
However, over-saturation of the cation exchange capacity by organic surfactant molecules
has been shown to cause clay particles to disperse due to the build up of positive charge
on their surfaces (Xu and Boyd, 1994, 1995a, b, c). Clay flocculation and dispersion can .
be viewed as affecting particle size dishibution, which in turn, affects the consolidation
behavior of the soil (Holtz and Kovacs, 1981). More importantly, this affects pore size
and pore size dishibution which influences conductivity of the soil mahix (Bowders,
1985). In fact, clay flocculation and dispersion are the factors that alter the hydraulic

conductivity of clay soils most significantly (Shackelford, 1994).

49

By changing the clay surface from hydrophilic to organophilic,
organomodification also affects the ability of clays to intercalate water (J aynes and Boyd,
1990, 1991a, b). In natural clays water hydrates the interlayer cations, causing a
separation of clay platelets which is commonly referred to as swelling. In conhast,
organophilic clays may swell in the presence of organic liquids but not water (Jordan,
1949). Thus the hydraulic behavior of natural swelling clays (e.g. Na-bentonites) and
organically modified ones are fundamentally different. In the former case, swelling and
clay dispersion can be caused by water, and in the latter by organic liquids.

The sorptive capacity of organoclay sorbent zones is likely to be directly
proportional to, and limited by their organomodified clay content (Boyd et al. , 1988a;
Lee et a1. , 1989; Hatfield et al., 1992). Therefore, adding more clay to the natural soil
composition could be used to increase its sorptive capacity. However, the associated
changes in hydraulic properties of the zone should be clearly defined in order to establish
its hydraulic feasibility. Recently, Wallace et al. (1995) investigated the hydraulic
feasibility of sorptive zone applications based on organomodification technology by
employing an actual aquifer type soil (subsoil). They reported an initial decrease of 9-fold
in hydraulic conductivity for heated soil compared to the unheated one. However, heated
soils became progressively more conductive than the unheated ones with increased
efi‘ective shess. The authors athibuted this behavior mainly to the flocculation of heated
clays and its effect on the effective pore radius of the soil mahix. It was concluded that
the lateral dimensions of the sorptive zone could be increased to accommodate the lower

conductivity of the heated soil at shallow depths.

50

Considering the early stage of organomodification technology in sorptive zone
applications, it is apparent that there is a paucity of information about hydraulic behavior
of modified aquifer soils. More research is needed on soils with different compositions,
clay contents, and heahnent levels to begin establishing practical criteria on the use of
this technology. The objective of this work was to evaluate the hydraulic feasibility of ex-
situ modified sorptive zones by investigating the effects of initial compaction, and
subsequent loading on hydraulic conductivity of heated and unheated aquifer type soils
with difl‘erent fines contents. Conductivity changes are interpreted in the context of
microscale processes such as flocculation of clays which affect the pore sizes of the soil
mahix.

M1 MRI/(LS AND METHODS

An Oshtemo Bt-horizon soil (a coarse-loamy, mixed, mesic Typic Hapludalfs)
from Hickory Comets, MI, was sampled at a depth of 3 to 4 ft. The soil was air dried,
sieved through a US #20 standard sieve, and thoroughly mixed by hand. Sieve analysis
and hydrometer tests showed particle sizes to be 87% sand, 10.5% clay, and 2.5% silt.
The major exchangeable cations were 83% Ca”, 9% Mg", 5% K‘, and 3% Na“, as
determined by Michigan State University Soil Testing Laboratory. The predominant clay
minerals were vermiculite, illite, kaolinite, and hydroxy-aluminum interlayered
vermiculite, as determined by X—ray diffraction analysis.

In order to compose soil samples with difi‘erent fines contents fiom the same
natural components, the Oshtemo soil was thoroughly washed with water on a us #200

standard sieve and the passed fine fiaction collected and air dried. Dried fines were

51

manually pulverized with a mortar and pestle and sieved through a US #100 standard
sieve. They were then recombined with the clean, air dried sand-sized portion of the
Oshtemo soil so that 6%, 12%, 18%, and 24% respectively of their masses were fine
material. Recombination was accomplished by thoroughly mixing dry fines and sand and
then adding 4 to 10% water by weight with additional mixing. Samples were then air
dried, gently ground with a mortar and pestle, and sieved through a US #20 standard
sieve. These comprised the unheated samples.

Organomodification of the reconstituted soils was accomplished by adding
HDTMA-Cl (25% w/w in water) in an amount equivalent to 0.8 times the cation
exchange capacity (CEC), to a 20:1 (w/w) suspension of 1 mM NaCl and soil, followed
by stirring at 120 rpm for 1 minute, then at 20 rpm for 30 minutes. Treated soil was
settled out of suspension and the clear solution was decanted. Subsequently, the heated
soil was air dried, gently ground with a mortar and pestle, sieved through a US #20
standard sieve, and kept in an open pan (hereafter referred to as heated samples). For the
pin-poses of this work, the CEC was determined by the first plateau of the HDTMA
adsorption isotherm (Xu and Boyd, 1995a). This reference heahnent level was found to
be 27 mole kg", for the soil with 12% fines.

Turbidity of soil suspensions equilibrated with different amounts of HDTMA was
measured utilizing 25-ml Corex tubes containing 1:10 mixtures of soil and 1 mM NaCl,
to which appropriate volumes of HDTMA solution, corresponding to 0.01 to 3.0 times
the CEC (0.01 to 3 CBC), were added and immediately mixed. The tubes were then

shaken for 4 days on a rotating shaker and allowed to stand for 0.5 hr. Samples of the

52

suspension (1 ml) were collected from 1 cm below the surface, diluted with water to
turbidities within the range of 10-100 nephlomehic turbidity units (NTU), for turbidity
determination by nephelomehy. The procedure employed was consistent with that used
by Xu and Boyd (1995a) and in compliance with ASTM D1889.

A fixed ring consolidometer was used to perform falling-head conductivity
measurements while a vertical shess was applied to the sample. A load frame equipped
with a precision displacement hansducer (0.0001 inch; 0.00025 cm) was used to apply
different effective shesses and measure the thickness of the sample‘at each shess. Head
loss in the consolidometer channels, connections, porous stones, and filter papers was
assessed by assembling the device without any soil in it and measuring its conductivity
(K.,...) at different effective shesses. Treated and unheated soil samples were packed by
pouring 130 g air dried soil into the consolidometer ring; Whahnan #54 filter paper was
placed on both ends of the soil column. Samples were then tamped slightly to flatten the
upper surface and compacted gently with a drop hammer. The entire consolidometer was
placed into a desiccator and vacuum saturated over night with an aqueous solution of 1
mM NaCl.

Saturated samples were placed in the load flame and subjected to an effective
shess (loading) sequence horn O to 766 kPa using the conventional consolidation
procedure (ASTM D243 5-90). At each load hydraulic conductivity (K) was measured by
the falling head method (ASTM D5084-90) and the thickness of the sample was
measured. An aqueous solution of 1 mM NaCl in deionized, deaired water was used as

the permeant. Porosity was calculated hour the measured thickness of the consolidated

53

sample and hydraulic conductivity of the soil sample was calculated fi'om the
predetermined Kblink and three to five conductivity measurements at that load, adjusted to
20 °C (Daniel, 1989).
RESUL 1SAND DISCUSSION

Treatment and clay structure

The amount of HDTMA added to soils with different clay contents determined the
turbidity of their suspensions (Figure 8). In unheated (0 CBC) and over saturated (>2
CEC) samples, clay particles were dispersed as indicated by turbid suspensions. Clays
were flocculated at heahnent levels near 1 CBC. The general U shape curve relating
turbidity to HDTMA added is consistent with findings of Xu and Boyd (1994, 1995a)
who discussed the microscopic mechanisms underlying this behavior. A heahnent level
of 0.8 CEC was selected for conductivity measurements as representative of the region
where the clays are fully flocculated and the majority of cation exchange sites are
occupied by HDTMA. These measurements were then conhasted to those from unheated
soils, where the exchange sites occupied predominately by Na+ and the clays were
dispersed. The turbidity results were consistent with qualitative observations made during
conductivity experiments that effluent fiem the consolidometer reservoir was clear with
heated samples and cloudy with unheated ones.

Porosity

Prior to saturation, dry-packed unheated samples were 5 to 13% more porous than
heated ones, depending on their clay content (Table 4). Average porosities of both heated

and unheated samples increased with their clay content with the magnitude of the

54

 

P
a

 

P
a

 

P
g

 

 

 

Relative Turbidity

O
N
5'2"

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 05 1 15 2 25 3
1 .- .
T- '-
£?QB I
Ens
P I
in
¢0.2 12%
O#*+.—va..l..r.
O 05 1 t5 2 25 3

Relative'l'tlbldlty

 

 

24%
0 05 1 15 2 25 3
mamaddadtoeoknanaludtotiu
CEC

Figure 8: Turbidity of soils with different fine contents

55

Egfiuztgpfiaer

on. 3 .o .8. .o canes. 88.28 .
ewe. Eon .u 3.3.3 3.833»

30. Ben E 5:23am 8 6.... ..

 

 

o. :8 8.3 :8 «one :8 «53 to 3.8 8.3 8.8 «o. .m 8.8 3.8 3
N. :8 3.3 3.8 8...... 3.8 3.3 no 88 8.3 .~.8 3.3 3.8 «3... a.
m... :8 3. 3 a8 8.3 :8 and. no :8 8.8 :8 «N3 88 8.3 N.
a... 8.8 3.9. 8.8 no. 3 3.8 «m. 3 «N 8.8 3.3 8.8 3.3 6.8 8.3 o
.4 .82 a... 8.9.28 .888 >5 .4 88. .8“. .3238 .888 an. 38 .828

 

ea. 2.8 8.8.. a. ...

much.

 

e8 2.8 8.8.2: 5 ...

8o. .2: o... .o

5:3..an 0. 0.... 980.60.. 3.3.2. new .32 .m... on. fine. 23 .m 8.3 .o emu. a 5.3.2. @9202 3 can...

56

increase being greater in the untreated samples. As the fine content increased fi'om 6% to
24%, porosity of untreated samples increased by 6.1%, while this increase was only 3.4%
in treated ones. The variation in porosity with clay content, especially in untreated
samples, highlights the role of the clay and its structure in the initial assemblage of the
soil matrices. Saturation with water increased the porosity of treated and untreated
samples less than 3%. Untreated samples remained more porous than treated ones afier
saturation.

With application of the first load a decrease in porosity was observed that was
much higher in untreated samples compared to the treated ones (Table 4). This decrease
for the untreated soils was approximately proportional to the clay content. This resulted in
the untreated samples having porosities approximately equal to or less than the treated
ones at the first load (24 kPa).

Clay particles may reside in the pore space created by sand grains, or in between
sand grains, in which case they are part of the load bearing mechanism of the matrix
(Kenney et al., 1992). It appears that in the untreated samples, dispersed clays resided
between the sand grains, especially at high clay contents. Furthermore, the clay platelets
may be poorly ordered relative to one another (e. g. in an edge-to-face arrangement). This
resulted in a relatively unstable but more porous matrix at zero load. Application of the
first load collapsed this unstable structure, resulting in a relatively large consolidation of
the soil matrix. This trend was more profound in samples with higher clay contents.
Treated samples were initially not as porous as untreated samples, apparently due to the

highly ordered, parallel structure of the flocculated clays. As a result, the first load only

57

slightly decreased the porosity; far less than the magnitude of the drop in their untreated
counterparts.

Subsequent load applications, following the first, produced additional
consolidation with no observable abrupt changes (Figure 9). In the 6% fine soil (Figure
9a), untreated samples were more porous on average, but relatively high variations
between samples were observed in both the treated and untreated case. Soils with higher
clay contents exhibited less variation and heated samples were generally more porous
than untreated samples except at the highest loads where they were similar (Figure 9b—d).
At low loads, unheated samples were less porous, and somewhat less compressible than
the treated ones. Apparently, the dispersed clays in the untreated soil facilitated
rearrangement of particles under a relatively low load causing the matrix to collapse and
lose more porosity than was the case with the heated soil matrix. Higher loads were then
sustained more effectively by the unheated soil matrix which had limited space for
further consolidation. Treated soils had more stable structures at zero load. The
magnitudes of their porosity loss due to the first load were comparable to their
compressibility throughout the loading sequence. This means that their flocculated clay
structure resulted in a relatively dense and stable initial pack which consolidated
gradually as the loading increased. Therefore, flocculated clays may have actually been
an essential part of the load bearing mechanism of the treated soil samples, thereby
producing higher porosities compared to unheated samples at low loads.

Differences in the porosities of treated and unheated samples decreased with load

such that nearly identical porosities were observed for the two materials at high loads.

58

Load (kPa)

1 000

 

 

 

 

 

 

 

 

 

 

 

 

 

10 100

44 i
.... A lUntreatedl 6%
2‘5 42 A
g .0 z
e [Treated
E. 38

36

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Porosity (96)

Porosity ($6)

Porosity (%)

10

Untreated

Untreated

Untreated

1 00
Load (kPa)

 

1000

Figure 9: Porosity of soils with 6%, 12%, 18%, and 24%

added fines at different loads

59

This similarity suggests that consolidation was ultimately conh'olled by the larger soil
particles which were identical in the two materials. At such high loads, clays evidently
accommodated rearrangement of granular particles, and hence consolidation, in the same
manner regardless of their shucture.

Fines content influenced the porosity of unheated and heated samples (Figure 10).
Unheated samples with 6% fine material were the most porous throughout the loading
sequence (Figure IO-a). They showed very low compressibility which is similar to the
response of a sand matrix. Apparently, sand particles were the main load-bearing fraction
in these samples. All samples with higher clay content were more compressible and in all
cases the compressibility increased with increasing clay content. This suggests that higher
clay content facilitated rearrangement of the soil particles. This behavior is also
consistent with the fact that in general, clay soils are more compressible than sands
(Freeze and Cherry, 1979).

Porosity of heated samples with difl‘erent clay contents reveals that at low loads,
average porosities of heated samples with 6 and 12% fines were similar (Figure lO-b).
Porosities of samples with 18 and 24% fines were similar as well, but larger than those of
the first group. This observation suggests that the load bearing mechanism was different
in the two groups.

Conductiviq

Conductivities of heated and unheated samples at zero load did not show a
consistent trend (Table 5). However, application of the first load (24 kPa) resulted in a

substantial decrease in the conductivity of each of the untreated soils, and a minor

Porosity We)

Porosity ('56)

60

Load (kPa)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 100 1000
46
44 I21
42 f
\ *-~
40
K N”
38 X
\\N
36 4 +696
34 d__ +12%
+18% \ "
32 -— +24%
so i
46
‘4 \g: 15%
42 \
I‘ \§
40 i:
38 N *x
36
+6% \NJ
34 +12% 5.
+18% \\
32 +24% \
30
1O LoadikPa) 100 1000

Figure 10: Average porosity of a) unheated and b) heated
soils with 6, 12, 18, 24% added fines

61

gggiv. \gggm

an... «a .0 82 .a 83...... 8.828 .

 

 

ewe. Sou 6 3353 3.82am .
3.. :8... a. E... :8... ”6...... we... .88.... “3...... 88.... 8...... cu
m. 38.... «8o... 88.... «m8... 2.. $88.... 8...... 88.... «8.... m.
3 :8... «m8... :8... «NS... 3.. 88...... “$8... 38.... «8.... a.
N. 68.... 3...... 38.... «8.... m... :83. ”5...... $8.... #8... o
.4 .82 a... .3828 a. .82 a... 8.928 e... .828 8c...

 

.7» Eu. =3 36...... c. X

 

Ave E... m_.om 352...: c. X

28. .2. o... .0

cos—8.3.5 0. one co... m.. new .23. .2: o... .32 Son .m 2.3 .0 1m... v. 5.2.39.8 0.35.3... emu.e>< ”m 2.3..

62

decrease for each of the treated soils such that all untreated soils were less conductive
than treated ones at that load. The drop in conductivities, with the initial load, depended
on the clay content in both treated and untreated samples. These results are consistent
with the observed decreases in porosity, so it is constructive to evaluate the extent to
which these two parameters are linked and what other characteristics of the soil matrix
might affect conductivity.

Hydraulic conductivity is a function of the soil matrix property, intrinsic

permeability (k), and the fluid properties of viscosity (u) and density (p);

K=k£§
p

To eliminate the possibility that the soil treatment altered fluid properties as the water
percolated through treated and untreated soils, viscosity and density of permeant water
from both materials was measured. Since less than a 1% difference was found in these
properties, it was concluded that conductivity difl'erences resulted exclusively from
changes in the soil permeabilities. The analogy of Poiseuille flow through a bundle of
capillary tubes can be used to describe permeability as a function of porosity (n) and an

effective pore radius (re) (Allred and Brown, 1994);

k = m;2
Using this relationship, the differences in porosities accounted for less than 17% of the
differences in conductivities of treated vs untreated soils at low loadings. We therefore
conclude that change in effective pore radius was the primary cause of the conductivity

difl‘erence observed in these soils.

63

The untreated soil formed a more porous matrix upon packing which was very
sensitive to external load in terms of its effective pore radius. Application of the first load
collapsed the matrix and apparently forced the clay particles from between the sand
grains into the interior regions of pores. This had the effect of partitioning the pores into
smaller sections (i.e. smaller pores) and lowered conductivity. Higher clay contents
initially produced a more porous matrix, and they resulted in higher drop in conductivity
with application of the first load. Application of the first load also dropped conductivities
of the treated samples, but to less of an extent. Flocculated clays in treated samples,
forming bigger aggregates, evidently were less easily forced into the interstitial spaces
and were less able to partition the pore space into small sections, manifesting smaller
drops in conductivity.

Changes in conductivities of treated and untreated soils with increasing load are
shown in Figure 11. Solid lines on the graphs represent two standard deviations on either
side of the data, established by the maximum of either the variance in the averages, or the
error propagated from estimated measurement uncertainties. It can be observed that all
treated samples had higher conductivities than untreated ones throughout the loading
sequence. These differences were sustained over the range of loads employed with similar
systematic decreases in conductivity with increased load for both types of soils. Treated
samples were 3 to 129 times more conductive depending on load and clay content.

At these higher loads, differences in porosities accounted for less that 11% of
observed differences in conductivities between treated and untreated soils, again

suggesting that treatment primarily alters the sizes of conductive pores, and that this is

Hyd. Conductivity (cmls) l-iyd. Conductivity (cmls)

Hyd. Conductivity (cmls)

Hyd. Conductivity (cmlc)

64
10 Load (kPa) 100 1000

Treated

1E+O
1E-1
1E-2

1 E‘3 Untreated
1 E-4

1 ES
1 E-6
1E+o
1 E-1
1 E-2
1 E-3
1 E-4
1 E5
1 E-6
1 BO
1 E-1
1 E-2
1 E-3
1 E-4 Untreated
1 B5
1 E-6

11:-j [Treated] HEEL]

15.2 : -

1E-3 O: E';

1E'4 mintreated
1E-5 a; -

\.
1E-6 1-
10 Load (kPa) 100 1000

Figure 11: Hydraulic conductivity of soils with different
lines contents

 

 

 

 

 

 

 

 

 

 

//

i‘
1
Ah]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

65

applicable to the full range of the loading sequence. It is interesting to note that at high
loads, where treated and untreated samples had identical porosities (Figure 9),
conductivities were substantially different. Because the clays were occupying the same
total volume in both treated and untreated samples, difi‘erences must result from large
significant difi‘erences in pore geometry. The larger pores sustained by the flocculated
treated clays produced conductivities up to 129 times higher.

Loss of conductivity with consolidation for untreated soils was directly related to
the amount of fine material (Figure 12-a). This is entirely consistent with the common
observation that higher clay contents correspond to lower conductivity and higher
compressibility. Higher amounts of clays partition the pore space to smaller sections and
facilitate easier rearrangement of particles. In contrast, the treated soils showed more
complex behavior as clay content increased (Figure 12-b). The 6% fines sample exhibited
high conductivity and low compressibility. With 12% fines, the conductivity at low loads
dropped substantially and slightly higher compressibility was observed. At 18%, the
initial conductivity was greater than the 12% material, nearly equaling the 6%, but a high
compressibility was observed. Finally, the 24% material had both low initial
conductivity, similar to the 12% material, and substantial consolidation due to the loading
sequence.

Previous studies have established that in Na-bentonite sand mixtures, as the clay
content increases, it first fills the pore space created by sand grains and then resides
between the grains (Kenney et al., 1992). At a certain “cut off” clay content, when clays

completely fill the pore space created by sand grains, a very low conductivity is achieved.

66

Load (kPa)

10 100 1000

1 B1

 

 

~1E-2
Q
E
2..
3‘ -
s 153
3
0
3
‘8
o 1E-4
O
15
>
I
1E-5
1E-6
1E-1
‘3
E
315-2
E‘
>
3
0
3
'U
C
O
3 15.3
>. 116%
3‘ n12%
918%
024%
1E-4

1 0 Load (kPa) 100 1 000

Figure 12: Hydraulic conductivity of a) untreated, and
b) treated soils with different fines contents

67

The minimal additional reduction in hydraulic conductivity caused by increasing the clay
content beyond that point has been used to infer the cut off clay content (Daniel, 1987;
Kenney et al., 1992).

These concepts can be used to help explain the porosity and conductivity data for
the treated samples (Figure lO-b and 12-b). In the 6% fine samples there was not enough
clay to fill the void space created by the sand grains of the matrix. High conductivities
were observed because the flocculated nature of the clays left relatively largeiconductive
pores and only minimal consolidation occurred because the load was supported primarily
by the sand matrix. At the 12% fines level, the fines filled this pore space and probably
some additional portion resided between the sand grains. In this arrangement, with the
fines mostly residing in the pore space created by sand grains, porosities similar to those
of 6% fine samples were obtained (Figure lO-b) but conductivities were much lower
(Figure 12-b). The increase in compressibility presumably resulted from the clays
between the sand grains which accommodated easier rearrangement of sand particles with
increased loading. The additional clay in the 18% fine samples was apparently placed
between some sand grains, creating new large pores, and hence increasing both initial
porosity (Figure lO-b) and conductivity. However, with the clays in between sand grains,
the sample was more sensitive to increased loading in terms of loss of porosity and
conductivity. Finally, at a level of 24% fine material, the new large pores created in 18%
fine samples, were filled with the additional fines whereby the porosity remained similar
to the 18% samples, but the conductivity was lowered substantially. Similar high

sensitivity of 18 and 24% fine samples to increased loading in terms of their porosity and

68

conductivity indicates that in both samples clays were dominating the response of the
matrix to load.

All hydraulic conductivities of samples with different clay contents increased as a
result of treatment with HDTMA (Figure 13-a). Ratios of conductivity in treated and
unheated soils W reveal that treatment produced a much greater increase in
conductivity for the 18 and 24% fine samples than for the 6 and 12% fine samples. This
provides another evidence that pore geometry, which controlled the conductivity, was
different between the two groups. In samples with 6 and 12% fines, the clays were mostly
residing in pore spaces created by the sand grains, and treatment increased their
conductivity by flocculating the dispersed clays and “opening up” these pore spaces. In
samples with 18 and 24% fines, clays were not only filling these pore spaces, but also
residing in between sand grains. Here, treatment opened up larger pores in both spaces,
which in turn, increased conductivity substantially.

IMPLICA HONS FOR FIELD APPLICA norv

In ex-situ applications of sorptive zone technology, soils down-gradient from the
contaminated region could be excavated, organomodified in a batch reactor, and replaced.
Feasibility of using such zones requires that the contaminated fluid pass through the zone
rather than around it. This, in turn, relies on the hydraulic conductivity of the treated soil
relative to the surrounding untreated soil. Furthermore, in ex-situ modification
technology, addition of clay to the natural soil in order to increase the net sorption

capacity can be considered as long as the zone remains hydraulically feasible.

69

 

10 “a“ (”’3’ 100 1000
1000
A 6%
I: 12%
e 18%
o 24%
1 00
x?
E?
1 0
1
1 E1
1 E-2
‘u?
'2
.9.
3. 1E-3
E
E
'0 0
1E-4 A 6 ,6 Treated
S o 12% Treated
0, e 18% Treated
'2, o 24% Treated
I 1E-5 A 6% Untreated
I 12% Untreated
9 18% Untreated
0 24% Untreated
1 E-6

10 Load "(9,, 100 1000

Figure 13: Comparison of average hydraulic conductivities
(K); a) ratio of treated and untreated soils, and
b) all samples‘

70

The data demonstrated that ex-situ organomodified sorptive zone technology was
hydraulically feasible as evidenced by a marked increase in conductivity of all treated
samples compared to untreated ones. This increase varied fi'om 3 to 129-fold depending
on the clay content and the load (Figure 13-a). In the design of sorptive zones developed
by batch treatment, such increases of conductivity could effectively draw in flow lines
from surrounding soil, and therefore increase the region of influence or capture area of
the sorptive zone. It is important to note that the baseline used for comparison was soil
permeated with 1 mM NaCl which would promote clay dispersion. Thus, the increase in
conductivity associated with HDTMA treatment may represent maximal values. In
addition, the data represent the increase in conductivity of recombined samples which
may have different matrices compared to natural soils. Therefore, for field applications, a
similar lab study of the particular soil under consideration is necessary prior to feasibility
determination of the technology. Nonetheless, the results suggest that there are no
fundamental hydraulic barriers to ex-situ soil modification using cationic surfactants.

Superposition of average conductivities of treated and untreated samples with
different clay contents shows that in general, treated samples with higher clay contents
had comparable conductivities to untreated samples with lower clay content (Figure 13-
b). This suggests that it may be possible to increase the sorptive capacity of a zone by
increasing its clay content prior to modification, without lowering its conductivity. In
fact, the data suggests that at shallow depths it would be possible to quadruple the

sorptive capacity by quadrupling the clay content of a 6% soil without lowering its

71

conductivity. At higher depths doubling the clay content of a 12% clay soil would not
lower its conductivity while this would double its sorptive capacity.
REFERENCES

Alther G. R., 1987. The qualifications of bentonite as a soil sealant. Eng. Geology,
23:177-191.

Allred, B., Brown, G. 0., 1994. Surfactant-induced reductions in soil hydraulic
conductivity. Ground Water Management and Remediation, 174-184.

Burris, D. R. and Antworth, C. P., 1992. In situ modification of an aquifer material by a
cationic surfactant to enhance retardation of organic contaminants. J. Cont.
Hydrol., 10:325-337.

Bowders, J. J ., Jr., 1985. The influence of various concentrations of organic liquids on the
hydraulic conductivity of compacted clay. Geotechnical Eng. Dissertation GT85-
2, Geotech. Eng. Center, Civil Eng. Dept, U Texas, Austin, TX, 219 pp.

Boyd, S. A., Lee, J. F., and Mortland, M. M., 19880. Attenuating organic contaminant
mobility by soil modification. Nature (London), 333:345-347.

Boyd, S. A., Mortland, M. M., and Chiou, C. T., 1988b. Sorption characteristics of
organic compounds on hexadecylhimethylammonium-media. Soil Sci. Soc.
Am. J. 52:652-657.

Boyd, S. A., Shaobai, S., Lee, J. F., and Mortland, M. M., 19886. Pentachlorophenol
sorption by organoclays. Clays and Clay Minerals, 36: 125-130.

Boyd, S. A., Jaynes, W. F., and Ross, B. S., 1992. Immobilization of organic
contaminants by organo-clays: Application to soil restoration and hazardous waste
containment. In R. A. Baker (ed.) Organic Substances and Sediments in Water,
Vol. 1, Publishers, Chelsea, ML, pp 181-200.

Corey, A. T., 1986. Mechanics of immiscible fluids in porous media, Water Resources
Publication, Littleton, CO., 255 pp.

Crochet, F. H., Guerin, W. F ., Boyd, S. A., 1995. Bioavailability of naphthalene sorbed to
cationic surfactant-modified clay. Environ. Sci. Tech. 29:2953-2958.

Daniel, D. E., 1987. Earthen liners for land disposal facilities. Proc. Geotech. Practice for
Waste Disposal, ASCE, New York, N. Y.

Daniel, D. B., 1989. A note on falling headwater and rising tailwater permeability tests.
Geotechnical Testing J ., ASTM, 12(4)308-310.

Evans, J. C., Pancoski, S. E., and Alther, G. R., 1989. Organic waste treatment with
organically modified clays. Res. Dev. [Rep.] EPA/600/9-89/072, Int. Conf. New
Front. Hazard Waste Manage, 3rd, pp. 48-58.

Freeze, R. A. and Cherry, J. A., 1979. Groundwater, Prentice-Hall, Inc., Englewood
Cliffs, N. J. 604 pp.

Hatfield, K., Burris D. R., Staufi‘er T. B., and Ziegler J., 1992. Theory and experiments
on subsurface centaminant sorption systems. Am. Soc. Civ. Eng., Environ. Eng. J.
118:322-337.

72

Holtz, R D. and Kovacs, W. D., 1981. An introduction to geotechnical engineering,
Prentice-Hall, Inc., Englewood Cliffs, NJ. 733 pp.

Jaynes, W. F. and Boyd, S. A., 1990. Trimethylphenylammonium smectite as an effective
adsorbent of water soluble aromatic hydrocarbons. J. Am. Waste Manag. Assoc.
40:1649-1653.

Jaynes, W. F. and Boyd, S. A., 1991a. Clay mineral type and organic compound sorption
by hexadecylhimethylammonium-exchanged clays. Soil Sci. Soc. Amer. J. 55:43-
48.

Jaynes, W. F. and Boyd, S. A., 1991b. Hydrophobicity of siloxane surfaces in smectites
as revealed by aromatic hydrocarbon adsorption from water. Clays and Clay
Minerals 39:428-436.

Jordan, J. W., 1949. Alteration of the properties of bentonite by reaction with amines.
Mineralogical Magazine and Journal of the Mineralogical Society, 28(205)598-
605.

Kenney T. C., Van Veen W. A., Swallow M. A., and Sungaila, M. A., 1992. Hydraulic
conductivity of compacted bentonite-sand mixtures. Canadian Geotech. J. 29:364-
374.

Mortland, M. M., 1970. Clay-organic complexes and interactions. Advances in
Agronomy, Academic Press, Inc, 22: 75- 117.

Mitchell, J. K., 1993. Fundamentals of soil behavior, 2nd Edition, John Wiley & Sons,
Inc., New York, 437 pp.

Nye, J. V. Guerin, W. F., Boyd, S. A., 1994. Heterotrophic activity of microorganisms in
soils treated with quaternary ammonium cations. Environ. Sci. Tech. 28:941-944.

Raussell-Colom, J. A., and Serratosa, J. M., 1987. Reactions of clays with organic
substances. in Chemistry of Clays and Clay Minerals, ed. Newman A. C. D., pp.
371-422, John Wiley & Sons, New York.

Shackelford, C. D., 1994. Waste-soil interactions that alter hydraulic conductivity.
Hydraulic Conductivity and Waste Contaminant Transport in Soil, ASTM STP
1142, David E. Daniel and Stephen J. Trautwein, Eds., Philadelphia, PA, pp. 111-
1 68.

Sheng G., Xu, S., and Boyd, S. A., 19960. Mechanisms controlling sorption of neutral
organic contaminants by surfactant derived and natural organic matter. Environ.
Sci. Tech. 30:1553-1557.

Sheng, G., Xu, S., and Boyd, S. A., 1996b. Cosorption of organic contaminants from
water by hexadecylhimethylammonium-exchanged clays. Water Res. 30: 1483-
1489.

Smith, J. A. and Jaffe, P. R., 1994. Benzene transport through landfill liners containing
organophilic bentonite. Am. Soc. Civ. Eng., Environ. Eng. 120:1559-1577.

Smith, J. A., Franklin, P. M., and Jafl‘e, P. R., 1992. Hydraulic conductivity of landfill
liners containing benzyltriethylammonium-bentonite. Am. Soc. Civ. Eng. proc.,
Water Forum ‘92, Environ. Eng. Session, Baltimore, MD., pp. 186-191.

Theng, B. K. G. 1974. The chemistry of clay organic reactions. John Willey & Sons, New
York.

73

Wanger, J. H., Chen, H., Brownawell, B. J ., and Westall, J. C., 1994. Use of cationic
surfactants to modify soil surfaces to promote sorption and retard migration of
hydrophobic organic compounds. Environ. Sci. Tech. 28:231-237.

Wallace, R B., Grant J. M., Voice T. C., Rakhshandehroo G. R., Xu S., and Boyd S. A.,
1995. Hydraulic conductivity of organomodified soils. Proceedings of 1995
National Conference on Environmental Engineering.

Xu, S. and Boyd, S. A., 1994. Cation exchange chemistry of
hexadecylhimethylammonium in a subsoil containing vermiculite. Soil Sci. Soc.
Am. J., 58:1382-1391.

Xu, S. and Boyd, S. A., 19950. Cationic surfactant sorption to a vermiculitic subsoil via
hydrophobic bonding. Environ. Sci. Tech., 29:312-320.

Xu, S. and Boyd, S. A., 1995b. Cationic surfactant adsorption by swelling and non-
swelling layer silicates. Langmuir, 11:2508-2514.

Xu, S. and Boyd, S. A., 1995c. An alternative model for cationic surfactant adsorption by
layer silicates. Environ. Sci. Tech., 29:3022-3028.

Xu, 8., Sheng, G., and Boyd. S. A., 1996. Use of organoclays in pollution abatement.
Adv. Agron., 59:25-61.

Chapter 3

HYDRAULIC FEASIBILITY OF IN-SI TU MODIFIED SORPTIVE
ZONES; HYDRAULIC CONDUCTIVITY OF SOILS
ORGANOMODIFIED TO DIFFERENT LEVELS IN A BATCH

SYSTEM

Answer

Hydraulic feasibility of in-situ organomodification of soils for sorptive zone
applications was studied by measuring hydraulic conductivity of a sandy loam with
different clay contents, batch treated to 0.1 (low-treated) and 2.2 (high-treated) times its
cation exchange capacity (CBC), and subjected to different effective stresses. The results
were viewed in combination with our previous study on the same soil treated to 0
(untreated) and 0.8 (mid-treated) times the CEC in order to provide a full spectrum of
treatment level consideration. Samples with 6, 12, 18, and 24% fines were prepared by
washing the soil on a Standard US #200 sieve and recombining different amounts of the
fines with the sand portion. Samples were treated with hexadecylhimethyl ammonium
(HDTMA) in a batch system and subsequently air dried and sieved through a US #20.
Upon initial dry packing, all high-treated samples were denser than low-treated ones.

Saturation with water induced less than 3% change in their porosities while their

74

75

conductivities didn’t exhibit a consistent trend. However, under the first load (0.25 tsf) all
high-treated samples showed higher conductivities with similar or higher porosities than
low-treated ones. Higher loads decreased the difference in porosity of high- and low-
treated samples while increased the difference between their conductivities. Turbidity of
suspensions from high- and low-treated soils was used to infer the clay structure in our
soils. Observed differences in conductivities were explained in terms of the role of clay
structure in controlling initial effective pore radius and its change during consolidation.
Treatment of soil to levels above its CEC, which can occur in creation of in-situ
organomodified sorptive zones is hydraulically manageable as evidenced by higher
hydraulic conductivity of high-treated samples compared to low-treated ones.
Imoovcnorv

It has been shown that large organic cations can be used to replace the metal
cations associated with the clay matrix of a soil in order to increase the limited sorptive
capacity for organic contaminants (Boyd et al., 1988a). Previous studies have suggested
that this increased sorptive capacity of organomodified soils could be managed to create a
sorptive zone which immobilizes the dissolved organic chemicals in contaminated ground
waters (Burris and Antworth, 1992; Wallace et al., 1995; Xu and Boyd, 1994, 1995). A
sorptive zone, coupled with subsequent bioremediation of contaminants within the zone,
could provide a comprehensive aquifer remediation scheme (Burris and Antworth 1990).
Hydraulic feasibility of ex-situ modified sorptive zones has been established by
measurement of their higher conductivity compared to untreated surrounding

(Rakhshandehroo et al., submitted).

76

Underground injection of an aqueous solution of organic cations through a set of
injection wells has been proposed to create an in-situ organomodified sorptive zone
(Boyd et al., 1988a). Hexadecylhimethyl ammonium (HDTMA) was identified as one of
the most effective cations for this purpose due to its excellent cation exchange properties,
effectiveness in modifying a variety of clay mineral types, and high sorptive capacity for
many organic contaminants (Lee et a1. ,1989; Boyd et al., 1988b,c; Jaynes and Boyd,
1990, 1991). Burris and Antworth (1992) simulated in-situ injection of HDTMA to their
soil columns and box model aquifer material and reported that effective sorptive zones
were created.

In-situ injection will likely result in a non-uniform distribution of the treatment
(Xu and Boyd, 1995). It may occur vertically due to the different overburden loads on the
soil at different depths which can result in different pore geometries, pore velocity of
treatment solution, and accessibility of sites to surfactant molecules. It may occur radially
fi'om the point of injection due to surfactant consumption with distance from the injection
point as well as changes in velocity. Field heterogeneity may also be responsible for
producing a non-uniform treatment level in an in-situ modification process. In general,
one may imagine that in regions far away fi'om point of injection the treatment level
correspond to less than CEC of the soil, and in regions close to the point of injection the
treatment would exceed the CEC of the soil. Burris and Antworth (1992) injected
HDTMA to their column and box model aquifer soils and found that such treatment
profiles were in fact created in the in-situ injection of the surfactant under their

experimental conditions.

77

Organomodification of clays to different levels has a major effect on their
flocculated or dispersed structure (Theng, 1974). Xu and Boyd (1994, 1995) found that in
originally dispersed soils, heahnent with HDTMA to levels far below and far above the
CEC resulted in clay dispersion, while heahnent to levels slightly below the CEC
produced flocculation in clays. The mechanisms of dispersion at low and high ends
however are quite different. In high-heated soils, excess HDTMA is hydrophobically
bonded on the clay surfaces, creating a net positive hydrostatic charge that causes
dispersion. On the other hand, dispersion in low-heated clays is due to the net negative
charge on their surfaces from unneuhalized exchange sites (Xu and Boyd, 1995). In any
case, clay dispersion and flocculation are demonstrates to significantly affect hydraulic
conductivity of soils (Shackelford, 1994).

While Burris and Antworth (1992) reported no apparent overall change in
hydraulic conductivity of their in-situ modified columns or box model aquifer, their
apparatus was incapable of separately measuring conductivity of sections heated to
different levels. Despite the fact that a typical column study with the soil packed and
heated in-place has practical appeal, a simpler experiment serving as the first indicator of
basic effects is more appropriate, considering the early stage of our understanding of this
relatively new material (i.e. organomodified soils). Soils may be heated to different levels
in a batch system and their conductivities measured under different shesses. Batch
experiments require less technical development and appear to have the advantages of

being simple, easily reproducible, and insightful. To our knowledge, no specific study of

78

the effects of heahnent level on hydraulic conductivity of soils has been previously
reported.

The effects of batch heahnent on hydraulic conductivity of soils with different
clay contents was studied by authors in a previous paper. The focus of that work was the
ex-situ application of sorptive zone technology, therefore a heahnent level of 0.8 times
the CEC was employed. It was shown that the heatnrent increases hydraulic conductivity
of all soil samples under the applied range of effective shesses and concluded that the
technology is hydraulically feasible (Rakhshandehroo et al., submitted).

In the present study, which focuses on the in-situ application of the technology,
authors wish to extend this understanding to soil samples heated to different levels. The
objectives of the study were to evaluate the hydraulic feasibility of in-situ heated sorptive
zones by (i) investigating the changes in conductivity of soils with different clay contents
induced by batch heahnent to levels far below and far above the CEC, (ii) monitoring this
change under different effective shesses and seeking its correlation with the microscale
processes such as clay dispersion (or flocculation) in determining the pore geomehy of
the soil mahix, and (iii) incorporating these results with those fi'om the previous study to
assess the feasibility of field application of the in-situ injection technique.

MA namrs AND METHODS

A sandy loam B-horizon soil (Oshtemo series; a coarse-loamy, mixed, mesic
Typic Hapludalfs from Hickory Corners, MI) from the depth of 34 ft was obtained for
these experiments. The soil was air dried, sieved through a US standard #20 sieve, and

carefully blended until a homogenized mixture was obtained. Particle size analysis

79

indicated the sample was 87% sand, 10.5% clay, and 2.5% silt based on sieve analysis
and hydrometer tests. Quantity and type of exchangeable cations were 83% Ca“, 9%
Mg", 5% K‘, and 3% Na‘, as determined by Michigan State University Soil Testing
Laboratory. The predominant clay minerals were vermiculite, illite, kaolinite, and
hydroxy-aluminum interlayered vermiculite, as determined by X-ray diffraction.

In order to conshuct soil samples with different clay contents from the same
natural components, the Oshtemo soil was thoroughly washed on a US standard #200
sieve and the passed fine fraction was collected and air dried. Dried fines were manually
pulverized with a mortar and pestle and sieved through a US standard sieve #100. They
were then recombined with the clean, air-dried sand portion of the Oshtemo soil at levels
of 6%, 12%, 18%, and 24% fine material. Recombination was accomplished by
thoroughly mixing dry fines and sand and then adding water (4 to 10% wat) with
additional mixing. Samples were then air dried, gently ground with a mortar and pestle,
and sieved through a US standard #20 sieve.

Organomodification of the reconstituted soils was accomplished by adding
HDTMA chloride (25% wt/wt in water) in an amount equivalent to 0.1 and 2.2 times the
CEC, to a 20:1 (wt/wt) suspension of 1 mM NaCl and soil, followed by stirring at 120
rpm for 1 minute, then at 20 rpm for 30 minutes. Treated soils were settled out of
suspension and the clear suspension was decanted. Subsequently, the heated soils were
washed fi'om the jars into a large container and allowed to air dry. Treated air dried soils
were then gently ground with a mortar and pestle, sieved through a US standard #20

sieve, and kept in open pans for conductivity experiments. Small amount of soils were

80

heated to a mid point (0.8 CEC) in the same manner as explained above, for comparison
in the turbidity experiments. Hereafter in this paper, low-, mid-, and high-heated soils
refer to the soils heated to 0.1, 0.8, and 2.2 CEC, respectively. For the purposes of this
work, the CEC was determined by the first plateau of the HDTMA adsorption isotherm
(Xu and Boyd, 1995). This reference heahnent level was found to be 27 mmole/kg for the
soil sample with 12% fines.

According to our previous study and the work of Xu and Boyd (1995) we
expected the low- and high-heated soils to have a dispersed clay shucture and the mid-
heated soil a flocculated one. A set of turbidity measurements was conducted to verify
that whether after the samples were air dried and rewetted with water the clay shucture
was changed. For these measurements 2.4 gr air dried sample from unheated, low-, mid-,
or high-heated soil was mixed with 25 ml aqueous solution of 1 mM NaCl in small Corex
tubes. The tubes were then shaken for 4 days in a rotating shaker and allowed to stand for
0.5 hr. Turbidity of the supernatant was determined by nephelomehy on the soil
suspension sampled from one cm below the surface in each tube. The procedure
employed was consistent with that used by Xu and Boyd (1995) and in compliance with
ASTM D1889.

A fixed ring consolidometer, which allows falling head conductivity
measurements, and a load frame equipped with a precision displacement hansducer
(0.0001 inch sensitivity) were employed for conductivity experiments. Treated and
unheated samples were packed by pouring air dried soils into the consolidometer ring

with filter paper (Whahnan #54) on the top and bottom of the sample. Samples were then

81

tamped slightly to flatten the upper surface, and compacted with a drop hammer to a
nominal initial thickness of 2.5 cm, and vacuum saturated in a desiccator with an aqueous
solution of 1 mM NaCl. Saturated samples were then subjected to an effective shess
(loading) sequence fi'om 0 to 8 tsf using a conventional consolidation procedure (ASTM
D243 5-90).

Hydraulic conductivity of the system at different effective shesses was measured
by the falling head method (ASTM D5084-90). Head losses in the tubes, connectors,
porous stones, and filter papers were quantified by assembling the consolidometer
without any soil and measuring its conductivity (K.,m) at different effective shesses.
Hydraulic conductivity of the soil sample was then calculated horn the predetermined
Kbunk and two or three conductivity measurements at that load and adjusted to 20 °C. At
each load, the sample thickness was measured and porosity was calculated from the
measured thickness of the consolidated sample.

RESUL rs AND DISCUSSION

Treatment level and clay shucture

Treahnent of soils with HDTMA to different levels determined the turbidity of
their suspensions (Figure 14). Closed symbols on Figure 14, which follow a general L
shape curve relating turbidity to HDTMA added, are the results from the present study.
For ease of comparison, turbidity results from our previous study are superimposed on
the figure and shown in open symbols. We reiterate here that the only difference in soil
samples prepared for turbidity tests in two studies is that in the previous study unheated

soil samples were equilibrated with HDTMA solutions (with different concenhations)

82

Relative mercury WIN"

Relative Turbidty

 

Relative mercury

 

o 0.5 1 1.5 2 25 3
AmomtofilJTMAaddedtocoflenonflz-dtotinCEc

Figure 14: Turbidity of soils with different fine contents

83

and in that wet condition, turbidity of their suspension was measured. In the present study
low-, mid-, and high-heated soil samples were first air dried and then rewetted with
HDTMA-flee water, and then turbidity of their suspension was measured.

As shown in Figure 14, in unheated and low-treated (0 and 0.1 CEC) samples, the
suspension was turbid indicating a dispersed clay shucture while in mid- and high-
heatedsamples (0.8 and 2.2 CEC) the clear suspension indicated a flocculated clay
shucture. High-heated soil obviously had a very different turbidity than the same soil in
the previous study. This result is also inconsistent with findings of Xu and Boyd (1994,
1995) who reported high turbidity for samples heated to levels far above the CEC.

It is known that sorbed HDTMA in excess of CEC forms a relatively weak
hydrophobic bonding to the clay surfaces, giving rise to easy desorption if the HDTMA
equilibrium concenhation in the liquid is perturbed (Xu and Boyd, 1994). This
phenomenon apparently manifested itself in redistribution of the heahnent profile in the
in-situ modified soil columns studied by Burris and Antworth (1992). They injected
HDTMA to a soil column and showed that most of the HDTMA’s were adsorbed close to
the column inlet. However, this profile was redishibuted to a more uniform one afier they
flushed the column with HDTMA-free water.

It is likely that in our heahnent process during washing, air drying, and saturating
the high-heated soils with HDTMA-flee water, some (or all) of the hydrophobically
bonded HDTMA molecules were desorbed, such that the clay particles retained a
flocculated shucture which normally corresponds to a heahnent level slightly below

CEC. Since the low- and high-heated samples prepared for conductivity tests were

84

processed in the same manner as the samples for turbidity (i.e. air dried and then
saturated with HDTMA-free water), it was believed that low-heated samples had a
dispersed clay shucture and conversely, high-heated ones had a flocculated clay shucture
in our conductivity tests. This is consistent with qualitative observations made during
conductivity experiments that eflluent hour the consolidometer reservoir was clear in all
high-heated conductivity experiments while cloudy in low-heated ones.

Porosity

Dry packed low-heated samples were 2 to 9% more porous than high-treated ones
depending on the clay content (T able 6). Average porosities of both low- and high-heated
samples increased with their clay content. However, this increase had a smaller overall
magnitude in low-heated samples than in high-heated ones. As the fine content increased
from 6% to 24%, porosity of low-heated samples increased by 3.7%, while this increase
was 5.4% in high-heated ones. The dependence of porosity on fine content (especially in
high-treated samples) highlights the role of clay content (and its shucture) in the initial
assemblage of the soil matrices.

Saturation with water changed the porosity of low- and high-heated samples less
than 3% (Table 6). All low-heated samples remained more porous than high-heated ones
after saturation. This minimal change, which was attributed to procedural handling of the
samples, verifies that the type of clays in this soil were primarily non-swelling. In
general, variation in porosities of hiplicates were small compared to their means,
indicating that the packing procedure was capable of establishing repeatable overall

volumes in replicate samples.

85

oflseeaaetgoaueeu

.o. .3 .o 8o. .8 8358 8.9.8.... .

 

 

coo. Ron 8 8368 8.8.28.
use. 0.3 8 5.8.23 2 .2... +
o. .3. «3e .3. «3e .3. «one o... .3. «3.. .3. «3e .3. «3... 8
E .3. «8.3 .3. «3.. .3. «3.. 3 .3. «3.. .3. «3e .3. «3.. m.
.... .3. «m. z. .3. «.3.. .3. «3.. 3 .3. «3.. .3. «3.. .3. «3.3 N.
3 .3. «3.. .331... .3. «3e 3 .3. «3.. .3. «3.. .3. «3... o
.2 .8o. .2... 8.2.80 .888 an. .4 .8o. 8.... .8853 .888 8.. 3... 8.80

 

.3. «.8 882.8... c. ...

 

.9... 2.8 8.8.32 e. ...

woe...

8o. 8... o... .o

..0...~0._...._.tv 0. 2... $3.08 3.8.0.. ecu fine. .8: o... .23. o.o~ 8 m_.om .0 ..em. a w. 5.8.0.. omEo>< .m 2.8..

86

With application of the first load, porosity of the low-treated samples dropped
much more than that of the high-treated ones (Table 6). This difference in drops increased
with the clay content such that at the first load, low-treated samples with 6% fine samples
were more porous than high-treated ones but at higher fine contents the situation was
reversed. The difference in porosities of low- and high-treated samples at zero load and
their different response to the first load suggest that there is a correlation between the
structure of clays and load bearing mechanism of the soil matrices which is different in
low- and high-treated samples.

The role of dispersed and flocculated clays in the consolidation behavior of low-
and high-treated soil samples can be conceptualized in order to explain the trends in the
data. In general, clay particles may reside in the pore space created by sand grains or
partly there and partly in between sand grains of a soil matrix which makes them part of
the load bearing mechanism of the matrix (Kenney et al., 1992). It appears that in our
low-treated samples, dispersed clays resided between the sand grains, especially in
samples with high clay contents. Furthermore, the clay platelets may be poorly ordered
relative to one another (i.e. in a “house of cards” arrangement). This resulted in a
relatively unstable but more porous low-treated soil matrix at zero load. Application of
the first load collapsed this unstable structure, resulting in a relatively large consolidation
of the matrix. This trend was expectedly more profound in samples with higher clay
contents. On the other hand, high-treated samples were not as porous initially, apparently

due to the highly ordered (parallel) structure of their flocculated clays. As a result, the

87

first load only slightly decreased their porosity; far less than the magnitude of the drop in
low-treated samples. O

Subsequent load applications (following the first), produced additional
consolidation but no abrupt changes were observed (Figure 15). In the 6% fine case
(Figure 150), low-treated samples were, on average, more porous than high-treated ones
but a relatively high variation was observed in all samples. The soils with higher clay
contents exhibited less variation and a clear distinction between low- and high-treated
samples was observed (Figure 15b~d). At low loads, low-treated samples were less
porous, and somewhat more resistant to consolidation than the high-treated ones.
Apparently in the low-treated matrix, the dispersed clays facilitated rearrangement of
particles under a relatively low load causing the matrix to collapse and lose most of its
porosity under the first load (0.25 tsf). Higher loads were then sustained by the packed
soil matrix which had limited space to consolidate. High-treated soils, on the other hand,
had a more stable structure at zero load. The magnitude of their porosity loss due to the
first load was comparable to their consolidation rate throughout the loading sequence.
This means that their flocculated clay structure resulted in a relatively dense and stable
initial pack which consolidated gradually as the loading increased. Therefore, flocculated
clays may have actually been an essential part of the load bearing mechanism of the high-
treated soil samples, resulting in higher porosities compared to low-treated samples at
low loads.

The differences between porosity of low- and high-treated samples decreased with

load such that nearly identical porosities were observed for the two materials at high

88

0.1 Load ([81) 1 10

 

 

 

 

LowTreat
42 ‘ L [6%

 

 

 

 

 

 

 

 

 

[Hignrreatl

 

M. n. (W
8
Mb;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8

 

 

 

 

 

 

 

 

 

a fJ—Ll
41 n HighTreat 12% 41
a : l
g: 39
g 3., .__[LowTreat} '
35 $4
33

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MAM
eraasets

 

MAM
88838886t$8

0.1 Load M 1 10

Figure 15: Porosity of soils with 6%, 12%, 18%, and 24% fines

89

loads. Identical porosities in low- and high-treated samples at high loads, suggest that
consolidation proceeded essentially independent of the differences between the structure
of the clays (i.e. dispersed or flocculated). In other words, consolidation was ultimately
controlled exclusively by the larger soil particles which were identical between the two
materials. At such high loads, clays evidently accommodated rearrangement of granular
particles (and hence consolidation) in the same manner regardless of their structure.

The same data may be viewed in combination with untreated and mid-treated
samples from previous study (Figure 16). Average values for porosity of low-treated
samples from the present study (closed symbols) and that of untreated samples fi'om the
previous study (open symbols) as a function of clay content are shown on Figure 16-a. As
it can be seen, samples with 6% fine material (untreated or low-treated) were the most
porous throughout the loading sequence. Their consolidation rate was very low which, in
general, is similar to the response of a sand matrix to the load. Apparently, sand particles
were the main load bearing fi'action of the soil in these samples. On average, samples
with higher clay contents showed steeper consolidation rates which suggest that more
clays facilitated easier rearrangement of particles. This behavior is consistent with the
common observation that clay soils are more compressible than sands (Freeze and
Cherry, 1979). In comparison of low-treated samples to untreated ones, samples with the
same fine contents had nearly identical consolidation behaviors. This is a manifestation of
the same dispersed clay structm'e in both soils.

A plot of the porosity of high-treated samples with different clay contents and its

comparison to the mid-treated samples, fi'om our previous study, reveals that porosity of

9O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.1 Loam 1 10
44
42
o .__
‘0 "N
N
n N \ ii
a 3" N
N
=: \‘
x N
\N
g +6%|0Treet #
34 +12%ioTnet \
+1emo'rmt \ \
32 +24%loTreet \
A 695M
I: 1295M \
3° 9 18%mtreeted ‘
O 24%mtreded
28 l I I I Tj T
$.—
\.+
8‘\
44 \N
N N
42
I.\
L- q
-40
t
‘2 38 +695 Hrrmr .
+12%Hfl’red \K
+18%HiTreet
38 +24%Hi‘l'red
A “Must \\
34 n 12%MidTreet \‘ _
o 18%MidTreet \
0 2495mm
32
[E30
0.1 Loam 1 10

Figure 16: Porosity of a)low-treated and untreated, and b)mid
and high-treated samples (untreated and mid-treat
data from the previous study)

91

corresponding samples from the two groups had very similar consolidation behaviors
(Figure 16-b). This demonstrates that they both had similar (i.e. flocculated) clay
structures. Typically, samples with higher clay contents showed steeper consolidation
rates under the load. This resulted in a reversed situation at high loads where greater clay
contents corresponded to lower porosities. Evidently, higher clay contents accommodated
easier rearrangement of particles and hence produced higher consolidation with each
load. The general trend in porosities of untreated and low-treated soils with the load
(Figure 16-a) was a diverging trend, while it was an intersecting trend in the case ofmid-
and high-treated soils (Figure l6-b). This fundamental difference in the response of these
soils to the load portraits the significance of the role of clay structure (i.e. dispersed in
untreated and low-treated soils and flocculated in mid- and high-treated soils) in
consolidation behavior of soils. As will be discussed in the following section, clay
structure also played an important role in conductivity of the these soils.

Conductivity

Hydraulic conductivity is a function of soil matrix property (permeability) and the
fluid properties of viscosity and density (Freeze and Cherry, 1979). This implies that a
difference in conductivity of two soils can be produced by a difference in water
characteristics as it permeates through the soils or by difference in soil structures. The
viscosity and density of the permeant (i.e. water) were measured to vary less than 1% as a
result of contact with low- and high-treated soils. Therefore, the observed difl‘erences in
conductivities were all attributed to the difl‘erences in permeability of the soils.

Permeability of a packed soil matrix, on the other hand, can be further characterized by its

92

total porosity and effective pore radius (Wallace et al., 1995). Based on the porosity
results, contribution of each component (i.e. porosity and effective pore radius) to the
measured difference in conductivity of low- and high-treated soils is established.

Conductivities of low- and high-treated samples at no load did not show a
consistent trend (Table 7). However, application of the first load (0.25 tsf) resulted in a
substantial decrease in conductivity of the low-treated soil, and a minor decrease in the
high-treated one. The drop in conductivity, with the initial load, depended on the clay
content in both low- and high-treated samples. These results are consistent with the
observed decreases in porosity, however, porosity decreases only made a minor
contribution (< 19%) to the observed difference in conductivities. We therefore conclude
that the change in effective pore radius was the primary cause of conductivity variation in
these soils.

The low-treated soil formed a more porous matrix upon packing which was very
sensitive to external load in terms of its effective pore radius. Application of the first load
collapsed the matrix and apparently forced the clay particles from between the sand
grains into the interior regions of pores. This had the effect of partitioning the pores into
smaller sections (i.e. smaller pores) and lowered conductivity. Higher clay contents
produced a more porous matrix initially, resulting in a higher drop in conductivity with
application of the first load. This also dropped conductivities of the high-treated samples,
however to less of an extent than in the low-treated ones. Flocculated clays in the high-

treated samples apparently formed bigger aggregates, which were less easily forced into

93

000228053! \ggga

E 03 .o .80. a 80......» 8.9200 .

 

 

08. as a 8......8 8.238 .
E .80.... «83 88.... «83 3~ .038... «.8... 88.0. «08... «u
3 a8... «83 .80.... «m8... 3. .883. «N8... :8... «83 2
0.. .83. «08... .23. «2...... 0.2 883. «08... 88.... «.03 N.
3 .80.... «83 .80.... «80... 0+ :8... «83 .08.... «83 0
a. .08. a... $2928 a. .08. a... .3928 .9... 2.280 8....

 

.88. 0.8 8.80-8... c. v.

 

.080. 0:00 0200.159 0. X

08. .9. m... .0

00:00:30 0. 000 00.0 a... 0:0 .002 .9: 05 .000. 0.0N .0 0:00 .0 am..." v. 5.2800000 0.30.0.5 000.024. K 0.00...

94

the interstitial spaces (as discussed in the Porosity section) and were less able to partition
the pore space into small sections.

Triplicate conductivities of low- and high-treated soils with different clay contents
are shown in Figure 17. Conductivity of soils with 6% fine showed slight unrealistic
trend of increase with the load and had a relatively high variation. This is likely due to
limitations in precision of the apparatus at such high conductivities. Solid lines on the
graphs represent two standard deviations established by the maximum of 1) observed
variance among triplicates, or 2) error propagated from measurement uncertainties. As
shown in Figure 17, all high-treated samples had higher conductivities than low-treated
ones throughout the loading sequence. These difl‘erences were sustained over the applied
range of loads with similar systematic decreases in both types of soils. High-treated
samples were 4 to 71 times more conductive depending on the load and clay content.

As indicated previously, the relative contributions of porosity differences to the
observed differences in conductivities were calculated and found to be less than 19%.
This suggests that we can extend the interpretation that treatment to different levels
primarily affected the sizes of the conductive pores to the full range of the loading
sequence. It is interesting to note that at high loads, where low- and high-treated samples
had identical porosities (Figure 15), conductivities were substantially different. Because
the clays were occupying the same total volume in both low- and high-treated samples,
difi‘erences must result from large significant differences in pore geometry. The larger
pores sustained by the flocculated high-treated clays produced conductivities up to 71

times higher.

1 E-01

Hyd. Conductivity (cm/s)

LowTreat

rs-oe LowTreat

Conductivity (cmls)
Fri
S

Hyd

 

E

A
In
is
A

 
 
 
  

E HighTreet EE-

3

 

Hyd. Conductivity (emit)

 

 

 

 

Hyd. Conductivity (culls)

 

 

 

l 1 4
15.03 i 1 1 1 ,
0.1 Load (an 1 10

Figure 17:Hydraulic conductivity of soils with 6%, 12%,
18%, and 24% fines

96

As with porosity, it is insightfirl to view the conductivity data over the full range
of treatment levels (Figure 18). On this figure, average values for conductivity of low-
and high-treated samples from the present study (closed symbols) and that of untreated
and treated samples from previous study (open symbols) as a firnction of clay content are
shown. In general, hydraulic conductivity of treated and high-treated samples were higher
than conductivity of untreated and low-treated ones. This is obvious from a comparison
of the number of log cycles on conductivity axes on Figures 18-a and 18-b. It can be seen
that the loss of conductivity with consolidation for both untreated and low-treated soils
was similar and directly related to the amount of fine material (Figure 18-a). This is
entirely consistent with the common observation that higher clay contents correspond to
lower conductivities and higher load sensitivities. Average conductivities of high-treated
samples with different clay contents show that the higher the clay content, the lower the
conductivity (Figure 18-b). It can be seen in Figure 18-b that a similar trend was also
observed in treated samples. Conductivities of samples with higher clay contents were
generally more load sensitive such that the difference in conductivities was magnified at
high loads.

IMPLICA norvs For! FIELD APPLICA norv

The introduction of cationic surfactants such as HDTMA into the subsurface
i could be utilized for creation of an in-situ sorptive zone. This could be achieved by
injection of an aqueous solution of HDTMA through a set of injection wells into the soil
down gradient fi'om the contaminated aquifer. The in-situ sorptive zone would then serve

to intercept and immobilize the dissolved organic chemicals as the contaminant water

97

1 E-01

1E-02

5

+696 loTreat
+12% loTreat
+ 1 8% loTreat
+24% loTreat
1E-05 A 6% untreated
u 12% untreated
o 18% untreated
O 24% untreated

'E

Hyd. Conductlvlty (am)

 

1E-06

1 E-O1

g

HlTroat
+12%HiTreai
+18%H1Treat
+24%H1Treai

A 8%MidTreli
D 12%me
o 18%M1dTrut
O 24%MidTl'ofl

E

Hyd. Conductivity (cm/s)

  

E 1E-O4
o.1 Load M 1 10

Figure 18: Hydraulic conductivity of a)low-treated and untrea
and b)mid- and high-treated samples (untreated an
mid-treated data from the previous study)

98

permeates through it. However, for the reasons discussed earlier, it is likely that the
treatment level of the soil in the regions close to the injection wells will exceed the CEC
of the soil. In contrast, in the regions far away from the injection well, it is reasonable to
expect a low treatment level in the soil, possibly far below the CEC. Treatment of the soil
to different levels has the potential to alter conductivity in expected ways and hence its
characterization is essential in determination of hydraulic feasibility of in-situ created
sorptive zones.

From the standpoint of field application, the combined results of this and the
previous study show that the creation of a sorptive zone by in-sr'tu injection of HDTMA is
hydraulically feasible. Soil samples treated to different levels of 0.1, 0.8, and 2.2 times
the CEC all had similar or higher conductivities when compared to the untreated soil
under the experimental conditions used in this study. This higher conductivity was up to
71 times greater in high-treated and up to 129 times greater in mid-treated samples in our
experiments. In an in-situ created sorptive zone, higher conductivities of heated soil
relative to the untreated soil in its vicinity could effectively draw in flow lines from
surrounding soils, and therefore increase the region of influence or capture area of the
zone.

While the presented data indicate that an in-situ created sorptive zone will
function favorably in groundwater flow situations, a general recommendation can be
made as to the injection scheme. In order to prevent the development of hydrophobically
bonded HDTMA (as would be expected at high treatment levels) a flushing period may

prove to be desirable after the injection of HDTMA. Hydrophobically bonded HDTMA

99

would likely be desorbed, washed down gradient, and re-adsorbed to the soils treated to
levels lower than their CEC. This would result in a more uniform and stable HDTMA
distribution and hence a more uniform and permanent sorptive capability. Furthermore, it
is likely that clays will retain a flocculated structure, following the desorption of
hydrophobically bonded HDTMA, which generally means a higher conductivity in the
treated soil.

CONCLUSIONS

The combined results of this and the previous study show that the creation of a
sorptive zone by in-situ injection of HDTMA is hydraulically feasible. Soil samples
treated to different levels of 0.1, 0.8, and 2.2 times the CEC all had similar or higher
conductivities when compared to the untreated soil under the experimental conditions
used in these studies. This is interpreted as resulting from neutralization of surface charge
on untreated soils, causing naturally dispersed clays to flocculate. Contrary to our
expectations deriving from previous reports, high-treated clays were also highly
conductive and the clays were apparently flocculated. It is believed that this resulted from
desorption of the hydrophobically bonded HDTMA at levels above the CEC during
washing with water.

High-treated samples displayed similar or higher porosities compared to low-
tr'eated ones. This was attributed to easier rearrangement of particles in the presence of
dispersed low-treated clays. Higher conductivity of all high-treated samples (with
different clay contents) compared to that of low-treated ones was attributed to their

flocculated clays resulting in larger effective pore radii in those samples.

l00

In an in-situ organomodification scheme, our data suggest that washing with
HDTMA-free water may desorb the weakly bonded excess HDTMA close to an injection
well and transport it to downstream regions. Expected differences in modification levels
that might result as the solution moves out from an injection point or from soil
heterogeneity, will be naturally dampend by this apparent tendency of the soil not to
retain HDTMA at levels far above the CEC. The process of creating a uniform zone of
modified soil may be as simple as following a period of injection by a flushing step.
REFERENCES

Burris, D. R. and Antworth, C. P., 1990. Potential for Subsurface In-situ Sorbent
Systems. Groundwater Management (Dublin, Ohio) 42527-538.

Burris, D. R. and Antworth, C. P., 1992. In Situ Modification of an Aquifer Material by a
Cationic Surfactant to Enhance Retardation of Organic Contaminants. J. Cont.
Hydrol., 10:325-337.

Boyd, S. A., Lee, J. F., and Mortland, M. M., 1988a. Attenuating Organic Contaminant
Mobility by Soil Modification. Nature (London), 333:345-347.

Boyd, S. A., Mortland, M. M., and Chiou, C. T., 1988b. Sorption Characteristics of
Organic Compounds on Hexadecylhimethylammonium-smectite. Soil Sci. Soc.
Am. J. 52:652-657.

Boyd, S. A., Sun, S., Lee, J. F., and Mortland, M. M. 1988c. Pentachlorophenol Sorption
by Organo-clays. Clays Clay Miner. 36:125-130.

Daniel, D. B., 1989. A Note on Falling Headwater and Rising Tailwater Permeability
Tests. Geotechnical Testing J ., ASTM, 12(4)308-310.

Freeze, R A. and Cherry, J. A., 1979. Groundwater, Prentice-Hall, Inc., Englewood
Cliffs, N. J. 604 pp.

Jaynes, W. F. and Boyd, S. A., 1991. Clay Mineral Type and Organic Compound 1
Sorption by Hexadecylhimethylammonium—exchanged Clays. Soil Sci. Soc. Am.
J ., 55:43-48.

Jaynes, W. F. and Boyd, S. A., 1990. Trimethylphenylammonium-Smectite as an
Effective Adsorbent of Water Soluble Aromatic Hydrocarbons. J. Air Waste
Management Assoc. (Pittsburgh, PA) 40:1649-1653.

Lee, J. F., Crum, J. R., and Boyd, S. A., 1989. Enhanced Retention of Organic
Contaminants by Soils Exchanged with Organic Cations. Env. Sci. Tech.
23:1365-1372.

Shackelford, C. D. 1994. Waste-soil Interactions That Alter Hydraulic Conductivity.
Hydraulic Conductivity and Waste Contaminant Transport in Soil, ASTM STP

101

1142, David E. Daniel and Stephen J. Trautwein, Eds., Philadelphia, PA, pp. 111-
168.

Theng, B. K. G. 1974. The Chemistry of Clay Organic Reactions. John Willey & Sons,
New York.

Wallace, R. ., J. M. Grant, T. C. Voice, G. R. Rakhshandehroo, S. Xu, and S. A. Boyd.
1995. Hydraulic Conductivity of Organomodified Soils. Proceedings of 1995
National Conference on Environmental Engineering.

Xu, S. and Boyd, S. A., 1994. Cation Exchange Chemistry of Hexadecylhimethyl
ammonium in a Subsoil Containing Vermiculite. Soil Sci. Soc. Am. J ., 58:1382-
1391.

Xu, S. and Boyd, S. A., 1995. Cationic Surfactant Sorption to a Vermiculitic Subsoil via
Hydrophobic Bonding. Environ. Sci. Technol., 29:312-320.

APPENDIX

APPENDIX

DETAILED MATERIALS AND METHODS

This appendix contains detailed explanation of the materials used and
experimental procedures followed in the course of this study. It includes material and
procedures used in all three chapters of this dissertation and should be used in
conjunction with them. Methods and procedures common in all chapters are explained
without reference to any specific chapter. Materials and methods specific to each chapter
are stated as such in this appendix.

SOIL PROCESSING

Approximately 100 lbs of Oshtemo soil was obtained from a pit located at the
Kellogg Biological Station (KBS) for experiments conducted in Chapter 1 of this
dissertation. The soil was taken from a depth of 1.5~2.5 it from the surface and hauled to
Engineering Research Complex (ERC) in 2 plastic buckets. The collected soil was spread
outside ERC building on plastic sheets and left on sunshine for 2 days until] air dried.
During air drying process the soil was occasionally mixed on the plastic sheets to ensure
a uniform drying and any plant root was taken out. Air dried Oshtemo soil was sieved
through a US standard #20 in a Ro-tap shaker for 5 minutes and then thoroughly mixed to
obtain a homogenized stock soil. In order to determine the quantity and type of
exchangable cations and also organic'mattcr content, approximately 0.5 lb of the stock

soil was sent to MSU soil testing laboratory located at the Plant and Soil Sciences

102

103

Building. Similar amount of the stock soil was sent to Dr. Boyd’s laboratory (Crop and
Soil Sciences Dept, MSU) for determination of CEC by HDTMA.

Original Oshtemo soil for experiments conducted in chapters 2 and 3 was
obtained from a difl'erent pit located at the KBS. Approximately 150 lb of the soil was
taken from a depth of 2~3 ft. The collected soil was air dried, sieved and mixed in
identical manners to the previously described procedures and the obtained soil was named
orginal soil. Exchangeable cations and organic matter content of the original soil was
determined by MSU soil testing laboratory. HDTMA batch iotherm and CEC
determination of the original soil was performed by J. Gellner (thesis, 1996).

Soil samples used in chapters 2 and 3 were prepared by recombination of the .
original soil’s components. The recombination process was as follows; original soil was
thoroughly washed by tap water on long wall US standard sieve #200 until the suspension
passing the sieve was clear. Meanwhile, the passing suspension was collected in big
plastic buckets and trays and left for 15 to 20 hrs. Clear supernatant was subsequently
decanted from the buckets and trays and fine materials were combined into one tray and
left for air drying. By this procedure the original soil was divided into clean sand and
fines fiactions. Air dried fines were manually pulverized with a pestle and mortar, sieved
through a US standard #100, and kept in closed plastic containers. The clean sand fi'action
was first spread on plastic sheets and left to air dry and then collected in a galvanized
bucket. Soil samples with 6, 12, 18, and 24% fines were prepared by dry recombination
of appropriate amounts of pulverized fines with clean sand. Each dry recombined soil

sample was first thoroughly mixed. Each batch of soil sample preparation usually

104

contained total of 400 gr soil. Therefore, 24, 48, 72, or 96 gr fines were mixed with 376,
352, 328, or 304 gr clean sand in order to make soil samples with 6, 12, 18, or 24% fines,
respectively. Dry recombined soil samples were then mixed with deaired deionized water
in order to create adhesion of the fines to the sand fraction. In this process 4, 6, 8, or 10%
water content was employed for soil samples with 6, 12, 18, or 24% fines, respectively.
Therefore in batches of soil sample preparation which contained 400 gr soil, total of 16,
24, 32, or 40 ml water was added to the soil samples with 6, 12, 18, or 24% fines,
respectively. These waters were added in three equal amounts and each time soil samples
were mixed until the added water was distributed evenly throughout the sample. The
recombined soil samples were then spread on plastic sheets and left for 15 to 20 hrs to air
dry. Subsequently, soil samples were gently crushed with a pestle and mortar, sieved
through a US standard sieve #20, and kept in clean containers. These samples constituted
the untreated soil samples with different fine contents which were later packed in the
consolidometer for conductivity tests in Chapters 2 and 3. Treated soil samples were
prepared fi'om these unheated soil samples by further processing.

Batch treatment of the soil samples for all three chapters was conducted in 2 liter
glass jars. In Chapter 1, HPLC grade water and in Chapters 2 and 3 DI water equilibrated
with 1 mM NaCl was used for this purpose. In treated soils prepared for Chapter 1, 60 gr
of the untreated soil was placed in each of the 6 jars and one liter water was added to it. In
separate 1 liter glass jars, one liter water was mixed with HDTMA solutions
corresponding to 1 CBC of the soil. While the jars with soil samples in them were being

mixed in a Phipps and Bird 7790-400 stirrer at 120 rpm, HDTMA solutions from 1 liter

105

jars were added to them. In treated soils prepared for Chapters 2 and 3 a different soil to
water ratio was employed. In these preparations, 131 gr of untreated soil was placed in
the 2 liter jars. Then 1365 ml DI water (equilibrated with 1 mM NaCl) was added to the
soils and the stirrer was turned on and set at 120 rpm. Subsequently, appropriate amounts
of HDTMA solution fi'om the manufacturer stock solution was added to jars. This amount
of HDTMA corresponded to treatment level of 0.8 CEC for the soil employed in Chapter
2 and to 0.1 and 2.2 CEC for the soil employed in Chapter 3. In all treatment procedures
the mixing process continued for one minute followed by a slow mixing of 20 rpm for 30
minutes. At this point the stirrer was stopped, paddles were taken out of the jars, and the
jars were allowed to sit for 15 to 20 hrs. At the end of this settling period, the clear
supernatant was decanted and the soil samples were all washed into a ceramic drying
bowl. Air drying usually took 1 to 3 days but at times blow drier was used to shorten this
time. In any case, the soil was frequently mixed, while drying, in order to prevent any
segregation of the soil fi'actions based on their settling velocity. Treated air dried soil was
sieved on a US standard sieve #4 and the retaining fraction was gently crushed with a
pestle and mortar and sieved again. This process was repeated few times until the entire
soil passed through the sieve. Treated air dried soils were kept in open plastic pans for
future experiments.
Tumor" TESIS

Fundamental turbidity tests were performed on all untreated soil samples. In these
experiments, a small sample of the soil, placed in 25 ml Corex tubes, was used to

determine turbidity of its suspension when difl‘erent amounts of HDTMA was added to

106

the soil. 2.4 gr air dried untreated soil was weighed into each of 24 tubes and 25 ml
solution of 1 mM NaCl water was added to each tube. Different amounts of HDTMA,
from the manufacturer’s stock solution, corresponding to 0, 0.01, 0.02, 0.03, 0.04, 0.06,
0.08, 0.1, 0.2, 0.3, 0.4, 0.6, 0.8, 1.0, 1.1, 1.2, 1.3, 1.4, 1.6, 1.8, 2.0, 2.3, 2.6, and 3.0 times
CEC of the soil sample were added to the tubes. The tubes were then sealed, agitated
gently by hand, and placed in a rotating shaker for 4 days. At the end of the 4 day period,
tubes were removed from the shaker and allowed to sit for 30 minutes. At the end of the
settling period, 1 ml of supemate was taken fi'om 1 cm below the surface. The sample
was then diluted with 20 mls of 1 mM NaCl water. Turbidity of the diluted solution was
then determined using a Hach 2400 turbidimeter. If turbidity of the diluted s'olution was
greater than the designated range in the turbidimeter, a 1 ml sample of the diluted
solution was taken and this sample was subsequently diluted with an additional 20 mls of
the 1 mM NaCl water.

The turbidimeter used in these experiments had 5 different ranges of
measurements (0-0.2, 0-1, 0-10, 0-100, and 0-1000 NTU’s) and a standard solution with
known turbidity for each range. Depending on the turbidity of the solutions which were
being tested, one range was selected. Prior to the turbidity measurements, the
turbidimeter was calibrated in the selected range using the standard solution (with known
turbidity) provided by the manufacturer. In each set of turbidity measurements only one
range was used and, as mentioned earlier, dilution of the soil suspension was employed in

order to keep the turbidities within the selected range. The turbidities were then

107

determined based on the dilution employed, and were recorded as the fundamental
turbidity data for that soil.

The second set of turbidity tests was performed on treated air dried soil samples.
Soil samples were obtained from the pans containing treated air dried soils explained
earlier. In this process, 2.4 gr soil was weighed into each 25 ml Corex tube and then 25
ml solution of 1 mM NaCl water was added to each tube. The tubes were then sealed,
agitated gently by hand, and placed in a rotating shaker for 4 days. At the end of the 4 day
period, the processes of settling time, sampling, dilution, and turbidity measurements
were followed in the same manner as explained earlier for the fundamental turbidity
experiments. These turbidities were recorded as turbidity of treated air dried soils in
water. An additional turbidity experiment was conducted only on the soil used in Chapter
1 with Soltrol as the fluid. In this additional experiment, 2.4 gr air dried treated or
unheated soil was weighed into each of two 25 ml Corex tubes and then 25 ml Soltrol
was added to each tube. Shaking the tubes, settling, sampling, and turbidity
measurements were performed similar to the fundamental turbidity experiments. These
turbidities were recorded as turbidity of treated and unheated soils in Soltrol. Therefore,
total turbidity measurements in the second set of turbidity tests included treated air dried
and unheated soils equilibrated with water or Soltrol (Chapter 1) and soils treated to 0.1,
0.8, and 2.2 CEC equilibrated only with water (Chapters 2 and 3).
CorvsouoouErER ASSEMBLA GE

The separate pieces of the consolidometer are shown in Figure 1A. Prior to any

assembly, each piece of the consolidometer was cleaned and air dried. In addition, the

108

Stand pipe

 

Consolidometer base plate
._ Top consolidometer plate
Effluent reservolr
Y Top metal disk

  
  
 
 
 
    
 
  

Top porous stone

 

Top firier paper

toc'rldlng Soil sample .
a
Bottom filter paper

 

Bottom porous stone
Consolidometer ring
- O-rlng

Locklng nuts

 

 

 

 

 

 

 

 

 

 

,,,,, i '

“—
., fix an . j "' ‘ {"1" '
_ a . .’\‘." 5"“ , . ... ‘, fl. ‘ _- ;._~"" 1.27:0! .

. __ _ .‘I,“_“ .. .. __ . ‘ . .‘. 7‘ -‘T...'_‘g>-'". 'gu‘lgl-H- . 1.‘v;;.:’- "'1‘ .‘._.._ 1,... .. .. ._ ~, .
' :s-e ‘.' .' min}.- :1i‘ffi'ég.$'.‘mi{ahma'Jmszn: :1 LL1M32¢$2 "Qgflfii‘ui‘éfihr .‘flvfit‘utfih Marina fink”, “

 

 

 

 

Figure 1A: Components at the consofldcmeier

109

upper and lower porous stones were sonicated in 0.01 N HCl (PH<2) for 12 minutes,
rinsed thoroughly, and dried. Two pieces of Whatrnan #54 filter paper (7 cm diameter)
were used for each assembly. The top filter paper was prepared by taking the
consolidometer ring and placing it in the center of the filter paper. The inside of the ring
was then h'aced using a fine point pen. The consolidometer ring was then removed, and
the filter paper was carefully cut along the outside edge of the traced circle. This h'acing
and cutting procedure produced a filter paper that would just fit inside the sample area of
the consolidometer ring and the top consolidometer plate.

After cutting the top filter paper, the consolidometer base plate was placed on the
ground. The lower porous stone was then placed in the base plate. Care was taken so that
no particulates were present when the lower porous stone was placed in the base. To
verify that no debris was lodged under the porous stone, it was spun with one hand. If no
particulates were present, the porous stone spun fieely, rotating several times with each
spinning motion initiated by hand. If the rotation was reshicted in any way, the stone was
removed and the base was checked for any particulates present. The aforementioned steps
were then repeated. After the porous stone was placed in the consolidometer base plate
and was in a “fiee spinning” condition, the lower filter paper (the one that was not cut)
was placed on top of the lower porous stone. The lower filter paper was slightly larger
than the lower porous stone and effectively covered the entire diameter of the surface
created by the upper surface of the lower porous stone and the consolidometer base plate.

The viton O-ring was then placed in the consolidometer base plate. The viton O-

ringhadadiameterthatwaslarge enoughtoallowatightsealbetweenthe O-ring andthe

110

consolidometer ring. The consolidometer ring was then seated in the consolidometer base
plate by placing it in the ceter of the O-ring and pushing gently downward. Proper seating
was achieved when the consolidometer ring slipped inside the O-ring and rested against
the filter paper resting on the metal surface of the base plate. Once the consolidometer
ring was seated in the consolidometer base plate, few grams of vacuum grease was placed
over the O-ring. The vacuum grease was applied by placing a small amount of vacuum
grease on one finger and then by running the finger around the base of the consolidometer
ring. This should have resulted in a continuous, beveled surface of vacuum grease which
completely covered the surface of the O-ring.

After application of the vacuum grease, the top consolidometer plate was placed
over the threaded dowels atached to the base plate. The bottom surface of the top plate
should have rested securely against the O-ring. The two locking nuts were then placed on
the threaded dowels and were tightened into place. Locking nuts were tightened
simultaneously to ensure proper formation of the vacuum grease seal not only on the O-
ring but also on the sealing surface of the consolidometer base plate.

The consolidometer was then ready for referencing and packing of a sample (soil
packing section). Once a sample was packed, or if a blank conductivity experiment was to
be performed, the top porous stone was screwed into the top metal disk. The top filter
paper was then placed on top of the sample (or placed in the middle of the consolidometer
if a blank system was to be tested) followed by placement of the top porous stone and

metal disk.

111

VACUUM SA rum nozv

Vacuum saturation of soil columns or blanks were achieved using a vacuum
dessicator and 4 liter vacuum flask, attached to a tap vacuum system. The entire vacuum
saturation system is shown in Figure 2A. Prior to any saturation procedure, the vacuum
flask was filled with 1 mM NaCl solution from a stock carboy. The 1 mM NaCl solution
was prepared by filling the carboy with distilled, deionized water to the level marked on
the outside of the carboy, corresponding to 17.3 liter of water. Then 1.013 gr ofNaCl was
added to the carboy and the solution was mixed manually. The vacuum dessicator was
emptied and cleaned using tap DI water. This cleaning process was repeated at least once
daily to remove any particulates and any possible build up of microbial growth. The
vacuum dessicator was then filled with 1 mM NaCl solution to a level just below the
ceramic plate (water level 1 in Figure 2A). If the dessicator was cleaned, excess water
was poured out of the dessicator so that the remaining water filled the dessicator to water
level 1.

The consolidometer containing the sample to be saturated (or blank
consolidometer) was then placed in the vacuum dessicator on top of the ceramic plate.
Two 500 m1 bottles filled with water were also placed in the dessicator to reduce the
volume of water needed to fill the dessicator to levels 2 and 3 (Figure 2A) during
saturation. The rim of the dessicator was coated with a thin film of vacuum grease to
ensure an air tight seal. The top of the dessicator was then placed on the rim and moved
laterally back and forth to ensure that a seal had formed. The appropriate tubes and

stopper were then placed in the top of the dessicator. The stopper and appropriate tubes

112

 
 

 

   

H C<>
G
0
(3)
@ C S tu
e m
_ ED 8 D
To vacuum A Q) TUb'nQ
A.-.“ an men-nu... - we»..-«.«..~.-.-a.--.-.unm.m. - .r..-.s-.«.~..1W ..
‘ (D D connect I
j E 6D
I Dessicaior

 

 

 

 

 

 

 

 

r; / water level 3
waiér level 2 © ©
069 J_L_ water level i / / \
I l Stirrers

/ © © 3%

 

Figure 2A: Schematic of the Vacuum Saturatlng set up

113

were also placed in the top of the vacuum flask, making sure that the stopper was
securely in place. The tubing leading from valve 2 to the vacuum flask was then put
together with a tubing connector as shown in Figure 2A.

The tube leading fiom the tap water vacuum system to valve 1 was then put into
the vacuum connection at the base of the tap water vacuum system. Valve 1 was moved
to position A, valve 2 was moved to position D, and valve 3 was moved to position G.
Moving valve 3 into position G shut ofi‘ any flow between the vacuum flask and the
dessicator system. The building tap was then turned on, allowing a vacuum to be applied
to both the vacuum flask and the dessicator. The stirrers under both the vacuum flask and
the dissicator were then turned on. During the deairing process, the vacuum' achieved by
the system was checked with a pressure tensiometer at the septum attached to valve 2.
The vacuum achieved should have exceeded 900 mbars. Failure to reach this vacuum
generally indicated a leak in the system.

The vacuum was applied to both vessels for at least one hour if a sample was to be
saturated or at least 30 minutes if a blank consolidometer was to be saturated. This time
period was intended to deair the water contained in the system. At the end of the deairing
period, valve 2 wastumed to position E. This disconnected the vacuum from the vacuum .
flask. The vacuum in the vacuum flask was then released by disconnecting the tubing
between valve 2 and the vacuum flask at the tubing connector. The stopper in the vacuum
flask was then loosened, but not removed, and the stirrer under the vacuum flask was
tinned ofl‘. Valve 3 was then moved to position H and water was allowed to flow fi'om the

vacuum flask to the dessicator. Valve 3 was left in position H until the water level in the

114

dessicator had risen to just below the lip of the consolidometer (water level 2). Valve 3
was then moved to position G, shutting off water flow fiom the vacuum flask to the
dessicator.

The system configuration was left in the indicated configuration (valve 3 position
G, valve 2 position E, valve 1 position A, tubing connector disconnected) until the
saturation period was complete. Saturation period was 30 minutes if a blank
consolidometer was being saturated, and overnight (at least 12 hours) if a soil sample was
being saturated. Overnight vacuum saturation of the consolidometer allowed the soil
sample to completely saturate under the vacuum. At the end of the saturation period,
valve 3 was moved to position H and the water level in the desiccater allowed to rise to
water level 3, above the rim of the consolidometer. Once the water level in the desiccater
rose to level 3, valve 3 was moved back to position G.

After inundating the consolidometer with water (water to level 3), the tap water
was turned off. This stopped the supply of vacuum to the system. Then the tubing which
connects the tap water vacuum system to valve 1 was removed from the connection to the
tap water vacuum system. This released the vacuum remaining on the dessicator. Valve 2
was then moved to position D, allowing further release of the applied vacuum. Once the
vacuum had completely dissipated, the top of the dessicator was carefully removed fi'om
the dessicator and the standpipe was attached to the consolidometer while the
consolidometer remained under water. Any residual air which was left in the fittings or
the canal of the consolidometer was then removed by applying a small vacuum on the

standpipe. Water was brought into the standpipe at least four times and until no air was

115

observed escaping from the standpipe during application of the vacuum. The plug was
screwed in the consolidometer base and secured. Teflon tape was applied to ensure the
seal and the consolidometer was removed fiom the dessicator. The remaining water in the
vacuum flask (if any) was used to supply the squeeze bottles used in the conductivity
experiments.

BLANK Corvoucnm'r MWURMNIS

Blank conductivity measurements, separate from the conductivity experiments,
were performed to establish the conductivity of the system without soil, K.,“, and to
establish a variance due to variation in the filter papers employed. Before assembling the
consolidometer, prelabeled stones were sonicated in 0.1 N HCl (PH<2) solution for 12
minutes. Stones were then rinsed with tap water and dried manually.

The consolidometer was then assembled without soil and placed in the vacuum
dessicator. The entire system was deaired by vacuum as described before for
approximately 30 minutes. The water in the dessicator was then brought up to a level just
below the lip of the consolidometer reservoir (water level 2 in Figure 2A). The system
was then allowed to saturate for 30 minutes under vacuum. Saturation was verified by
observation of water inside the consolidometer reservoir.

After saturation, the blank consolidometer system was inundated with water
(water level 3) and the vacuum was released. While the consolidometer was under water,
the standpipe was attached to the consolidometer system, ensuring that no air entered the
system while being attached. Any residual air left in the canal beneath the bottom porous

stone was purged fi'om the consolidometer by applying a small vacuum on the standpipe

116

while the consolidometer remained below the water surface with both valves open.
Vacuum was applied by a rubber tube attached to the tap water vacuum system. Water
was drawn into the standpipe at least four times and until no air was observed in the '
standpipe during application of the vacuum.

Once all of the air was purged from the consolidometer system, the plug on the
consolidometer base was screwed in and secured. Teflon tape was applied to ensure the
seal. The consolidometer was then removed from the dessicator and the external parts
dried manually. The loading ball was then placed in the top metal disk and the
consolidometer was placed in the load frame. With the middle load shaft well above the
loading ball and the low pressure regulator opened until a negative reading was observed
on the pressure hansducer readout, the regulator selector was moved to the low position
and the load/unload valve moved to the load position. The load fi‘ame pressure was then
tarred by slowly increasing the applied pressure using the low pressure regulator and
observing the stage movement relative to a reference line marked on the stage shafi
beneath the stage. The low pressure regulator was then adjusted until no movement was
observed, relative to the reference line on the stage shaft. Once no movement was
observed, the load/unload valve was moved to the off position and the pressure output
tarred by pressing the down arrow button on the pressure transducer readout. The
pressure reading on the pressure transducer readout then read 0 psi. The load/unload
valve was then moved to the unload position and the stage allowed to lower to its

lowermost position. For detailed insh'uctions on the operation of the load flame and

117

digital readouts the reader is referred to the manuals for both provided by the
manufacturer.

The middle load shaft was then lowered by turning the shaft with two fingers. The
consolidometer was manually adjusted to a cenhal position while lowering the middle
load shaft. The cenhal position was verified by lifting the loading ball slightly with two
fingers to the position in the receptacle of the middle load shaft and taking note of the
gaps between the set position in the top metal disk. The consolidometer was then moved
laterally on the stage until sufficiently centered. This centering process was performed
several times as the middle load shaft was lowered. The middle load shaft was seated on
the consolidometer by lowering the middle load shaft with two fingers until a small
resistance was felt. Again, the centering of the consolidometer was checked by attempting
to raise the loading ball from the top metal disk. Proper centering was indicated by no
movement of the ball. This configuration represented a O tsf condition during the blank
conductivity measurements. Once the middle load shaft was sufficiently seated on the
consolidometer, the load/unload valve was moved to the off position. The reservoir was
filled with 1 mM NaCl water from a squeeze bottle until the water just spilled from the
top of the effluent reservoir.

If the old load fiame was used for the blank conductivity measurements, a slight
modification of the aforementioned procedure was used. After the consolidometer was
placed on the stage, the load frame was tarred as before, the load/unload valve was moved
to the unload position and the stage was allowed to move to its lowermost position. The

load/mload valve was then moved to the off position. Then, a small metal ring was

118

placed over the loading ball of the consolidometer. At this point, conductivities were
determined as before and corresponded to the 0 tsf condition. After determination of the
conductivity the load/unload valve was moved to the load position. The low pressure
regulator was then turned slowly to slightly increase the pressure on the pressure
hansducer readout. The stage was then allowed to raise util the metal ring which was
positioned on the load ball touched the load arm and became snug. The load/unload valve
was then moved to the off position. The same procedure as described in the following
paragraphs was then used for the rest of the experiment.

Conductivity was then measured at the 0 tsf condition by the following procedure.
Temperature measurements were then taken fiom both the squeeze bottle and the effluent
reservoir and were recorded. These values were averaged to obtain the temperature value
used for correction of conductivity to 20°C. Once the temperature readings were taken,
conductivity was measured 12 times by quickly filling the standpipe with the squeeze
bottle and measuring the time of drop of water between two prerecorded reference
heights. Each time value was recorded individually. After twelve replicates were
performed, the water in the standpipe was allowed to drop to its lowermost position (i.e.
equilibrium) and this value was recorded as the datum. The procedure for measuring the
conductivity conformed to the conventional falling-head permeability test with constant
tail water level (ASTM D5084-90).

The load on the consolidometer was then increased to the next increment (0.25
tsf) by adjusting the load pressure regulator to the predetermined values prescribed by the

manufacturer while the load/unload valve was in the off position. Once the reading of the

119

pressure readout indicated the needed pressure, the load/unload valve was moved to the
load position. This applied the 0.25 tsf load to the consolidometer system. After the load
was applied to the consolidometer, the aforementioned procedure for measurement of
conductivity was repeated. Afier measurement of conductivity at that load, the
load/unload valve was moved to the off position and the pressure increased with the low
pressure regulator to the next pressure corresponding to the desired effective stress.
Conductivity measurements were taken at each of the following additional load
conditions (0.5, 1, 2, 4, and 8 tsf). Afier the measurement of conductivity at the 2 tsf
condition, the load/unload valve was moved to the off position as before. The regulator
indicator was then moved from the low to the high position. The high pressure regulator
was then used to adjust the system pressure to its desired value. While adjusting the high
pressure regulator to the desired system pressure, the low pressure regulator was loosened
to relieve the pressure behind the low pressure regulator valve. When the low pressure
regulator moved easily, the pressure was sufficiently reduced so as not to put undue stress
on the internal regulator for the course of the experiments. After hydraulic conductivity
was measured at each of the effective shesses, the load/unload valve was moved to the
unload position and the stage allowed to lower to its lowermost position.

Conductivity was calculated from each of the twelve time readings at each load.
These values were averaged to obtain an average conductivity for each load. A total of
ten experiments were performed on each set of stones for characterization, the stones
sonicated before each experiment as before. At each load, a value of hydraulic

conductivity of the stones (K.,M) was established by taking the average of the values of

120

conductivity determined from the ten experiments. An uncertainty value was also
established by taking twice the standard deviation of the conductivity values determined
from the ten experiments. The average and uncertainty were propagated through the
calculations fi'om the conductivity experiments to establish hydraulic conductivities of
the soil columns. Average hydraulic conductivity of the stones (Km) and its uncertainty
for three sets of stones, two types of filter papers, and permeated with water or Solhol are
presented in Table 1A.

After each conductivity experiment at least one experiment with new filter paper
was performed to verify the blank conductivity value assigned to the stones. Stones were
sonicated in acid solution, as before, prior to the experiment. Initially, one experiment
was performed and the average conductivity values at each effective shess were .
determined as before. The values calculated were then compared to the characterization
values from the ten initial experiments. If the average values fiom the experiment fell
within the range of conductivity values determined by the initial characterization, the
blank conductivity of the stones was deemed acceptable. The initial characterization
values were then used for the next experiment. If the blank conductivity did not fall
within the range, two additional experiments were performed with the same procedure as ,
before. The conductivities fiorn the three new experiments were then averaged to
establish a new blank conductivity value for the stones. The higher uncertainty cuculated
fiom the three experiments and the ten initial experiments was carried as the uncertainty

for the blank system.

121

 

3 3 N... 3 N... 3 a... e... 3.008
3...... am. ... 3.... 3.8.. EN. $02.. 3.2.. .88. 3....

m e N . 3 m3 .. ca. 08..
.0.=_0m 5...» 08005.0.— .3 .59... 08E .3.. .8 0:0... .0.— 000025» 8. 0:0 .52! 03.5.3.

 

 

 

 

 

3. 3 3 I. m... 3 3 .s. EVEN
Wmva. _ Wmth m-m0m.m 9mg...” Wmcmé rum—mm... ”Wm—3.0 €0.00. «nev—
.. e N . 3 m3 .. ca. 2.3
.803 5...» 08:05.0.— .5 some; .3.—u .3“ .00 2.0.0 .0.— eoaata.» 8. 05. 15v— 09035.

_.: 0.0. 2: 0.0_ 0.3 m...— ..2 03. .2..ng
Tm _ v.0 n-5, _ .m mefim m-mmm.0 «$30.0 m-mm0.0 mum—30 3.00. 33M
w v N . 3 m3 .. ca. 08..

.833 5...» 08:08.5.— .vmu 00.....— 085 .9... «on 0:02. 00. 0000...: 3. 0:0 .150— 0929?.
m0 vs m.m a... fin ”.0 0.0 §.. 30w
Wm: m... Tuna... Wmmmd Tum—no.0 Wma _ .0 m-mmm.0 m-mm0.0 380. Ed.
w v N _ m... 23 0 W3 0004

 

 

.803 5.3 08008.3.— .vma. .3.—0.. .3.—c .N... «on 0:80 000 0000...: 3. 0:0 seaw— 0wEo><

 

3:08.330 .0..=0m 0:0 .80.: a. 00...: 800... 00 305000000 0:00.03— 3: 030,—.

122

San. PACKING

Prior to assembling the consolidometer system, the stones (prelabeled and
characterized as described previously) were sonicated and dried as in the characterization
experiments. The consolidometer system was then assembled, including everything
except the top filter paper, top metal disk and porous stone, and loading ball. A reference
puck (664 gr; 6.35 cm diameter; 2.54 cm thickness) was then inserted in the
consolidometer and allowed to slide along the walls of the consolidometer ring until
seated against the bottom filter paper. The top filter paper, out to just fit in the to ring,
was then place on top of the puck, followed by the top metal disk and porous stone
assembly. The consolidometer was placed in the load frame with the load/unload valve in
the unload position and the stage at its lowermost position. Prior to placing the
conslidometer in the load fiame, the load arm was checked to verify that it was level. The
load arm was leveled using a small, two way level that could be placed on top of the load
arm. Adjustments were made by loosening the jam nuts on the sides of the load arm and
using the load arm position nuts to raise or lower the sides of the load arm until level.
Once level, the jam nuts were tightened.

The LVDT shaft was lowered by loosening the mounting wing nut holding the
shaft. The consolidometer was placed on the center of the loading stage while lowering
the LVDT shaft. The LVDT shaft was then lowered until its stem touched the top leveled
circular surface of the top metal disk and the LVDT transducer readout was
approximately 0.0100-0.0200”. The mounting wing nut was then tightened and the

LVDT reading was recorded. The consolidometer was then turned two complete rotations

123

and at each rotation four LVDT readings were recorded at perpendicular locations along
the circumference of the top leveled circular surface of the top metal disk. The average
value of these eight LVDT readings was recorded as the LVDT,,_.f reading. The LVDT
stem was then gently pushed up by hand until the consolidometer could be removed fi-orn
the load fiame.

After removing the consolidometer, the top metal disk, top porous stone, and top
filter paper were removed. The consolidometer was then inverted slightly, being careful
to place a hand over the opening of the consolidometer to catch the puck, until the puck
began to slide out of the consolidometer. After the puck was removed, the puck was
attached to the compaction assembly if the puck was used for both referencing and
packing. The compaction assembly consisted of the puck, screwed into a threaded metal
dowel, and a compaction hammer which slid along the metal dowel.

Soil packing process started with weighing 130 gr air dried Oshtemo B soil in to a
~5 V2” aluminum pie plate. The soil was then poured in the consolidometer (the
consolidometer without top filter paper, top metal disk, and top porous stone). The soil
was poured into the consolidometer by slightly bending the pie plate so as to pour the soil
over a smaller width and placing the bottom of the pie plate against the top rim of the top
ring of the consolidometer. As the soil was poured into the consolidometer, the
consolidometer was slowly turned to keep the surface of the soil in the consolidometer as
level as possible. The consolidometer was turned approximately 2 to 4 complete

rotations, depending on the speed with which the soil was poured. Once all of the soil

124

was poured into the consolidometer, the pie plate was gently tapped over the
consolidometer to ensure that all of the fines were removed.

The consolidometer was then tapped gently fiom four positions (once at each
position) to further level the soil. The compaction assembly was then carefully placed in
the consolidometer until the puck (screwed onto the end of the threaded, metal dowel)
rested on top of the soil. Both hands were then used to turn the entire compaction
assembly four times while the puck rested on the soil. One turn consisted of a
comfortable rotation achieved without removing either hand fi'om the dowel of the
compaction assembly. After four turns, the hammer of the compaction assembly (1.47 kg)
was dropped onto the puck 4 times from an upper position on the dowel (7.5 cm of drop),
marked by masking take on the dowel. The compaction assembly was then carefully
removed from the consolidometer so as not to disturb the compacted soil.

The top filter paper, porous stone and tap metal disk were thenplaced on top of
the soil in the consolidometer. Caution was used to ensure that disturbance of the upper
surface of the soil did not occur during placement. The consolidometer assembly was
then placed in the load frame, again with the load/unload valve in the unload position and
the stage at the lowermost position. The LVDT stem was then pushed up and the
consolidometer was slid under the LVDT shaft such that the LVDT stem seated on the
top leveled circular surface of the top metal disk. The consolidometer was then turned
two complete rotations, as before, and at each rotation four LVDT readings were recorded
at perpendicular locations along the circumference of the top leveled circular surface of

the top metal disk. The average value of these eight LVDT readings was recorded as the

125

LVDTpack reading. The LVDT stem was then gently pushed up by hand until the
consolidometer could be removed freely from the load frame.

The consolidometer was removed fi'om the load flame and was placed in the
vacuum dessicator. The vacuum system was sealed and the water deaired under vacuum
for at least one hour. After the water was deaired, the water was brought into the
dessicator until the level was just below the lip of the consolidoeter (water level 2, Figure
2A). The sample was then allowed to saturate for at least twelve hours (overnight) under
vacuum. Vacuum saturation was verified after the saturation period by observation of
water in the consolidometer reservoir above or around the top metal disk. At the end of
saturation, water was allowed to enter the vacuum dessicator until the water level was
above the rim of the consolidometer (water level 3, Figure 2A). The vacuum was then
released and the top of the dessicator removed. While the consolidometer was still under
water, the standpipe was attached to ensure that no air entered the consolidometer system.
Any residual air which was left in the fittings or the canal of the consolidometer was then
removed by applying a small vacuum on the standpipe while the plug was not screwed in.
Water was brought into the standpipe at least four times and until no air was observed
escaping fi'om the standpipe during application of the vacuum. The plug on the
consolidometer base was then screwed in and secured. For more details on the vacuum
saturation procedure see the Vacuum Saturation section in this appendix.

LOADINGAND Commons/In MEASUREMEMS
After the consolidometer was taken out of the dessicator with the standpipe and

the plug attached to it, the external parts of the consolidometer were dried with paper

126

towels. The consolidometer was then carefully placed in the load frame so as not to hit
the standpipe or the LVDT shaft. The consolidometer was oriented so that the standpipe
was just in fi'ont of the left load arm support. As before, the load/unload valve should
have been in the unload position and the stage at its lowermost position. The LVDT stem
was then pushed up and the consolidometer was slid under the LVDT shafi such that the
LVDT stem seated as before on the top leveled circular surface of the top metal disk. The
LVDT reading fiom the displacement readout was then recorded. The consolidometer
was carefully turned such that two more LVDT readings were recorded at locations along
the circumference of the top leveled circular surface of the top metal disk. The process of
LVDT reading was duplicated such that total of six LVDT readings were made. The
average value of these six LVDT readings was recorded as the LVDT“, reading. The
LVDT shaft was then raised such that the stem no longer rested on the top metal disk and
the consolidometer was taken out of the load flame. The load/unload valve was then
moved to the off position and the regulator selector moved to the low position. The low
pressure regulator was then adjusted until a negative pressure readout was observed.

The loading ball was then placed on the top metal disk and the consolidometer
was placed in a flat plastic bowl which served to catch the water (or Soltrol) spill from
the effluent reservoir of the consolidometer. The consolidometer (with the bowl) was then
placed back in the load frame. The load frame was then tarred by moving the load/unload
valve to the load position, slowly increasing the pressure with the low pressure regulator,
and observing the stage movement relative to a reference line marked on the stage shaft

beneath the stage. The low pressure regulator was then adjusted until no movement was

127

observed, relative to the reference line on the stage shafi. Once no movement was
observed, the load/unload valve was moved to the off position and the pressure output
tarred by pressing the down arrow button on the pressure transducer readout. The
pressure reading on the pressure transducer readout then read 0 psi.

The loading arm was then lowered by loosening the jam nuts on the sides of the
arm and using the loading arm position nuts. The consolidometer was manually adjusted
to a central position while lowering the loading arm. The central position was verified by
lifting the loading ball slightly with two fingers to the position in the receptacle of the
loading arm and taking note of the gaps between the set position in the upper metal disk.
The consolidometer was then moved laterally on the stage until sufficiently centered.
This centering process was performed several times as the loading arm was lowered. The
loading arm was seated on the consolidometer by loosening the loading arm position nuts
until a small resistance was felt. Again, the centering of the consolidometer was checked
by attempting to raise the loading ball fi'om the top metal disk. Proper centering was
indicated by no movement of the ball. This configuration represented a O tsf condition.
Prior to tightening the jam nuts, the load arm was checked to verify that it was level. The
load arm was leveled using a small, two way level that could be placed on top of the load .
arm. Adjustments were made by loosening the jam nuts on the sides of the load arm and
using the load arm position nuts to raise or lower the sides of the load arm until level.
Once level and sufficiently seated on the consolidometer, the jam nuts were tightened.
The LVDT shaft was then lowered until the stem rested on the locking nut of the

consolidometer and the LVDT reading was 0.0100~0/0200”. The mounting wing nut of

128

the LVDT shafi was then tightened and the displacement readout was recorded as the
LVDT of the sample at O tsf. The reservoir was filled with 1 mM NaCl water fiom a
squeeze bottle until the water just spilled fiom the effluent reservoir.

Conductivity was determined at O tsf by falling head (ASTM D5084-90). Prior to
measurement, temperatures of both the effluent reservoir and the permeant to be used in
the test were determined with a conventional thermometer and recorded. Water ( 1 mM
NaCl) or Soltrol, that had been deaired for at least two hours under a vacuum of
approximately 900 mbars, was used as the permeant. Then, 4 falling head measurements
were performed by filling the standpipe with the squeeze bottle and measuring the time
for the standpipe fluid level to drop approximately 12 cm. Each time was recorded
individually. After the four tests, the fluid level in the standpipe was allowed to drop to
its equilibrium level. This level was recorded as the datum at this load.

After the datum was recorded for the 0 tsf load and with the load/unload valve on
off position, the low pressure regulator was used to increase the pressure to the pressure
reading corresponding to the 0.25 tsf efl'ective stress condition. The load/unload valve
was then turned to the load position and the sample allowed to consolidate under the load.
The consolidation procedure was a modified form of the conventional consolidation test
(ASTM D2435-90) and was justified by Grant (thesis 1995). The samples were allowed
to consolidate under each load for approximately 1 hour, after which the LVDT reading
was recorded and the conductivity and temperature were measured using the
aforementioned procedures. The pressure was then increased to the next pressure

corresponding to the next effective stress state, allowing the sample to consolidate under

129

that load. The samples were consolidated incrementally to 0.25, 0.5, l, 2, 4, and 8 tsf
stress states using progressively larger loads as described in Blank Conductivity
Measurements section. After the measurement of conductivity at the 2 tsf condition, the
load/unload valve was moved to the off position as before. The regulator indicator was
then moved fi'om the low to the high position. The high pressure regulator was then used
to adjust the system pressure to its desired value. After hydraulic conductivity,
temperatures and LVDT were measured at 8 tsf effective stress, the experiment was
concluded and the load/unload valve was moved to the unload position allowing the stage
to lower to its lowermost position.
CONSOLIDOMEIER DISASSEMBL Y

Once the load/unload valve was moved to the unload position and the stage was
moved to its lowermost position, the LVDT shaft was loosened and moved up to a
position that allowed free movement of the consolidometer without the possibility of the
consolidometer hitting the LVDT shaft. Both pressure regulators on the load fi'ame were
then loosened until a negative pressure was observed on the pressure readout when the
regulator selector was in either the high or the low position. Then the consolidometer was
removed from the load flame and taken to the sink for cleaning. The plug on the
consolidometer base was unscrewed and allowed to fieely drain. The standpipe was then
removed , rinsed, and allowed to air dry.

The loading ball was then removed from the top metal disk, rinsed, and allowed to
dry. The top metal disk and porous stone were then carefitlly removed and were

separated. The top metal disk was rinsed and dried. The top porous stone was then rinsed

130

thoroughly to remove any fines particle on its surface (if any). Upon removing and
cleaning the top metal disk and porous stone, the locking nuts of the consolidometer were
loosened and removed. The seal between the base plate and top plate of the
consolidometer (created by the vacuum grease) was then broken by prying the top plate
with a spatula. After the seal was broken, the top plate of the consolidometer was
carefully removed. The consolidometer ring containing the soil column, the bottom filter
paper, and the O-ring sometimes remained connected to the top plate, so that removing
the top plate also removed these pieces. In this case, caution was used so as not to let the
consolidometer ring fall on the floor accidentally.

The soil sample and filter papers (top and bottom) were visually checked for any
unusual segregation of particles or discoloration and, upon not finding any, were
subsequently discarded. The O-ring was then removed from around the consolidometer
ring, cleaned, and dried. The remaining pieces of the consolidometer were then rinsed
thoroughly with building DI water to remove residual soil particles and then were cleaned
with paper towels to remove vacuum grease. The cleaned pieces were then placed in a
storage area prior to the next experiment. The bottom porous stone was rinsed completely
to remove any particulate and then was placed, along with the top porous stone, in a 500
ml beaker. The stones were then covered with 0.01 N HCl solution and were sonicated
for 12 minutes. Subsequently, the stones were thoroughly rinsed with tap water, dried,

and stored prior to further testing.

131

CALCULA n0Ns AND DA TA REDUCTION
All hydraulic conductivities determined by falling head measurements (blank and

total conductivities) were calculated using the following equation:

a.L.C, 10g (hl-hdm)
A-t ' (hz-hdaz)

K : hydraulic conductivity, [L/T]

K = 2.303

a : cross sectional area of the stand pipe, [L2]

L : total length through which flow occured, [L]

Ct: correction factor to adjust the conductivity to 20° C, []

A : cross sectional area of the sample, [L2]

t : time for the fluid drop from height h 1 to h 2, [T]

h 1: height of the initial hydraulic head, [L]

h 2: height of the final hydraulic head, [L]

hdati height of the datum at equilibrium, [L]

For the consolidometers used in this study, the cross sectional area of the stand
pipe was measured to be 0.335 cm2 and the cross sectional area of the consolidometer
ring (which was the same as the cross sectional area of the soil sample) was 31.7 cmz.
The total length through which flow occured (L) including the soil sample, porous stones,
and filter papers in soil conductivity tests, and only the porous stones and filter papers in
blank conductivity measurements. The coefficient 2.303 appears in the above equation to
account for the use of “log” rather than “In” in the equation. In all cases the hydraulic
conductivity calculated here represented either Ktotal if soil was in the consolidometer, or

Kblank if soil was not used.

132

It was found experimentally that the conductivity of the stones and filter papers,
Kblank, exerted some influence on the hydraulic conductivity of the system especially at
low loads where conductivity of soils were relatively high. Therefore, it was necessary to
separate the effects of the blank system to accurately determine the hydraulic
conductivity of the soil column. Based on this, when conductivities of the porous stones
and filter papers were being measured, the hydraulic conductivity was labeled as Kblank-
When hydraulic conductivity was measured with the soil sample packed in the

consolidometer, the calculated conductivity represented conductivity of the total system,

 

including the soil, stones and filter papers, and therefore was labeled as Ktotal-

In order to separate the effects of porous stones and the filter papers fi'om the
measured total conductivities, the equation for a layered soil system was employed.
Consequently, the hydraulic conductivity of the soil sample was calculated using the

following equation, derived fiom a layered soil system (Freeze and Cherry, 1979):

Lsog - 190g - Kbla_n_k
[( Ltotal' K blank ) - (Ktotal - Lblank )]

Ksoil =
where K3051 was the conductivity of the soil sample under various effective stresses, L 50,-]
was the length of the soil sample at the time of conductivity measurement, Ktotal was the
hydraulic conductivity of the entire system, Kblank was the hydraulic conductivity of the
blank, and Ltotal was the length of the stones and filter papers.

Lengths of the system and the soil were based on LVDT readings taken at various
times during the experiment. The length of the porous stones and filter papers (i.e.

Lblank) was assumed to remain constant throughout the experiments. Lblank was

determined to be 1.618 cm, based on measurements of the thicknesses of stones with

 

133

caliper and the manufacturer standards for the filter papers. The length of the soil prior to
any consolidation was based on the reference LVDT reading (LVDTmf) taken with the
machined brass puck in the consolidometer assembly and the LVDT readings taken after
packing (LVDng or after sample saturation (LVDTm). The length of the soil sample
and the length of the entire system (cm) were calculated from the following equations:
Lsoll( pack, sat) = 2.54.[1 + (LVDTMM - LVDT"; )] '

Lm:( pack, sat) = Lsoil( pack, sat) + 1.618

Based on the method used to determine the lengths of the system, the length of the

 

soil sample (and entire system) at O tsf was equal to the corresponding length after
saturation, determined from equations above. Therefore, the length of the soil or entire
system at any desired point was determined fiom the LVDT reading at that point
compared to the LVDT at 0 tsf (recall that the LVDT shaft was moved afier recording
LVDT", and was therefore different fi'om the LVDT at O tsf). The following equations
were used to calculate the length of the soil and lengths of the entire system (cm) at any
time during the experiments:
Lsoil( Otsf) = Lsou(sat)
Lsoil(load) = Lsoil(0tSf) - [254.(LVDTIoad - LVDTW )]
Lmz(load) = Lmu(load) + 1.618

Porosity of the soil sample was calculated under each load. It was calculated

based on the length of the soil sample using the following equation:

134

M
A.Lsoi '-
n = E =( I) G: - MS

_ 1 .. ——
V: (A. Lsoil) Gs. A. Lsoil

 

where n was the porosity (obviously without units), Vv was the total volume of the pore
space [L3], Vt was the total volume of the soil sample [L3], A was the area of the sample
(31.7 cm’), Lu,“ was the length of the soil sample in the direction of the flow [L], Ms was
the mass of solids used in the experiment (130.0 gr), and GS was the specific gravity of
the soil solids (measured to be 2.65).

Uncertainty was propagated through all calculations based on statistical analysis.

 

This propagation, while straightforward in its application, involved relatively complex
equations. Therefore, the exact equations used for each variable is not mentioned here.
The reader is referred to calculation sheets included in records kept during the

experiments and to the spreadsheets used in the computations.

   

"‘111111111111111: