ACKNOWLEDGEMENTS. I would like to extend my sincere appreciation to my thesis advisor Dr. Gayle

THE USE OF TIME DOMAIN REFLECTOMETRY (TDR) TO DETERMINE AND MONITOR NON-AQUEOUS PHASE LIQUIDS (NAPLS) IN SOILS A Thesis presented to The Faculty of t...
Author: Geoffrey Ellis
9 downloads 0 Views 5MB Size
THE USE OF TIME DOMAIN REFLECTOMETRY (TDR) TO DETERMINE AND MONITOR NON-AQUEOUS PHASE LIQUIDS (NAPLS) IN SOILS

A Thesis presented to The Faculty of the Fritz J. and Dolores H. Russ College of Engineering and Technology

Ohio University

In Partial Fulfillment of the Requirements for the Degree Master of Science

by Nabil M. Quafisheh March, 1997

ACKNOWLEDGEMENTS

I would like to extend my sincere appreciation to my thesis advisor Dr. Gayle Mitchell for her guidance, comments and suggestions regarding this thesis. Special thanks to Dr. Shad Sargand and Dr. Larry Snyder for serving on my thesis committee.

I cannot give enough thanks to my parents, brother and sisters, whose emotional support through the years have made the completion of this thesis possible.

11

TABLE OF CONTENTS

ACKNOWLEDGMENTS TABLE OF CONTENTS

11

LIST OF FIGURES

v

LIST OF TABLES

viii

NOTATION

IX

CHAPTER 1: INTRODUCTION 1.1

Scope

1

1.2

Objectives

4

1.3

Summary

4

CHAPTER 2: LITERATURE REVIEW 2.1

Principles of TDR

6

2.2

Dielectric Constant and Water Content

8

2.3

Electrical Conductivity

10

2.3.1

Determination of Electrical Conductivity by TDR

11

2.4

Solute Transport

13

2.5

Calibration for Solute Transport

19

2.5.1

Direct Approach

19

2.5.2

Indirect Approach

19

III

2.5.2.1 Vertically Installed Probes

19

2.5.2.2 Horizontally Installed Probes

19

2.6

Definition of NAPLs

21

2.7

Sources of NAPLs

22

2.8

Migration of NAPLs in Subsurface

22

2.9

Properties of NAPLs

24

2.9.1

Saturation

24

2.9.2

Interfacial Tension

25

2.9.3

Wettability

25

2.9.4

Capillary Pressure

25

2.9.5

Residual Saturation

26

2.9.6

Solubility

26

2.9.7

Volatilization

27

2.10

2.11

2.9.8 Viscosity

28

Methods of Measuring NAPL in Subsurface

28

2.10.1 Soil Gas Analysis

29

2.10.2 Drilling Investigation

29

Application of TDR to Monitor NAPLs in the Subsurface

30

CHAPTER 3: EXPERIMENTAL SETUP AND PROCEDURES 3.1

Introduction

35

3.2

Data Acquisition

35

IV

3.3

Calibration of TDR Probe

36

3.4

Experimental Procedure of Phase I

37

3.5

Experimental Procedure of Phase II

41

3.6

Clean-up and Disposal of Waste

45

CHAPTER 4: DISCUSSION OF RESULTS 4.1

Dielectric Constant Results in Phase I

46

4.2

Summary and Discussion of Results of Phase I

56

4.3

The Results of Phase II

63

4.4

Summary and Discussion of the Results of Phase II

67

CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS 5.1

Conclusions

75

5.2

Limitations

78

5.3

Recommendations

78

BIBLIOGRAPHY

ABSTRACT

80

v

LIST OF FIGURES

2.1.1

Schematic diagram of a typical signal of the TDR showing the travel time

7

2.3.1

TDR trace showing the locations of the four significant voltages

12

2.4.1

Schematic of the TDR trace showing the location of the measurement of the impedance load, RL

17

2.10.1 Results of Redman's LNAPL experiment

31

2.10.2 Results of Redman's DNAPL experiment

34

3.3.1

Calibration curve of TDR for fine sand

38

3.3.2

Calibration curve of TDR for coarse sand

38

3.5.1

System setup for phase II

42

3.5.2

A photograph of the system setup for phase II

43

4.1.1

The results of fine sand with gasoline at different water content

47

4.1.2

The results of fine sand with diesel at different water content

47

4.1.3

The results of fine sand with PCE at different water content

48

4.1.4

The results of coarse sand with gasoline at different water content

48

4.1.5

The results of coarse sand with diesel at different water content

49

4.1.6

The results of coarse sand with PCE at different water content

49

4.1.7

Different NAPLs with fine sand at zero water content

51

4.1.8

Different NAPLs with fine sand at 5.6% water content

51

4.1.9

Different NAPLs with fine sand at 11% water content

52

4.1.10 Different NAPLs with fine sand at 17-20% water content

52

VI

4.1.11 Different NAPLs with fine sand at saturation

53

4.1.12 Different NAPLs with coarse sand at zero water content

53

4.1.13 Different NAPLs with coarse sand at 5.6% water content

54

4.1.14 Different NAPLs with coarse sand at 11% water content

54

4.1.15 Different NAPLs with coarse sand at 17-20% water content

55

4.1.16 Different NAPLs with coarse sand at saturation

55

4.1.17 Silty sand (10%) with gasoline at different water content

57

4.1.18 Silty sand (30%) with gasoline at different water content

57

4.1.19 Silty sand (50%) with gasoline at different water content

58

4.2.1

The value of VnN s when the dielectric constant change for different NAPLs with fine sand

59

The value of VnN s when the dielectric constant change for different NAPLs with coarse sand

59

4.2.3

Comparison between fine and coarse sand at zero water content

61

4.2.4

Comparison between fine and coarse sand at 5.6% water content

61

4.2.2

4.2.5 Comparison between fine and coarse sand at saturation 4.3.1

62

Dielectric constant versus time after the addition of diesel to coarse sand

64

Dielectric constant versus time after the addition of PCE to coarse sand

64

4.3.3

Dielectric constant versus time after the addition of diesel to fine sand

66

4.3.4

Dielectric constant versus time after the addition of PCE to fine sand

66

4.4.1 TDR water content after the end of each spill of diesel to coarse sand

68

4.3.2

4.4.2

TDR water content after the end of each spill of PCE to coarse sand

68

VII

4.4.3

TDR water content after the end of each spill of diesel to fine sand

69

4.4.4 TDR water content after the end of each spill of PCE to fine sand

69

4.4.5 Percent of water displaced in the coarse sand with diesel experiment

72

4.4.6 Percent of water displaced in the coarse sand with PCE experiment

72

4.4.7 Percent of water displaced in the fine sand with diesel experiment

73

4.4.8

Percent of water displaced in the fine sand with peE experiment

73

VIII

LIST OF TABLES

3.4.1

Properties of NAPLs used in this study

39

3.4.2

Media, contaminants and water content used in phase I

40

3.5.1

Media and contaminants used in phase II

43

IX

NOTATION

a.

Attenuation coefficient

y

Cell constant

E

Relative dielectric constant of the medium

E'

Real part of the relative permittivity or dielectric constant

E"

Imaginary part or dielectric loss

Eo

The permittivity of vacume

e

Volumetric water content, m3/m3

Jl

Medium's relative magnetic permeability, henrys/m

Jlo

The magnetic permeability in free space

p

The voltage reflection constant

o

The zero frequency conductivity, siemens/s

c,

Soil bulk electrical conductivity, siemens/s

cr w

Electrical conductivity of the soil solution, siemens/s

u

Propagation velocity of the TDR pulse, mls

ill

The angular frequency, radians/s

c

Velocity of light in free space, mls

C

Pore-water concentration, kg m"

I

Length of the wave guide, meter

L

Depth of soil, meter

M

Total mass of the solute per unit area, Kg m- 2

x RL

Load resistance (impedance),

Ro

The initial impedance load,

o

n

RL(tO) The measured impedance load after the solute has been applied but before any has moved past depth L, Q

RL(t j)

The measured impedance load just before the tracer pulse is applied,

t

Measured transit time, seconds

v

0

Q

Amplitude of the TDR pulse, volts

V1

Amplitude of the signal after partial reflection from the start of the probe, volts

V2

Amplitude of the signal after reflection from the end of the probe, volts

Vf

Amplitude of the reflected signal after a very long time, volts

1

CHAPTER 1

INTRODUCTION

1.1

Scope Fellner-Feldegg [1969] introduced time domain reflectometry (TDR) as a

means of measuring the complex dielectric permittivity of liquids. Since that time, TDR has been applied to the measurements of dielectric properties of many materials including soils. Since 1980, time domain reflectometry has become widely used to measure water content of soils. With TDR, dielectric soil properties are measured and correlated with soil water content and its ionic concentration. TDR is still being evaluated, and much literature is available on its principles, calibration, and practical use.

The need for rapid, reliable, and routine techniques (the three R's of field measurement) for monitoring in situ volumetric soil water content, soil electrical conductivity and organic contaminants in the fields of hydrology, agriculture, site remediation, and various areas of civil engineering is obvious and needs no elaboration (Zegeline et aI., 1989). Time domain reflectometry (TDR) appears to have the potential to provide these three R's, particularly for determination of water content, electrical conductivity, and perhaps organic contaminants as well [Dalton et aI., 1984;Dasberg and Dalton, 1985; Topp et aI., 1988; Redman and DeRyck, 1994].

2 Environmental waste site investigations have needs that time domain reflectometry technology are capable of meeting. Three basics steps are followed in waste investigations (Knowlton et aI., 1994): (1) characterization or assessment of the site to determine if it posseses a possible threat to human health and the environment,

(2) identification of cleaning alternatives, if the site poses an unacceptable threat, and (3) implementation of a remedy, with possible post-closure monitoring of the performance of the remedy.

In the assessment phase, hydrological information is needed to establish flow and transport properties, as well as to monitor the transient behavior of contaminant movement beneath a waste site. These data are ultimately used in a risk assessment of the site to determine if a possible threat exists. TDR technologies offer several solutions to these needs. For example, during post-closure monitoring, TDR can present unique capabilities for evaluating the performance of a landfill cap in reducing infiltration.

Essential to any assessment and remediation efforts is a thorough understanding of local geologic and hydrologic conditions and the expected behavior of a contaminant under various physical influences (Illangasekane et aI., 1995). A majority of the research in this area has focused on development of mathematical models that are used in remediation design. Many modeling efforts incorporate innovative approaches and complex numerical schemes in the prediction of transport and behavior

of solutes and immiscible bulk-phase organics in both saturated and unsaturated zones of aquifers. However, limited attempts to evaluate models based on experimental data have appeared in the literature. To a large extent, these efforts were dependent on qualitative measures and their usefulness in terms of model validation or calibration under heterogenous aquifer conditions is limited (Illangasekare, et.al., 1995).

In the past few years, as hazardous waste sites have been studied more often and in more detail, immiscible fluids have been encountered in the subsurface with greater frequency (Mercer et al., 1990). These Non-Aqueous Phase Liquids (NAPLs) behave differently than dissolved solutes in the subsurface. Their behavior depends on fluid properties such as interfacial tension, viscosity and density. Nonaqueous phase liquid (NAPL) is a term used to describe a class of organic liquid contaminants that incorporates both DNAPLs (dense NAPLs) that are more dense than water and LNAPLs (light NAPL) that are less dense than water (Redman et al., 1994). Although NAPLs are considered immiscible with water, large plumes of dissolved-phase ground water contamination are produced because NAPLs have solubilities that are well above the drinking water standards. Typical examples of DNAPLs are the chlorinated solvents, which make up one of the largest classes of DNAPL; other DNAPLs of concern include PCB oils, coal tars, and creosote. Typical LNAPLs are the hydrocarbon fuels (eg., gasoline, and kerosene).

Typically, chlorinated solvents have a higher density than water, a low

4

solubility in water, low viscosity, high volatility, and are mechanically immiscible in water. This combination of properties allows such DNAPLs to exist as a distinct liquid phase within the subsurface where they may give rise to evolving dissolved phase plumes for several decades or more, before they are depleted through natural dissolution. To remediate sites contaminated with NAPLs, some important factors must be considered such as locating the contaminants in the subsurface and understanding the effectiveness of remediation efforts.

1.2 Objectives The objectives of this research are as follows:

* Study the potential

use(s) of Time Domain Reflectometry (TDR) to monitor Non-

Aqueous Phase Liquids (NAPLs) in the subsurface

* determine

the effect of water content on the monitoring of NAPLs with TDR

* investigate the effect of soil types

on the monitoring of NAPLs with TDR

* investigate the effect of the presence

of silt in soils in the monitoring of NAPLs

with TDR

* study

the difference in behavior of DNAPLs and LNAPLs in the subsurface with

TDR.

1.3 Summary TDR is short for Time Domain Reflectometry, a technique which has been

5 used since 1980 to measure water content and electrical conductivity of soils. In this study, the principles of TDR are discussed. Also, the applications of TDR such as, the determination of water content, electrical conductivity, solute transport and NAPLs migration are discussed.

The experimentation strategy of this research emphasizes the study of the potential use(s) of TDR to monitor NAPLs migration in subsurface. This is achieved by using different soils and types of NAPLs at different water contents.

6

CHAPTER 2

LITERATURE REVIEW

2.1 Principles of TDR Time Domain Reflectometry (TDR) is now well accepted as a method for measurement of soil water content. The soil water content is obtained from the measurement of velocity of propagation of high-frequency electromagnetic pulse signals which can be determined from the time axis of the TDR trace. Recent studies have proposed and promoted the simultaneous measurement of both water content and electrical conductivity of soil using TDR. The determination of electrical conductivity is possible by measuring the attenuation of the reflected signal, making use of the reflection coefficient or amplitude axis of the TDR trace.

At the end of each TDR wave guide, the launched electromagnetic pulse is reflected back to its source (see Figure 2.1.1). Therefore, the pathlength is twice the length of the wave guide, / (in meters), and the measured transit time, t (in seconds), gives the propagation velocity of the pulse (in meters per seconds) (Dalton et al., 1984): v

= 2//t

(1)

if dispersion is negligible, then v can be given simply in terms of relative dielectric

7

~ (5

>

oj" Ol

ro o

+J

5

c: o :.;:;

(I -J

\

o

Q)

c;:::

Q)

a::: Q)

> :.;:;

ro

Q)

a:::

~ t

Time- (nsec)

Figure 2.1.1 Schematic diagram of a typical signal of the TOR showing the travel time ( Dalton et aI., 1984)

8 constant of the medium,

E,

and the velocity of light in free space, c (meters per

seconds)

v ==

1 C/E / 2

(2)

Therefore, the relative dielectric constant is given by: E

= (Ct/2l)2

(3)

Since the dielectric constant of water is much greater than that of the air and the soil grains, the presence of water should be easily detectable.

2.2 Dielectric Constant and Water Content The dielectric characteristics of a material can be represented by a complex dielectric or relative permittivity,

E,

by the relation (Ledieu et aI., 1986):

E = E' - j(E"+cr/Eoro) where:

E' =

the real part of the relative permittivity;

E"=

the imaginary part or dielectric loss;

(4)

o == the zero frequency conductivity (in siemens/m); Eo = 0)

j

the permittivity in free space (in farads/m);

== the angular frequency (in radians/sec);

==

v-I, the imaginary number.

The permittivity expresses the polarization of the material submitted to an electrical field. The polarization tends to reduce the field strength. Two main classes of phenomena cause the polarization induced by an electric field (Ledieu et aI., 1986): (1) an electronic, atomic and molecular distortion of non-polar molecules, causing an

9 indirect dipolar moment in the direction of the field; and (2) a rotation of dipolar elements, tending to align them with the field. In the imaginary part of E, the term

E"

represents the dielectric losses due to the

vibration or rotation of the molecules. On the other hand the term

0'/8 000

is

characteristic of materials containing free charges and represents the losses due to conductivity. The imaginary part of E results in a phase shift between the applied electric field and the resulting polarization (Ledieu et al., 1986). At low frequencies the ions can move to the limits of the conducting microdomains to create an extra polarization, which can be greater than the original dielectric permittivity. At high frequencies the term

a/EoO)

becomes negligible.

For water, and at high frequencies frequency and also be represented as

E.

E"«

In this case

E'.

E'

E

is no longer strongly dependent on

is nearly real and

constant~

so that it can

For water this assumption is valid between 1001tIHz and 3 or

4GHz (Ledieu et al., 1986).

The empirical relationship between relative dielectric constant,

E,

and soil

volumetric water content, 8, was found to be (Topp et al., 1980):

e = -5.3 X

10-2 + 2.92 X 10-28

5.5 X 10-48 2 + 4.3 X 10-68 3

-

(5)

It was found that this relationship was nearly independent of soil texture, soil density, temperature and salt content. It was noted, however, that soil salinity did influence

10

the TDR signal in the soil medium.

2.3 Electrical Conductivity Soil may conduct current through (Nadler et aI., 1980): (1) the interstitial water which contains dissolved electrolytes and (2) via the exchangeable cations that reside near the surfaces of charged soil particles and are electrically mobil to various extent. The relative contributions of exchangeable cations to electrical conduction is small at high solution concentrations. However, at low concentration they may play an important role in determining the bulk soil electrical conductivity.

In addition, the actual soil conductivity depends on the water content, chemical composition of soil solution and exchangeable ions, percent clay in the soil, and the interaction between the bulk and exchangeable ions. [t is assumed that the specific electrical conductivity of soil containing dissolved electrolytes (salts) in the soil solution can be represented by a conductance model consisting of three elements in parallel (Rhoades et aI., 1989): (1) conductance through alternating layers of soil particles and interstitial soil

solution(a solid-liquid series-coupled element), (2) conductance through or along the surface of the soil particles (primarily associated with exchangeable cations) in direct contact with one another (a solid element),and (3) conductance through the interstitial soil solution (a liquid element).

11 It was found that the soil bulk electrical conductivity,

O'a'

is related to the

impedance by (Kachanoski et al., 1994): (6) where R is the load resistance (n), which is essentially equal to the load impedance at low frequencies or larger times, and y is a cell constant that depends on probe geometry.

2.3.1 Determination of electrical conductivity by TDR Figure 2.3.1.1 shows the four significant voltage values that are routinely measured on a TDR trace that appears as an output on the screen. Yo, VI' V 2, and V r, represent amplitudes of the TDR pulse, signal after partial reflection from the start of the probe, signal after reflection from the end of the probe, and reflected signal after a very long time (t--> (0), respectively (Nadler et aI., 1991).

According to electromagnetic field theory, the amplitude of a signal traveling a distance 2/ in an electrically conductive medium with an attenuation coefficient,

U,

diminished exponentially according to (Dalton et aI., 1984): (VI -V2)

= VI exp(- 2af)

(7)

where the attenuation coefficient,o, is approximated by: a =

where:

c,

=

(ja

/2 (Jlf.lo/EEo)

(8)

the electrical conductivity of the medium, siemens per meter

Jl = the medium's relative magnetic permeability, henrys per meter

is

13 Eo==

the dielectric constant in free space, farads per meter

u,= the magnetic permeability in free space, henrys per meter. The medium electrical conductivity can be presented as : (ja

= E 1/2/( 120nl) In(V I/(V 1-V2) )

(9)

where V I (Figure 2.3.1.1) is the magnitude of the signal that enters the TDR probe and V2 is the magnitude of the reflected signal. Later work showed that, especially with media of low conductance, multiple reflections occur. Dalton et al. (1984) used V Ie-2ul as an approximation for the reflected pulse, while a direct consideration of the reflected pulse after one round trip was suggested by Topp et al. (1988), who obtained an approximation of O'a by using the expression VI + (V2-Vl)e-2al for the reflected pulse magnitude, a being the attenuation coefficient resulting, as shown by Zegeline et al. (1989), in (10)

2.4 Solute Transport Measurement of R (load impedance or resistance, Q), obtained by TDR, depends on the soil water content and electrical conductivity of the soil solution (the soil not included), crw' for constant volumetric water content, 8, and solute concentration levels of practical concern. For relatively low concentration of an electrolyte, the value of o., is linearly related to the concentration of the electrolyte in the water. Thus, at constant 8, there should be a linear relationship between average pore-water concentration, C, of a particular solute and

O'a.

The major task in any study

14 of solute transport is the estimation of the calibration relationship. The method of obtaining the calibration, as well as the choice of dispersion model, varies with probe orientation, although under certain boundary conditions, the need for a calibration can be eliminated from the analysis.

Assuming that a linear relationship exists between c, and the average porewater concentration, C (Kg m'), of a particular solute at constant water content, 8, then o, can be given by:

era where

l(

and

=

KC +

f3

(11 )

p are empirical constants.

For a particular depth of soil, L, the total mass of the solute per unit area (Kg m"), M L , can be given by (Kachanoski et aI., 1992):

M L = CL 8 L L

(12)

CL = the average solute concentration for soil depth L, Kg m"

where

8L

=

the average volumetric water content for soil depth L, rrr' m"

Thus, the c, becomes: (13)

where

O'a.L

is the TDR estimate of c, for the soil depth L. Therefore, at constant 8

there is a linear relationship between the specific mass (Kg m") of the solute and the "[DR measured

O'a"

15

Consid er a particu lar soil colum n or field plot, where a consta nt surface flux density of water, q (rrr' m-2

S·I),

is applied. After steady state (dS/dt = 0, t=tJ has been

attained within the measu remen t depth, L, a pulse of solute (electr olyte) with a total specific mass MT,L (Kg m') is applied at time t

= to. According to the equation above,

as long as all of the specific mass of the solute applied is within L, the value of MT,L can be given by (Kach anoski et aI., 1992): MT,L = (Le/K)[cr~L(to) - cr~L(ti)]

(14)

where cra,L(to) is the TDR estima te of the electrical condu ctivity after the pulse of solute has been applied, but before any of the solute moves past L, and cra,l,(tJ is the TDR estimate of the electrical conductivity before the solute pulse has been applied but after steady state. In other words, the total specific mass of the solute tracer applie d is equate d with the difference in O"a,L before and after the solute pulse is applied. Since the specific mass of the solute applie d is known, the value a can be directly obtain ed from the readin gs of c, before and after the pulse is applied.

As the solute tracer moves throug h the soil and starts to disperse, the measu red value of craL should not change as long as the total mass of solute is still within L. As the solute moves past L, the value of (ja,L will decrease. The specifi c mass of solute tracer ML(t) remain ing within depth, L, at any time, t, can be given by (Kachanoski et aI., 1992): (15) where aa,L(t) is the TDR estimate of c, at any time t (t > to). The relative specific

16

mass of solute MR,L(t)

==

ML(t)IMT,L remaining within the measurement depth is

obtained by dividing the previous two equations: (16)

The equation above gives only a TDR estimate of the electrical conductivity and no calibration constants.

In addition, a TDR estimate can be obtained from the impedance load, RL(Q), of the TDR probe measured at a fixed distance along the TDR trace after multiple reflections. Impedance is the total opposition to flow of electrical energy in the transmission line. It is composed partly of the reactance (opposition to alternating current, AC) and partly of the resistance (opposition to direct current, DC). At zero or low frequencies, RL is dependent only on the DC component and is equal to the resistance. Since the lowest frequency of the TDR signal is associated with the longest travel time, R L should be measured at the longest time. Figure 2.4.1 shows a typical trace from the TDR and the location used in measuring RL • The value of RL is obtained from (Kachanoski et al., 1992) (17)

where p is the voltage reflection constant, related to the ratio of the reflected-wave amplitude to the TDR input-wave amplitude, and Ro is the initial impedance load.

17

---,

!

_I~\

/._-----/

\ \, ~

/

»;

/

/

Measurement location for RL

~

Time (nsec)

Figure 2.4.1 Schematic of TOR trace showing the location of the measurement of the impedance load, RL (Kachanoski et aI., 1992)

18

In this study, a Tecktronix cable tester (Model 1502B Tecktronix, Beaverton, OR) was used, which has a built-in menu option that automatically calculates the value of RL, at the location where the trace intersects the screen cursor line, and then displays the value in the upper right comer of the screen. Thus, once the cursor line is set at the selected measurement time, t m , along the trace, the value of RL can be recorded with time directly from the screen. The value of O'a is estimated from o a = KRL -1

(18)

where K is a calibration constant. Thus, the specific mass of the solute applied within length, L, is: (19) where: RL(to) = the measured impedance load after the solute has been applied but before any has moved past depth L, Q RL(ti )

=

the measured impedance load just before the tracer pulse is applied, Q.

In a similar manner, the relative mass can be given by (Kachanoski et aI., 1992): (20)

where, as before all calibration constants are cancelled. As long as a linear relationship exists between the electrical conductivity of the soil solution,

O"w'

and

O"a'

then all calibration constants are not present in the final equation for the relative solute mass.

19

2.5 Calibration of TDR for Solute Transport 2.5.1 Direct Approach Direct calibration relates measured load resistance to the actual concentration of the electrolyte tracer at specified soil water contents and is usually obtained from an experiment different from the actual transport experiment (Kachanoski et al., 1994). This technique is best suited to vertically homogeneous soils (repacked or undisturbed). Briefly, samples of the soil of interest are equilibrated with different amounts of water of known concentration.

2.5.2 Indirect Calibration Approach 2.5.2.1 Vertically Installed Probes Indirect calibration relates R- 1 to the applied solute mass or concentration. Thus, calibration does not require a separate experiment. This approach is well suited to in situ measurements in heterogeneous systems and undisturbed soil columns (Kachanoski et al., 199). In these conditions, the calibration relationship has been shown to vary with lateral and vertical probe location. The relative solute mass is given by equation 21.

2.5.2.2 Horizontally Installed Probes In vertically heterogeneous soils and soil cores, it may be more useful to use horizontal probes to obtain information on the variation of transport parameters with depth. Under a constant-flux surface boundary, the movement of a surface-applied

20 electrolytic tracer to a depth,

Z,

may be tracked by measurement of load resistance,

R(z.,t), obtained from horizontal probes installed at depth, z. Estimates of resident concentration as a function of time at a fixed depth C(z,t) are then obtained from the following equation (Kachanoski et al., 199):

(21) where R j = the impedance before any tracer is added, Q R= impedance measured after a tracer has been added,

n

P(8)= slope calibration constant The calibration constant,

p, becomes a function of depth, z;

and must be calculated for

each probe. There are two methods of obtaining the calibration constant.

1. Step function input One approach is to use a step increase of solute at the surface under steady state conditions (dS/dt =0). At large time, t F, after the solute dispersion front has passed a horizontal probe located at depth, z, the resident concentration between the

'fDR rods will equal the input concentration, Co' Thus, rather than equating the maximum change in R-1 to the specific mass of the applied tracer, as with the vertical probes (Eq. 20), it is equated to the input concentration such that (Kachanoski et aI., 1994):

(22) Since Co is known, and ~(tF) and ~(t) are directly measured, the calibration constant

is easily obtained. This assumes that solute is uniformly distributed throughout the

21 soil solution.

2. Convolution of a Solute Pulse Response Although transport parameters could be obtained from experiments conducted with a step increase at the surface, in many instances a pulse input of tracer is more desirable. Certain features, such as bimodal transport, are more accurately measured with a pulse input than with a step input.

2.6 Definition of NAPLs In the past few years, as hazardous waste sites have been studied more often and in more detail, immiscible fluids have been encountered in the subsurface with greater frequency. These nonaqueous phase liquids (NAPLs) behave differently than dissolved solutes in the subsurface. This behavior depends on fluid properties such as interfacial tension, viscosity and density. Non-aqueous phase liquid (NAPL) is a term used to describe a class of organic liquid contaminants which are considered to be immiscible with water (Redman et aI., 1994). They incorporate both DNAPLs (dense NAPLs) that are more dense than water and LNAPLs (light NAPLs) that are less dense than water. Although NAPLs are considered immiscible with water, large plumes of dissolved-phase groundwater contamination are produced because NAPLs have solubilities that are well above drinking water standards. Typical examples of DNAPLs are the chlorinated solvents (eg., tetrachloroethylene), and typical LNAPLs are the hydrocarbon fuels (eg., gasoline, and kerosene).

22 2.7 Sources of NAPLs Nonaqueous phase liquids (NAPLs) have been discovered at numerous hazardous waste site. Moreover, NAPL often is identified with contamination problems associated with underground storage tanks. According to Villaume (1984), typical chemical and industrial processes that may involve NAPLs include (Mercer et

al., 1990): *transformer oil containing polychlorinated biphenyl *trichloroethane and related chlorinated hydrocarbon *coal tars from manufacturing gas plants *steel industry coking operations *wood treating operations, and *petroleum products.

2.8 Migration of NAPLs in Subsurface Although chemical properties and site conditions vary from site to site, the basic principles governing the fate and transport of NAPLsare the same. These principles may be used to understand the contamination problem and to evaluate remediation. NAPL migration in the subsurface is affected by (Mercer et al., 1990): (1) volume of NAPL released (2) area of infiltration (3) time duration of release (4) properties of the NAPL

23 (5) properties of the medium (6) subsurface flow conditions When LNAPL is introduced into the subsurface, gravity causes the LNAPL to migrate downward through the vadose zone as a distinct liquid. This vertical migration also is accompanied to some extent by lateral spreading due to the effect of capillary forces and due to the medium spatial variability (eg., layering).

As the NAPL progresses downward through the vadose zone, it leaves residual liquid (residual saturation) trapped in the pore spaces. This entrapment is due to surface tension effects. In addition to migration of NAPL, some of the immiscible fluid may volatilize and form a vapor extending beyond the NAPL.

Denser-than-water nonaqueous phase liquid (DNAPL) will displace water and continue its migration under pressure and gravity forces. Preferential spreading will occur where DNAPL encounters relatively permeable layers, fractures, or other pathways that present less capillary resistance to entry than underlying less permeable strata (Mercer et al., 1990). Given sufficient volume, DNAPL will continue its downward migration until it encounters barrier layers upon which it may continue to flow under pressure and gravity forces. As in the vadose zone, some of the DNAPL will be held in the pore space within the saturated zone. This residual DNAPL will serve as a chemical source to the flowing groundwater depending on the aqueous solubility of the organic sources compounds. In the vadose zone, infiltrating rainwater

24

may dissolve organic vapors or the residual DNAPL and transport these organic components to the saturated region (Mercer et aI., 1990).

Below the water table, ground water will flow through regions containing DNAPL residual and pools, giving rise to dissolved-phase contamination. Because many DNAPLs have a relatively low solubility and because most natural ground-water velocities are low, it may require up to several decades and possibly centuries before residual and pool zones are depleted by natural dissolution alone. Since the drinking water guidelines for most DNAPL compounds are orders of magnitude below their aqueous solubility, it follows that the concentration of the associated contaminant in the plumes will be large, and that the majority of water in these plumes will be out of regulatory compliance (Kueper et aI., 1990).

2.9 Properties of NAPLs The subsurface transport of NAPLs is governed by various factors, some of which are different from those for dissolved (or miscible) contaminants. In this section, properties associated with NAPL flow are discussed.

2.9.1 Saturation The saturation, s, of a fluid is the volume fraction of the total volume occupied by that fluid (Mercer et al., 1990). Saturation varies from zero to one, and the saturations of all fluids sum to one. Saturation is important because other properties,

25

such as capillary pressure and relative permeability, are represented as a function of saturation.

2.9.2 Interfacial Tension Liquid interfacial tension is equal to the free surface energy at the interface formed between two immiscible or nearly immiscible liquids (Mercer et aI., 1990). It is caused from the difference between the mutual attraction of like molecules within each fluid and the attraction of dissimilar molecules across the fluid interface. Liquid interfacial tension is directly related to the capillary pressure across a NAPL-water interface and is a factor controlling wettability.

2.9.3 Wettability Wettability describes the preferential spreading of one fluid over solid surfaces in a two-fluid system; it depends on interfacial tension (Mercer et aI., 1990). Whereas the wetting fluid will tend to coat the surface of grains and occupy smaller spaces in porous media, the nonwetting fluid will tend to be constricted to the largest openings.

2.9.4 Capillary Pressure Capillary pressure is a property that cause porous media to draw in the wetting fluid and repel the nonwetting fluid (Mercer et aI., 1990). If capillary pressure is assumed positive, it is defined as the difference between the nonwetting fluid pressure and the wetting fluid pressure. For a water-NAPL system with water being the

26 wetting phase, capillary pressure, Pc, is defined as (Mercer et al., 1990): P, == PN

-

Pw

(23)

where PN is the NAPL pressure; and P w is the water pressure.

2.9.5 Residual Saturation Residual saturation (s.) of NAPL is the saturation (V NAPL/V voids) at which NAPL becomes discontinuous and is immobilized by capillary forces under ambient groundwater flow conditions (Mercer et aI., 1990). Residual saturation of NAPL has important consequences in the remediation of subsurface contamination because drinking water standards for many NAPL's are orders of magnitude less than their solubility limits. In addition, residual saturation results from capillary forces and depends on several factors, including (Mercer et aI., 1990): (1) the medium pore size distribution (2) wettability (3) fluid viscosity ratio and density ratio (4) interfacial surface tension (5) gravitylbuoyancy forces, and (6) hydraulic gradients.

2.9.6 Solubility The aqueous solubility of a chemical is the maximum concentration of the chemical that will dissolve in pure water at a particular temperature (Mercer et aI.,

27

1990).

NAPL 's vary widely in their aqueo us solubility. Nonpo lar hydrop hobic compo unds are less soluble than polar hydrophilic compounds. Solubi lities may be measured experimentally or estimated based on empirical relationships developed betwe en solubility and other chemical proper ties such as partition coefficient and molec ular structure. Several factors influen ce solubility, including temperature, cosolvents, salinity and dissolv ed organic matter.

2.9.7 Volati lizatio n Volatilization refers to the transfer of matter from liquid and soil to the gaseous phase (Mercer et aI., 1990). Therefore, chemi cals in the soil gas may indicate the presence of NAPL or dissolved chemicals. Chemical properties affecti ng volatilization are vapor pressure and solubility in water. Moreover, other factors affecting volatilization rate include: concentration in the soil, soil moisture, soil-air movement, sorptive and diffusive characteristics of the soil, temperature, and bulk properties of the soil such as organic carbon content, porosity, density and clay conten t (Mercer et al., 1990). Volatilization increases with soil air movement. Also, volatilization losses in the subsurface from NAPL s are expect ed where NAPL s exist close to the ground surface or in the dry pervious sandy soil, or where a NAPL has a very high vapor pressure.

28 2.9.8 Viscosity Viscosity is the internal friction within a fluid that causes it to resist flow (Mercer et aI., 1990). After a spill, a product of low viscosity will penetrate faster into the soil than a product with higher viscosity. For example, fresh crude oils with volatile components become increasingly viscous as they evaporate.

Also significant is the NAPL-water viscosity ratio, which is part of a term used in the petroleum industry known as the mobility ratio. In a water flood, the mobility ratio is defined as the mobility of the displacing fluid (relative permeability/viscosity for water) divided by the mobility of the displaced fluid (relative permeability/viscosity for NAPL) (Mercer et al., 1990). Mobility ratio of> 1 favors the flow of water whereas the ratio of < 1 favors the flow of NAPL.

2.10 Methods of Measuring NAPLs in Subsurface

NAPL source locations are often poorly defined. Characterizing the presence, composition and properties of mobile and residual NAPLs is fundamental to determining the nature, extent and rate of chemical migration during a site assessment (Mercer et aI., 1990). Data collection typically occurs in phases using many of the same techniques employed at contaminated sites where NAPL is not present. However, special precautions and considerations must be considered where a NAPL is present. Here are some of the field techniques which are used to characterize NAPL transport.

29

2.10.1 Soil Gas Analysis For volat.ile NAPLs, soil gas analysis may be a screening tool for the location of potential sources of contamination and for siting monitoring wells.

Grab and passive sampling techniques are used to collect soil gas. During static grab sampling, samples are collected from a quiescent soil gas sample (Mercer et aI., 1990). In dynamic grab sampling, samples are collected from moving soil gas as it is pumped through a hollow probe. Grab samples can be analyzed on site during the sampling episode or shipped to a laboratory. Passive sampling provides an integrated measure of VOC concentrations over time. Charcoal sorbents to trap solutes that diffuse through the soil gases are buried up for one month and then retrieved for analysis (Mercer et aI., 1990).

Although soil gas analysis has been used successfully at many contamination sites, it can provide misleading results if subsurface conditions are not understood adequately.

2.10.2 Drilling Investigation Borings and monitoring wells are the primary means for evaluating the distribution of subsurface NAPLs (Mercer et aI., 1990). Drilling at NAPL contaminated sites occurs most frequently using hollow-stem augers with split-spoon samplers in soils and wireling-coring or rotary drill rigs in bedrock. The presence and

30 density of NAPL relative to water are easily identified in the field by shaking free liquid or NAPL contaminated soil in a jar with water and watching for phase separation.

Free liquids or NAPL extracted using centrifugation or other methods can be tested for fluid properties such as density, interfacial tension and viscosity. Also, monitoring wells are used for the characterization of NAPL. Well screens in LNAPL monitoring wells should be long enough to ensure that the entire free LNAPL layer will be within the screened interval during seasonal water-table fluctuations (Mercer ey aI., 1990). Misleading LNAPL measurements will result from a well that is screened through perched LNAPL in the vadose zone to a deeper water table. At DNAPL sites, a well that is fully screened from above the DNAPL-water interface to the top of the underlying stratigraphic barrier is usually appropriate. Sandpacks should be coarser than surrounding media to ensure the movement of NAPL to the well.

2.10 Application of TDR to monitor NAPLs in the subsurface There have been few attempts to use TDR for the purpose of monitoring NAPLs migration in the subsurface. However, Redman and DeRyck (1994) have proposed multilevel time domain reflectometry probes for monitoring the subsurface distribution of non-aqueous phase liquids (NAPLs). The idea behind this method is that the volumetric fraction of NAPLs can be estimated with TDR probes because, in general, they have a much lower electric permittivity than the water they displace.

31

Two separate NAPL injection experiments were conducted to study the use of TDR to monitor NAPLs migration. In the first experiment, a multilevel TDR probe was designed in an LNAPL (Kerosene) injection within a large cylindrical polyethylene test cell packed with a local sandy soil. The probe consisted of 65 brass rods spaced every 2.5 em from the surface of the cell to a depth of 1.6 m. The multilevel probe was installed in the test cell before packing. Sand was packed in the cell in layers of 10 to 15 em at a residual volumetric water content of about 6%. Each layer was raked and tamped, and particular care was taken to ensure uniform packing around the brass rods. Then, kerosene was injected in a number of separate injections to give a total injected volume of 343 liters. The transport of kerosene was monitored with the multilevel TDR probes over a 4-month period before, during" and following all injections. The results are presented as vertical profiles of TDR estimated water content and change in water content (Figure 2.10.1). The profiles show a reduction in water content, principally within the capillary zone. This reduction in water content is roughly equivalent to the kerosene content.

The second experiment was to test DNAPL (tetrachloroethylene) injection in a natural, saturated sandy aquifer isolated within sheet pile walls. The multilevel probes, as in the LNAPL experiment, were designed to have minimal influence on the geophysical techniques, and therefore could not incorporate permanent wiring. The probe consisted of 21 twin-wire transmission lines mounted on the outside of a PVC access tube. The transmission lines consisted of IS-em long brass bars ( 3 mm by 3

32

WATER CONTENT 0

30

20

10

(~~)

VOLUMETRIC FLUID CONTENT (%)

40

50

-2

0

}

t

0.2

..

I'

~..

0.4

0.2

Saturated I!, I

2

4

6

8

10

,2

14

16

•. • *

\i"b

/;t

0

I

J ~

0.4

.-i ~

Kerosene content from core

Background '~

II f~

II I( .!

~

I.

0.6

~

(I

0.6

I,

~o~

\J ? L ~ 100l~:'/' 0

m v

0.8

/ ../

200

-1

J:

)~ L'~ \ ~ .......

~

343

.-j

".":''1:

0.8

I: .~ '.~ ~

,))

(

t.

.i->:".. ~~ .... \ l , :} \ .

'1i

Water table

Water table

\1

l:

1.2

I 1.4

~l

1.

Water content loss (TOR)

1.6

1 .8

1.8

(a)

(b)

Figure 2.10.1 - (a) TDR derived water content before kerosene injections and following each injection. Total volume of injected kerosene is indicated. (b) Water content loss after final spill compared with kerosene content measured in core samples (Redman et aI., 1994).

33 mm) with a I-cm space between the ends of each twin-wire line. A brass screw was attached at the top of each brass bar and passed through the wall of the PVC pipe. The multilevel probes were installed using the following procedure. A steel casing, with a disposable drive point on its lower end, was driven into the aquifer to a depth of 3.7 m. The drive point was knocked off, the casing was filled with water, and then the TDR probe was placed inside the casing. The casing was pulled out while holding the TDR probe in place, allowing the soil to collapse around the outside of the probe. This resulted in an annular zone of disturbed soil, approximately 0.8 ern thick.

In this experiment, 770 liters of tetrachloroethylene were injected into the fully water saturated cell over a period of 70 h. Profiles of DNAPL saturation, estimated from the observed change in dielectric permittivity using the Bruggeman-Hanai-Sen (BHS) model, show the development of DNAPL pools at depths of 1.25 m and 3.0 m (Figure 2.10.2).

In short, the changes observed in dielectric permittivity profiles, measured on multilevel TDR probes, have proven to be very useful in monitoring NAPL migration and distribution in large-scale injection experiments.

34

O. a j--------,.-----...-----..,..---------.--....--.-----------

0.5

r-

69

0.2

0.3

0.4

0.5

0.6

0.7

G.?

DNAPL Saturation Figure 2.10.1 Estimated DNAPL saturation from multilevel TDR probe data, before and during the 70 h injection (Redman et al., 1994).

35 CHAPTER 3

EXPERIMENTAL SETUP AND PROCEDURES

3.1 Introduction In this research there were two phases. In the first one, a static study on the effects of the addition of NAPL to soils was conducted using TDR. Also, the effects of types of soil, contaminant and water content were studied. In the second phase, the potential use of TDR to study the migration of NAPLs in the subsurface was conducted. Five different soil types and three NAPLs were used to achieve this goal.

3.2 Data Acquisition The Tektronix 1502B cable tester was used in this research to obtain the TDR traces for all the tests performed. The Tektronix 1502B Time Domain Reflectometer is a short range metallic cable tester capable of finding the faults in metal cable. It sends an electrical pulse down the cable, and detects any reflections. Also, it is sensitive to impedance changes. The 1502B is a highly accurate tester, easy to use, and provides fast, accurate measurements.

The 1502B generates a rapidly rising step signal, applies it to the cable under test, and detects and processes the reflected voltage waveform from the cable. These reflections are displayed on the 1502B's Liquid Crystal Display (LCD) where distance

36 measurements may be made using the cursor. Impedance information is obtained by interpreting waveform amplitude. The waveform may be temporarily stored within the 1502B and recalled. In this research, the waveforms are printed out using the dot matrix strip chart recorder, which is installed into the front panel option port.

There are a lot of different types of TDR probes. In this research, the three rod probe was used (20 em in length, 2.5 ern spacing between the rods), becuase it was less destructive and allowed water movement to occur perpendicular to the probe rods. Also, this probe design is cost efficient due to the absence of a balun transformer.

3.3 Calibration of TDR Probe The TDR probe was calibrated for the fine and coarse sand used in this research for water content. In the range of water content of interest, from 3% to 30%, Topp's equation works well. In theory, the TDR requires no calibration and the soilmoisture determination is independent of soil texture, structure, salinity, density, or temperature. In reality, different calibrations may be necessary for compensation of the TDR system configuration and the type of soil being sampled. In calibrating for a particular soil type, a series of samples were analyzed, typically varying in moisture content from 5% to 35% moisture by volume, both by oven-drying and with TDR. Paired values of oven-dried moisture determinations (assumed as actual) versus the ~rDR

calculated moisture were used to develop a relationship, through regression

37 techniques, between these values such that the TDR response is calibrated for the soil of interest. The regression statistics yielded

r

values around 0.948 for fine sand, and

0.886 for coarse sand. See Figures 3.3.1 and 3.3.2 for the calibration values of TDR for fine and coarse sand. The water content in the silty sand was determined by TDR.

3.4 Experimental Procedure of Phase I In this part of the study, there were two types of sand used, fine and coarse. Silty sand also was used in the following percentages of 10, 30, 500/0 by weight mixed with coarse sand. The NAPLs that were used in this experiment were: gasoline and diesel for LNAPL and tetrachloroethylene (Pf'E) for DNAPL (see Figure 3.4.1 for the properties of these NAPLs). Table 3.4.2 summaries each soil type, contaminant and water contents that were tested.

In this phase, plastic containers ( 12" (30 em) long, 6" ( 15 em) wide and 6" \

(15 em) deep) were used in the testing. Each test started with cleaning and drying the soil. Then, the soil was weighed and wetted to the desired water content. The soil was then left overnight to reach equilibrium. The water content was measured by TDR and calculated using the calibration curves mentioned previously. Then, the measured volume of NAPL was added and mixed thoroughly. The soil was then covered and left for 15 to 20 minutes to reach equilibrium. Then, the TDR probe was placed into the soil with special eare taken to cover the whole probe with the soil. A TDR reading was then performed. Additional measured volume of the NAPL was

38

.. c

0.35 0.3

8

0.25 0.2

Sc

.s ;

0::

~

0.15-+0.1 0.05

0

o

0.2

0.1

0.4

0.3

Gravim etric w ate r conte nt

Figure 3.3.1 Calibration curve of water content for fine sand.

0.35 .. e 13 c

0.3 0.25

S

0.15

0:: ~

0.05

...8

;

0.2

0.1

o

+ ...-----+-------t-----r----+----j---r------1

o

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Gravim etric w ater Content

Figure 3.3.2 Calibration curve of water content for coarse sand.

39

Table 3.4.1 Properties of the NAPLs used in this research. Liquid

Specific gravity

Absolute viscosity (m.Pa's)

Surface tension (mN/m)

Interfacial tension

Dielectric constant*

rce

1.63

1.932

32.86

44.4

3.68

Gasoline

0.732

0.45

21

50

4.08

Diesel

0.8

1.1

25

50

4.08

Water

1.00

1.00

71.99

*Measured by TDR.

---

70

40

Table 3.4.2 Media, contaminants and water content used in phase I Media

Contaminant

Water content (%)

Fine sand

gasoline

0,5.6,11,20,35

Fine Sand

diesel

0,5.6,11,20,29

Fine Sand

peE

0,5.6,8.3,17,32

Coarse Sand

gasoline

0,5.6,8.3,20,34

Coarse Sand

diesel

0,8.3,11,20,32

Coarse Sand

peE

0,5.6,11,17,29

Silty Sand (10%)

gasoline

0,14,20,29

Silty Sand (30%)

gasoline

0,14,26,29

Silty Sand (50%)

gasoline

0,14,31,35

41

then added in the same manner as explained previously until the soil was completely saturated with the NAPL. TDR readings were taken after each addition. All tests in this phase were performed following the same procedure explained in this section.

3.5 Experimental Procedure of Phase II The experimental setup for the NAPL transport test consisted of the following: 1) a cylindrical polyethylene test cell (12" (30 ern) in diameter by 16" (40.5 em) deep) 2) a plastic cover with uniform openings to simulate an even and uniform spill 3) three TDR probes inserted into the cell with 5" (12.7 em) spacing between them 4) 1502B TDR cable tester for data acquisition A system schematic is presented in Figure 3.5.1 and Figure 3.5.2. The system was setup in Room 203 of Stocker Center. It was placed under a ventilation system to remove any possible vapor generated by the evaporation of NAPLs.

Two types of sand were used in this phase, fine and coarse. The NAPLS that were used were diesel for LNAPL and tetrachloroethylene for DNAPL. Table 3.5.1 summaries the soil types and contaminants that were used in phase II.

Three TDR probes were used in the NAPL spill experiments within the cylindrical test cell which would provide three measurements points. The first probe was 5" from the surface, and each probe was spaced 5" (12.7 em) apart. The third probe was placed 1" (2.5 ern) from the bottom of the test cell.

42

I

I

8

~ri ! I

15028 TOR I I

\

... ... ... .... ~.-... II

D

II

D

A

iI C

A

8

c

o

12"

1"

5"

Figure 3.5.1 System setup for phase II.

43

Figure 3.5.2 A photograph of the system set-up for phase II.

44

Table 3.5.1 Media and contaminants used in phase II Media

Contaminant

Coarse Sand

Diesel

Coarse Sand

PCE

Fine Sand

Diesel

Fine Sand

PCE

I

45 In this phase, coarse and fines sand were used. The sand was cleaned and dried. Then, it was weighed and wetted with water to saturation. The sand "vas packed into the test cell, and the TDR probes were then inserted during the packing. Sand was packed in the cell in layers of 10-15 em. Each layer was raked and tamped, and particular care was taken to ensure uniform packing around the TDR probes. After steady-state water content conditions had been reached, as determined by the constancy of the apparent length of the TDR trace, a fi.rst spill of the NAPL was performed by pouring 600 mL of the NAPL into the openings of the plastic cover. Then, the dielectric constant measurements were taken frequently until there was no change in the profile. Readings were taken more frequently in the first 5 to 6 hours and less after that. Then, a second spill was performed by pouring an additional 400 mL of the NAPL into the test cell (the 1000 mL volume was determined in comparison to the ratio of NAPLs to the total volume used in Redman's experiment, 1994). More TDR readings were then taken in the same manner as in the first spill. All the tests in this phase were performed in the same way described in this section.

3.6 Clean-up and Disposal of Waste When each experimental run was completed, the container was emptied and the soil waste contaminated with NAPL "vas disposed into buckets and sealed. The Environmental, Health and Safety (EHS) group at Ohio University was then contacted for the removal of the hazardous waste.

46

CHAPTER 4

DISCUSSION OF RESULTS

4.1 Dielectric constant results in phase I After each addition (50 - 100 mL) of the NAPL to the soil, a measurement was taken of the dielectric constant using TDR. On the average, there were about 4 to 5 data points. Figures 4.1.1 - 4.1.19 show the results obtained for the soils and NAPLs that were used in phase I. In these graphs, Vg is the volume of gasoline, Vp is the volume of PCE, V d is the volume of diesel and VnN s represents the ratio of the volume of the NAPL added to the volume of the soil.

Figure 4.1.1 shows the results obtained for fine sand with gasoline at different water contents. There was an increase in the dielectric constant from 0% to 20% range of the water content as more of the gasoline was added to the sand. However, there was a drop in the dielectric constant as more gasoline was added when the sand was saturated at 35% water content. Figures 4.1.2 and 4.1.3 show the results obtained for fine sand with diesel and peE, respectively. The results observed were the same as in the fine sand with gasoline, where there was an increase in the dielectric constant when the sand was not saturated and a drop in the dielectric constant when the sand was saturated as more of the NAPL was added. Figures 4.1.4 to 4.1.6 show the

47

25

...., t: J9 tn

20

0

15

e

1--=+=6~~O%;~t~~~l i

content

i

I

(.)

I

i

--.- 5.6% water content

(.)

....,

"L:

e

10

---.- 11 % water content

(1)

Gi

C

5

--*- 20% water 0

content 0

0.14792

0.29584

0.44375

VgNs

~

35% water content

Figure 4.1.1 The results of fine sand with gasoline at different water content.

.....c .....eneu c

0 CJ CJ

r-=-.-=O~O~/;water

20

content 15 10

_5.6% water content

5

--.- 11% water content

".: ..... CJ

(1)

G)

s

~ 20%

0 0

0.1055

0.2461

0.457

water content

~29%water

VdNs

content

Figure 4.1.2 The results of fine sand with diesel at different water content.

48

...., c: ....,C'G

1---+---- -- ---- -1

20

rn e 15

____ 5.6% water

0

e

o

10

eCI)

5

is

0

0.0% water content '

i

content

....,

"i:

---.- 8.3% water content

a;

0

~

C\.I

M 0'> 0 M

C\.I

v

0

0

0

0

LO

f'...

M C\.I

f'...

~

~

f'...

M

170/0 water content

--*- 32% water

VpNs

content

Figure 4.1.3 The results of fine sand with PCE at different water content.

...., 30 c J9 25

--.- O.O°t'o water

e 0 e e

_

tn

"i: ....,

u

eu Gi

is

content

20 15

5.6°t'o water content

10 5 0

--.- 8.3% water content 0

0.136

0.272

0.408

--*- 20°t'o water content

---*- 34°t'o water VgNs

content

Figure 4.1.4 The results of coarse sand with gasoline at different water content.

49

r-=.-

O.OOk water

I

content

' _ 8.3% water

content --+-- 11okwater content ~ 200k water content

--T- 32°k water content

VdNs

Figure 4.1.5 The results of coarse sand with diesel at different water content.

..... c ca ..... e

15~~

CJ CJ

10

0

0

--+- O.eJ>JOwci.er

2)

cxrtErt

_

-t:

..... CJ CD

5.ffVowci.er cxrtErt

--A- 11 010wci.er cxrte1

5

Gi

C

0

---1

~-+----+-

0

O.Cm3

Vpvs

~

1?JOwci.ercxrte1

0.3131 ~ 2fJ>JOwci.er cxrte1

Figure 4.1.6 The results of coarse sand with PCE at different water content.

50 to 4.1.6 show the results obtained for coarse sand with gasoline, diesel and peE at different water contents. The outcome of the results for the coarse sand was the same as in the fine sand. Figures 4.1.7 to 4.1.11 show the results obtained for fine sand after the additions of gasoline, diesel and PCE at different water contents.

In general, there

was an increase in the dielectric constant as more NAPL was added; except when the fine sand was saturated (Figure 4.1.11). The saturation of the sand seemed to have an effect to reduce the dielectric constant as more of the i'TAPL was added. Also, it is evident from this graph that it took more volume of the DNAPL, PCE, to change the dielectric constant than the LNAPLs. The values of V n/Vs, when the dielectric constant changed, of peE were larger than of those of .gasoline and diesel at the different water contents. In addition, gasoline had a larger Vn/Vs ratio than diesel, especially at lower water content.

Figures 4.1.12 to 4.1.16 show the results obtained for coarse sand after the additions of gasoline, diesel and peE at different water contents. There was still an increase in the dielectric constant as more NAPL was added to the soil; except when the coarse sand was saturated, there was a reduction in the dielectric constant. Again, PCE still had a higher value of V n/V S' when the dielectric constant changed, than gasoline and diesel at the different water contents. It means that more peE was needed to change the dielectric constant than gasoline and diesel. Also, gasoline had a higher Vn/V s than diesel.

51

.....

c 5 J! ~ 4 8 3 ~~~~a.------.~ 2 U 1 +-----------------(1) (i) 0 is 0 0.2 0.4 0.6

-+-FCE

_[lesel

--l-----------------

--.-

Suggest Documents