oEPA
United States
Environmental Protection
Agency
Geoelectrical Response
of Surfactant Solutions
in a Quartzitic Sand
Analog Aquifer
RESEARCH AND DEVELOPMENT
-------
-------
EPA/600/R-10/041
April 2010
www.epa.gov
Geoelectrical Response
of Surfactant Solutions
in a Quartzitic Sand
Analog Aquifer
Contract No. EP08D000104
Prepared for
Dale Werkema, Ph.D.
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Environmental Sciences Division
Characterization and Monitoring Branch
944 E. Harmon Ave.
Las Vegas, NV 89119
Prepared by
Meghan Therese Magill, M.S.
Department of Geoscience, College of Science
University of Nevada, Las Vegas
December 2009
Although this work was reviewed by EPA and approved for publication, it may not necessarily reflect official
Agency policy. Mention of trade names and commercial products does not constitute endorsement or
recommendation for use.
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
22138cmb10
-------
ABSTRACT
Geoelectrical Response of Surfactant
Solutions in a Quartzitic
Sand Analog
Aquifer
by
Meghan Therese Magill
In this project, the resistivity and phase shift of ten surfactant aqueous solutions in a
sand matrix were measured using spectral induced polarization (SIP). In addition,
specific conductivity, pH, dissolved oxygen, and dielectric constant measurements of the
solutions were also evaluated. The frequency range assessed was 0.091-12000Hz. The
surfactants, which are typically used in the remediation of tetrachloroethylene, were
Aerosol MA 80-1, Dowfax 8390, and Steol CS-330. The surfactants were mixed into
solutions of both deionized and tap water at varying concentrations and injected into a
closed system of silica sand. The surfactant treatments altered resistivity, specific
conductivity, and pH to varying degrees. Increased real and specific conductivities
associated with surfactant presence support the work of Werkema (2008), and the
correlation between real and specific conductivities indicates that the primary electrical
conduction mechanism in quartz sand-water environment. A decrease in the pH response
associated with high concentration surfactant solutions could impact subsurface
organisms, potentially affecting bioremediation. Phase, dissolved oxygen, and dielectric
constant response to surfactant showed little change from the control. The positive
results suggest that geoelectrical changes may be an applicable property to map and
monitor surfactant floods in the subsurface. In order to better understand how the
ii
-------
geoelectrical response of surfactant solutions would respond in a field situation, it will be
necessary to increase the complexity of the experimental set-up. Increasing the
heterogeneity of both the solid materials and pore fluid through the addition of clays and
chlorinated solvents are potential avenues to follow.
111
-------
ABSTRACT
TABLE OF CONTENTS
111
ACKNOWLEDGEMENTS vii
CHAPTER 1 INTRODUCTION 1
CHAPTER 2 BACKGROUND INFORMATION 7
Dense Non-aqueous Phase Liquids 7
Surfactants and SEAR 9
Geophysical Methods 13
Water Quality Measurements 22
Statistics 24
Experimental Design 28
CHAPTER 3 MATERIALS AND METHODOLOGY 32
Experimental Column 32
Spectral Induced Polarization 34
pH, Dissolved Oxygen, Specific Conductivity 36
Time Domain Reflectometry 38
Statistical Analysis 39
CHAPTER 4 RESULTS 42
Real Conductivity 43
Imaginary Conductivity 46
pH 49
Specific Conductivity 51
Dissolved Oxygen 54
Dielectric Constant 56
Results Summary 58
CHAPTER 5 DISCUSSION 59
Geoelectrical Measurements 59
Water Quality Measurements 65
Dielectric Constant 67
CHAPTER 6 CONCLUDING REMARKS 69
EXHIBITS FIGURES AND TABLES 72
APPENDIX A SYSTEMATIC ERROR RESULTS 131
APPENDIX B RAW DATA 144
REFERENCES CITED 169
VITA 177
v
-------
vi
-------
ACKNOWLEDGEMENTS
This work was completed under the US EPA Student Services Contract No.
EP08D000104 and was include as part of the requirements for a MS degree at the
University of Nevada - Las Vegas. Many people should be acknowledged for the
completion of this work. These people included; Dr. Dave Kreamer (UNLV), Dr. Dale
Werkema, and John Zimmerman at EPA. Additional thanks to Lisa Hancock, Ryan
Joyce, Alan Williams, and Danney Glaser for logistical help in the laboratory. I would
like to give particular acknowledgement to Maria, Liz, Kathryn, Rainee, and Joy in the
UNLV Geoscience Department office and Marion Edison in the EPA QAL office.
This work was partially funded through the U.S. Environmental Protection Agency
through the student services contracts, EP08D000104. Although this work was reviewed
by EPA and approved for presentation, it may not necessarily reflect official Agency
policy. Mention of trade names or commercial products does not constitute endorsement
or recommendation by EPA for use.
I would also like to extend my thanks to Anthony Endres and Alle van Calker for
their insight into time domain reflectometry and dielectric constants, as well as to Tino
Radic for developing the SIPLabll instrument and assisting in its proper operation.
vii
-------
CHAPTER 1
INTRODUCTION
Groundwater contamination is a growing concern both domestically and
internationally. Dense non-aqueous phase liquid (DNAPLs) comprise one category of
contaminants. DNAPLs are immiscible, denser-than-water fluids that include creosote,
common metal degreasers, and solvents used as dry cleaning fluid. Once released to the
environment, DNAPLs are difficult to remediate, but some methods have been used to
varying degrees of success, including the use of surfactants floods. While seemingly
effective in field studies, there are questions and concerns regarding surfactant-enhanced
aquifer remediation that suggest the need for further research. These include potential
problematic behavior in the subsurface and possible impacts on future remediation.
There is a potential to use geophysical methods, particularly geoelectrical, to monitor
surfactants in the subsurface to better understand their behavior. Previous work has
indicated that measurements of pH and specific conductivity that were taken in surfactant
solutions of deionized water had a different response than control samples (Werkema,
2008). This chapter will introduce the concepts fundamental to this thesis, with further
details to follow in subsequent chapters.
Surfactants are amphiphilic monomers composed of a head and tail, generally a
functional group and carbon chain, respectively. They are potential groundwater
remediators of non-aqueous phase liquid (NAPL) through multiple mechanisms (e.g.,
U.S. EPA, 1996; Dwarakanath etal., 1999; Londergan et al., 2001). Their amphiphilic
nature can both increase a contaminant's solubility through micellar solubilization as well
as decrease the interfacial tension between the non-aqueous and aqueous phases.
1
-------
Surfactant-enhanced aquifer remediation (SEAR) is a promising technology that utilizes
water and surfactant solution floods to remove residual DNAPL from an aquifer (Brown
et al., 1999; Dwarakanath et al., 1999; Londergan et al., 2001; U.S.EPA, 1996). This
technique appears to be a viable method but has not experienced widespread acceptance.
There are several criticisms of SEAR which require further research to resolve.
Understanding of surfactant behavior in the subsurface is limited, and many surfactants
are known to produce uneven wetting surfaces. In addition, the use of surfactants to
decrease the interfacial tension between a contaminant and pore fluid may result in
unwanted downward migration of the contaminant (Longino and Kueper, 1995). Finally,
the current inability to monitor surfactants' behavior in the subsurface makes it difficult
to determine whether surfactants are reaching the DNAPL-contaminated areas (Conrad et
al., 2002), potentially resulting in less efficient use of surfactants.
Geoelectrical methods including direct current (DC) resistivity, induced polarization
(IP), spectral induced polarization (SIP), and ground penetrating radar (GPR) have been
used to successfully map DNAPL in the subsurface (Adepelumi et al., 2006; Brewster et
al., 1995; Brewster and Annan, 1994; Grimm et al., 2005; Sogade et al., 2006). To
address the possibility of monitoring subsurface surfactant floods with non-invasive
geophysical techniques, Werkema (2008) tested several physicochemical parameters of
various surfactant aqueous solutions, without considering the contributions of solid
materials to these responses. This work found that solutions with surfactants showed an
increase in specific conductivity over solutions containing no surfactant. Dissolved
oxygen (DO), pH, temperature, and density were also tested, with dissolved oxygen, pH,
and specific conductivity showing the most predictable response. Because of the positive
2
-------
response of conductivity to surfactant presence, this research focuses on geoelectrical
methods and the parameters that affect them.
In an effort to further previous research, the inclusion of solid materials in the
experimental design added more complexity to the conditions, as well as introducing a
more realistic situation, closer to what would be encountered in a field application of
SEAR. The working hypothesis of the research is that the addition of surfactants will
result in a measurable geoelectrical response in analog aquifer materials that can be
directly or indirectly detected with SIP and time domain reflectometry (TDR). This
anticipated response may enable the use of non-invasive geoelectrical methods to map the
subsurface distribution of a surfactant flood. The ability to detect the surfactants or the
impact of those surfactants used in SEAR could reduce monitoring uncertainty and
increase the technique's use, resulting in more effective clean-up of groundwater.
The hypothesis was tested through a series of 30 experiments. Resistance, phase, pH,
DO, specific conductivity, and dielectric constant measurements were made and
analyzed. The measured parameters and reasoning for including them are found in
Chapter 2, Background Information. A simplistic analog aquifer was created using
quartz sand saturated with ten testable surfactant solutions packed into an 18cm long
column, 3.5cm in diameter. In this research, the term "analog aquifer" refers to a
simulated aquifer environment made with clean quartz sand acting as aquifer solid
material and the experimental solutions acting as pore fluid. The construction and further
details of the analog aquifer are described in Chapter 3.
The research presented in this paper investigates the SIP response of surfactants in a
quartz sand-water matrix, in addition to select water quality measurements and dielectric
3
-------
constant, measured through time domain reflectometry (TDR). The overall objective of
the research was to gain additional insight into the geophysical and physico-chemical
responses that could occur during the use of surfactants in groundwater remediation in
order to determine the feasibility of using geophysical methods, particularly geoelectrical,
to monitor those surfactants in the subsurface. In addition, water quality parameters can
indicate changes in physicochemical conditions that may affect geophysical responses of
the subsurface or impact remediation efforts.
The ultimate goal of efforts in this field of research is to non-invasively monitor and
map surfactants in the subsurface that have been introduced as part of a field application
of SEAR. Eventually, it may also be possible to determine the effectiveness of SEAR
remediation at a particular site by monitoring where the surfactants are located in the
subsurface and whether the DNAPL contaminants are being effectively remediated.
In order to realize these goals, the geoelectrical response to surfactants must be
characterized in a laboratory setting to isolate the response of the surfactant from the
responses to conditions that will be encountered in the field. Performing small-scale
experiments in the lab allows conditions to be adjusted and monitored in order to isolate
and scale the experimental response. To enable proper scaling of response, complexity
must be incrementally added to the system until it is well understood.
The research described in this thesis represents an early stage of characterizing the
geoelectrical and water quality responses of surfactants used in SEAR. Previous work
has not included a solid matrix in experiments to represent the solid materials in the
subsurface. This research has implemented a simple matrix of 20-30 sieve-size silica
sand. The absence of clay in the matrix material was purposeful as clays introduce an
4
-------
additional level of complexity to the system beyond simple quartzitic sand. Clays have a
complex geoelectrical signature, in addition to reacting with surfactants in the subsurface.
The near-uniform grain size and mineral composition limits the heterogeneity and any
anomalous response associated with changing environmental conditions. In short,
changes in geoelectrical and water quality response should indicate changes in the pore
fluid, as opposed to changes in packing method and mineral composition.
The experiment plan was designed with the help of Design Expert 7.0 (Stat-Ease,
2007), an experimental design statistical software used mainly in manufacturing and
industrial engineering to optimize performance through combinations of factors
(Anderson and Whitcomb, 2000). The design utilized in this project is a General
Factorial, more specifically a two factor interaction (2FI). As the name implies, this
project utilizes two categorical factors, surfactant and water type, with five measured
responses: resistivity, phase, specific conductivity, dissolved oxygen, and dielectric
constant. Three repetitions of each surfactant and water treatment were performed,
resulting in 30 experiments. Further description of experimental design is found in
Chapters 2 and 3. One of the goals of this project is to develop a statistical model for
each of the tested responses (i.e., the dependent variables) of real and imaginary
conductivity, pH, dissolved oxygen, specific conductivity, and dielectric constant due to
the experimental factors. The independent variables in this research are the surfactant
treatment and water type.
Investigating the anticipated geophysical response using SIP may enable the use of
non-invasive or partially invasive geoelectrical methods to map the subsurface
distribution of a surfactant flood. The ability to detect the surfactants or the impact of
5
-------
those surfactants used in SEAR could reduce monitoring uncertainty and increase its use,
resulting in more effective clean-up of groundwater. This research addresses this issue
by investigating the geophysical response to select surfactants in a saturated quartzitic
sand matrix.
6
-------
CHAPTER 2
BACKGROUND INFORMATION
The goal of this chapter is to provide the reader with background information relevant
to this thesis. Topics covered include dense non-aqueous phase liquids, surfactants and
groundwater remediation techniques utilizing them, and geophysical methods.
Information on relevant water quality measurements, as well as the experimental design
and statistics used in analysis, are also located in this chapter.
Dense Non-Aqueous Phase Liquids
Dense non-aqueous phase liquids are chemical compounds that are generally
immiscible, only slightly soluble in water, and have specific gravities greater than one
g/cm3. Common DNAPLs include chlorinated solvents like tetrachloroethylene (PCE),
as well as polychlorinated biphenyls (PCBs), coal tar, and creosote (e.g., Brewster et al.,
1995; Reynolds and Kueper, 2000). Many DNAPLs are carcinogens and possible
teratogens. As such, they are a threat to human health when released to the environment
(U.S.EPA, 1991).
Transport of DNAPL in the subsurface is complex and primarily driven by gravity
and capillary forces (Figure 1). A typical DNAPL contaminant plume will flow through
the vadose zone to the transition zone and associated capillary fringe, where the capillary
forces of the pore fluids can inhibit its further downward movement into finer-grained
materials. In this situation, the contaminant will flow horizontally or build vertically
until its fluid pressure overcomes the capillary pressure in the pore spaces (Zhong et al.,
2001).
7
-------
Upon penetration of the saturated zone, DNAPL will continue to move downward,
displacing groundwater until it reaches an impermeable barrier or a fining textural
interface. Textural interfaces include changes in pore size, permeability, wettability, and
capillary pressures (Bradford et al., 1998).
The dense nature of DNAPL can result in its pooling at low spots in the aquifer base
and migration against the groundwater gradient, resulting in up-gradient contamination
beyond expected diffusion. In addition, any heterogeneities of the aquifer, including
changes in porosity, permeability, capillary pressure, or groundwater flow will alter the
migration of the contaminant (National Research Council of the National Academies,
2005).
Traditional pump-and-treat remediation methods do not appear to be completely
effective in removing DNAPL from the subsurface due to complex migration and the
DNAPL physical characteristics (Kueper et al., 1993; Londergan et al., 2001; Mackay
and Cherry, 1989; Mercer and Cohen, 1990; Qin et al., 2007; Zhong et al., 2001).
Sinking DNAPL displaces fluids from the pore spaces. After the bulk of the DNAPL
volume has moved through an area, the in-situ pore fluid reinvades and fragments the
DNAPL into free-phase pools and disconnected ganglia (Zhong et al., 2001). The
disconnected DNAPL is referred to as residual, which implies that the DNAPL is trapped
in the pore spaces as a result of high interfacial tensions and pore size. Residual DNAPL
is a problem because it can be a source of long-term contamination.
An interface is the boundary between two phases that are immiscible or have low
miscibility. Interfacial tension is defined as the amount of work required to expand an
interface between two phases by a unit area (e.g., Rosen, 2004). If one of the two phases
8
-------
is a gas, this is generally referred to as surface tension. If both phases are liquid, it is
simply termed interfacial tension. This term can also be used to describe the dissimilarity
between the two phases. In general, two similar phases have lower interfacial tensions
than two less similar phases (e.g., Rosen, 2004).
High interfacial tension inhibits the DNAPL from easily transitioning into the
aqueous phase. The interfacial tension between groundwater and DNAPL has been
measured at 20-50 dynes/cm (e.g., Mercer and Cohen, 1990), although interfacial tension
of coal tar has been measured at 0.6 dynes/cm above a pH environment of 9.1 (Barranco
and Dawson, 1999). While free or dissolved phase contaminant may be removed using
traditional pump-and-treat techniques, the removal of residual contaminant requires
impracticably high hydraulic gradients to overcome the capillary pressure of the aqueous
pore fluids (Zhong et al., 2001).
Although residual DNAPL has proven to be resistant to non-traditional groundwater
remediation methods, there are alternanative treatments. Potentially effective non-
traditional methods of removing DNAPL include enhanced bioremediation, air sparging,
in-situ chemical oxidation, and steam enhanced extraction, and surfactant-enhanced
aquifer remediation (SEAR). The work presented here builds on earlier studies
(Werkema, 2008) directed at evaluating the potential of SEAR for DNAPL remediation.
Surfactants and SEAR
A surfactant is a surface active agent, a chemical compound that acts at the interface
between aqueous and non-aqueous fluids (Figure 2). Surfactants are amphiphilic
monomers composed of a hydrophilic head and a hydrophobic tail. The hydrophilic
9
-------
group is ionic or highly polar and determines the classification of the surfactant into
anionic, ionic, nonionic, or zwitterionic. The hydrophobic group is generally composed
of a carbon chain (Figure 3). As a result, surfactants are soluble in both water and
organic solvents (Lowe et al., 1999; Mercer and Cohen, 1990; Sabatini and Knox, 1992;
West and Harwell, 1992).
Surfactants have the potential to be successful DNAPL remediation agents because of
their ability to interact with a NAPL contaminant in two ways. First, surfactants can
decrease the interfacial tension between the aqueous and nonaqueous phases (e.g.,
Adamson and Gast, 1997), thus lowering the force required for the DNAPL to displace
water from a saturated pore (National Research Council of the National Academies,
2005), resulting in increased contaminant mobility. Second, surfactants can also increase
the solubility of nonaqueous contaminants through the formation of micelles (Adamson
and Gast, 1997; Harwell, 1992; Londergan et al., 2001; Lowe et al., 1999; U.S.EPA,
1996). The addition of surfactants to a system above the critical micelle concentration
(CMC) may result in the growth of surfactant monomers into micelles through
aggradation. A micelle is a grouping of monomers of surface active agents (e.g., Rosen,
2004). Fifty to two hundred of these monomers may cluster together to form structures
with hydrophobic interiors and hydrophilic exteriors (Harwell, 1992). NAPL
contaminant molecules can collect in the micelle interiors, while the micelle itself is
soluble in the aqueous phase (Figure 4). This process effectively increases the solubility
of the contaminant by creating a macroemulsion that can be extracted from the
subsurface (Lowe et al., 1999). An emulsion is a suspension of molecules of a liquid that
lies within a second, immiscible liquid in the presence of an emulsifying agent. A
10
-------
macroemulsion refers to the relatively large size of the particles in the suspension, which
must be greater than 400 nm (e.g., Rosen, 2004).
Surfactant-enhanced aquifer remediation (Figure 5), is a promising technology that
utilizes water and surfactant solution floods to remove residual DNAPL from an aquifer,
although it is not yet widely used (Brown et al., 1999; Dwarakanath et al., 1999;
Londergan et al., 2001; Qin et al., 2007; Robert et al., 2006; U.S.EPA, 1996). A typical
surfactant-enhanced pump-and-treat remediation effort begins after the majority of free-
phase DNAPL has been removed from the target area. This removal of free-phase
DNAPL can be achieved through well skimming, vacuum-enhanced recovery
(bioslurping), or water flooding (Lowe et al., 1999). Water flooding is often the most
practical option in preparation of SEAR, as the same equipment can be utilized for the
surfactant floods. It is important to note, however, that every site must be evaluated to
determine the best method of remediation. After the free-phase contaminant has been
removed, surfactant solutions are injected into the subsurface so that they will sweep
through the target area. The surfactant floods increase contaminant solubility as they
sweep through the subsurface. After surfactants have had time to equilibrate, water
floods typically follow in order to flush the system of solubilized DNAPL and surfactant
solution. Multiple pore volumes of surfactant solutions and flood cycles may be
necessary depending on the swept volume (Lowe et al., 1999).
The length of time required to successfully complete a SEAR application will depend
largely on the target zone permeability and heterogeneity, the number of pore volumes
required to treat the area, and spacing between the delivery and recovery wells. While a
full scale operation could take over a year to reach completion, it is believed that the
11
-------
amount of contaminant removed using SEAR is larger than can be removed using another
enhanced pump-and-treat or natural attenuation in the same amount of time (Harwell,
1992; Lowe et al., 1999).
Design of the SEAR process requires identifying the chemical make-up of the
contamination, as well as determining subsurface geology and hydrogeology.
Understanding the chemical system aids in the selection of the surfactant(s), while
understanding the subsurface will help in understanding and predicting behavior of the
surfactant floods (Harwell, 1992; Lowe et al., 1999).
There are several criticisms of SEAR which have limited its use thus far. As with
most pump-and-treat remediation methods, the efficacy of treatment is a function of the
hydraulic conductivity at the site (Fountain et al., 1996). Because of this, sites with low
or heterogeneous hydraulic conductivity will continue to be difficult to remediate,
although some laboratory experiments suggest that the addition of polymer to the
surfactant solution can diminish these problems (Dwarakanath et al., 1999; Martel et al.,
1998; Robert et al., 2006). Most surfactants display uneven wetting surfaces or fronts in
the subsurface. These preferential flow paths, along with the present inability to monitor
surfactant behavior in the subsurface, make it difficult to determine whether surfactants
are reaching the DNAPL-contaminated areas. In addition, most field studies have treated
relatively low amounts of contaminant at a small scale (Londergan et al., 2001), leaving
uncertainties about the effectiveness of using surfactants at larger-scale sites. In addition,
some surfactants can act as bactericides, inhibiting microbial activity and biodegradation
in the subsurface (Bramwell and Laha, 2000; Willumsen et al., 1998). This may affect
ongoing and future bioremediation at a site remediated with surfactants.
12
-------
One of the principal criticisms is related to a surfactant's ability to decrease
interfacial tension. In order to mobilize a contaminant, the interfacial tension between
the aqueous and non-aqueous phases must be lowered to a high degree. This decrease
could result in downward migration of the contaminant through low permeability
barriers, fractures, or faults which had previously not acted as DNAPL conduits due to
the high interfacial tension (Longino and Kueper, 1995). Research in this area has
suggested that surfactant choice and mixture can decrease this problem. In general, a
surfactant that is engineered to increase contaminant solubility will not necessarily result
in a large decrease in surface tension (Harwell, 1992; Pope and Wade, 1995).
Some controlled field studies have experienced significant successes with SEAR,
reporting over 85% reduction in NAPL mass (Fountain et al., 1996; Martel et al., 1998),
and up to 98.5% (Brown et al., 1999; Londergan et al., 2001). There is some indication
that SEAR is not as effective with increasing complexity of a mixed NAPL contaminant,
although evidence for this statement is sparse (Jawitz et al., 1998).
Geophysical Methods
All materials have inherent geophysical and compositional properties which can be
measured with proper instrumentation. These properties include, but are not limited to,
density, electrical and magnetic fields, temperature, and chemical make-up. A wide
range of geophysical methods and techniques for measuring some of these properties
have been developed for application throughout the various branches of geoscience.
Gravity, corresponding to density, and magnetic surveys can be used to locate large or
small-scale anomalies in the subsurface due to density or magnetic property contrasts.
13
-------
Seismic surveys, which utilize acoustic wave properties, can be used to identify
subsurface structure (Lowrie, 2003; Telford et al., 1990).
Scientists have taken advantage of the electrical properties of many groundwater and
soil contaminants in order to monitor the location and behavior of contaminant plumes.
DNAPL plumes have been identified through GPR, IP, and resistivity surveys (Brewster
et al., 1995; Brewster and Annan, 1994; Grimm et al., 2005; Hwang et al., 2008; Sogade
et al., 2006). In particular, geoelectrical methods have been found useful as many
contaminated areas show altered electrical conductivity relative to uncontaminated areas
after the introduction of some pollutants.
Additionally, previous work (Werkema, 2008) indicated that the measured
geoelectrical parameter of specific conductivity showed a larger response to surfactant
presence, while density failed to respond substantially.
Aside from the Werkema 2008 EPA report, there is little in current peer-reviewed
literature that indicates that the geophysical responses of surfactants used in SEAR have
been or are being investigated. There is some indication of research within the petroleum
industry, however. Specifically, the use of high resolution resistivity has been used to
monitor surfactant floods, among other things, in deep formations (Black et al., 2007).
Geoelectrical Methods
Conductivity
All materials have inherent electrical properties including electrical conductivity or
resistivity. Conductivity and resistivity are material properties that are independent of a
material's thickness or geometry. Conductivity is the ability of a material to allow
current to flow through it. Resistivity is its inverse, a relationship defined in Equation 1:
14
-------
a is conductivity and Q is resistivity. Both parameters are measured in per unit
length.
Conductance is an object property, as opposed to a material property. It is also
termed the thickness-conductivity product (Telford et al., 1990) and is the conductivity of
a material that has been corrected for the size and geometry of the object.
Electrical conduction is a broad category that encompasses several types of
mechanisms. Common electrical mechanisms include ionic or electrolytic, surface, and
electronic conduction.
Electrolytic, or ionic, conduction refers to electrical current flow via the pore fluid of
a material and is the most common form of conduction in low-clay, uncontaminated,
water-saturated environments (Figure 6).
Surface conduction refers to the transfer of electricity along the fluid-grain interface
and the electrical double layer (if present), and is a function of surface charge density,
grain surface area, and ion mobility (Endres and Knight, 1993; Lesmes and Frye, 2001;
Marshall and Madden, 1959; Revil and Glover, 1998; Schwarz, 1962; Vinegar and
Waxman, 1984). The surface conduction mechanism moves current through the
electrical double layer (EDL), a small region adjacent to the grain surface (Figure 7).
The EDL is often associated with clay materials and can also develop in the presence of
biodegradation (Aal et al., 2004; Atekwana et al., 2004). The double layer is composed
of two layers: a single fixed layer of ions adhered to a grain's surface and a diffuse layer
of ions that exists adjacent to the grain's surface. Ions can move across the diffuse layer
15
-------
in a process called surface conduction, which tends to be a slower mechanism than ionic
conduction.
Clays and other phyllosilicates, unlike quartzitic sand, are not electrically inert. They
are capable of cation exchange due to their sheetlike mineral structure (Schoen, 1996),
and have the potential to impact geoelectrical response. Because of this structure, there is
typically water trapped between the sheets, affecting conductivity. In addition, clays
typically have a negative surface charge, which enables them to adsorb ions at their
surfaces. Depending on the charge balance of the clay and the ions available for
adsorption, the clay may form an electrical double layer instead (Schoen, 1996).
Biodegradation has several mechanisms by which surface conduction can be
increased. These include excess charge build up in the fluid-grain interface and the
potential for the microbes themselves to become polarized. Additionally, the increased in
microbial colonies may result in a build-up of organic acid in the subsurface. The
organic acid can increase etching of the grains, likely resulting in an increase in surface
area, which is a partial control on surface conduction (Aal et al., 2004; Atekwana et al.,
2004). It should be noted that the phenomenon of microbial-enhanced surface
conduction is not well understood at this point in time.
Electronic, or ohmic, conduction can occur in the presence of metallic ions as a result
of vibrations in the lattice (Howarth and Sondheimer, 1953). The free electrons in
metals, and sometimes crystals, acquire a common drift velocity when an electrical field
is applied. This slows and directs the electrons in the direction of the field. Resistivity
by this mechanism is determined by the free time between collisions of the electrons into
16
-------
the metal atoms. More frequent collisions means that resistivity is higher, while fewer
collisions result in lower resistivity (Lowrie, 2003).
In the absence of clays, DC resistivity measurements can be related to the
geoelectrical response of pore fluids to that of the matrix through Archie's Law (Eq. 2):
Pe=a* Pw*~m (2)
pe is bulk resistivity, pw is resistivity of the pore fluid, (p is the porosity of the matrix,
and a and m are empirical parameters relating to cementation (Archie, 1942).
In this context, the matrix refers to the solid materials in an aquifer, as well as the
chemical and physical properties due to the solids. These properties include porosity,
grain size, shape, composition, and sorting. While Archie's Law appears to be valid
when conduction is primarily through pore fluids, it does not describe the role of surface
conduction in bulk resistivity.
Archie's Law can be rearranged (Eq 3) to create a formation factor, FF, which is the
portion of bulk resistivity that incorporates the matrix.
FF = — = —(3)
~ J.v y
Pw
As stated previously, Archie's Law assumes conduction is through the pore fluid
alone, known as electrolytic conduction. However, surface and electronic conduction are
also common methods of conduction and capacitance (i.e. charge storage). Archie's Law
is often modified to include a surface conduction term because of these additional
conduction mechanisms, as well as the presence of clays in many aquifer materials
(Waxman and Smits, 1968).
17
-------
Alternating current conductivity (complex conductivity) is a complex parameter (Eq.
4), with real and imaginary components, as it consists of both a magnitude and a
direction. The direction is referred to as phase (e.g., Zonge et al., 2005).
a*=a'+ia" (4)
This equation describes the relationship between complex conductivity and the real
and imaginary components, o* is the complex conductivity, o' is the real component, and
o" is the imaginary component of conductivity. Real conductivity is, in essence, the
total or bulk conductivity of the system. It takes into account electrolytic conduction,
surface conduction, and any electronic conduction. Imaginary conductivity is the
component of the measured bulk conductivity that results from polarization of ions at the
fluid-grain interface. When a current is applied to some materials, polarization at the
fluid-grain interface occurs, separating the anions from the cations (Figure 8). When the
current is turned off, the ions re-equilibrate along the interface. In the time domain, this
polarization and re-equilibration appears as a decay curve over time. In the frequency
domain, the polarization appears as a frequency-dependent phase shift, or change in
angle, of the received sine wave relative to the transmitted signal (e.g., Zonge et al.,
2005).
Spectral Induced Polarization
During spectral induced polarization (SIP), current is induced in the subsurface or
experimental sample over a range of frequencies. SIP measures the resistivity magnitude
and phase as functions of frequency, which can then be used to calculate real and
imaginary conductivities (Eq. 5 and 6).
18
-------
The calculations for real and imaginary conductivities utilize a phase shift term. The
phase shift is the difference between the phase of the transmitted sine wave signal and the
phase of the received sine wave signal. A large phase shift, suggests that some of the
transmitted current was attenuated during conduction. A delayed current could indicate a
change in conduction mechanism from electrolytic conduction to surface conduction or
polarization because of their slow speeds relative to electrolytic conduction. It could also
correlate to a change in chemistry or materials encountered during testing (e.g., Zonge et
al., 2005). SIP assesses the frequency dependence of this response, which may be
indicative of a specific material or set of conditions in the subsurface (e.g., Zonge et al.,
2005).
Equation 5 (e.g., Zonge et al., 2005) shows the calculation of real conductivity from
the conductivity magnitude and phase shift from a SIP reading, o' is the real conductivity
component and
-------
Dielectric Constant and Time Domain Reflectometry
Electrical permittivity is a dimensionless term that describes an ion's ability to
transmit charge or polarize due to an applied electric field in a particular medium (e.g.,
Lowrie, 2003). At some frequencies of alternating current, polarization occurs, resulting
in a modification of effective conductivity (Lowrie, 2003). Normally, electrons are
distributed symmetrically around an atom's nucleus. When an electrical field is applied,
the electrons are displaced in an opposite direction to the field, while the nucleus shifts in
the same direction as the field (Lowrie, 2003). As a result, the permittivity of the
material is different from that of free space.
Dielectric constant, often referred to interchangeably with relative permittivity, is a
dimensionless term that describes the relationship between electrical permittivity of free
space and electrical permittivity of a medium (Lowrie, 2003) (Eq. 7).
s = K * s0 (7)
s is the permittivity of a medium other than free space, s 0 is the permittivity of free
space, and K is the dielectric constant. Alternately, K can be represented by s r,
indicating the relative permittivity. Equation 7 is then rearranged and substituting s r for
K yields the ratio between permittivity of free space versus permittivity of another
medium (Eq. 8) (Lowrie, 2003)
— = er (8)
^0
The relative permittivity, or dielectric constant, can be represented by the complex
parameter K*, which, as in complex conductivity, consists of real and imaginary
components. The real component, K' describes energy storage, while the imaginary, K"
describes energy loss. K* has been shown to be frequency dependent in some
20
-------
environments (Kelleners et al., 2005), as described in Eq. 9 (Topp et al., 1980). This
frequency dependence also modifies the effective conductivity through the following
relationship:
K* is the complex dielectric constant, K' is the real component of the dielectric
constant, K" represents the loss due to frequency-related relaxation mechanics, oac is the
zero-frequency conductivity, and / is frequency.
In direct current and low frequency environments, dielectric effects are considered
negligible. In an environment with an alternating electrical field however, polarization
changes with frequency, thus resulting in fluctuating polarization and effective
conductivity.
Time domain reflectometry (TDR) is a method used to measure the apparent
dielectric constant (or electrical permittivity) of a medium by sending a pulse of
electromagnetic energy at 746 kHz through a transmission line embedded in the medium.
During travel, the beam reflects off of discontinuities in the host material. When the
pulse reaches the end of the line, it reflects most of the remaining energy (Dalton et al.,
1984; Soilmoisture Equipment Corporation, 2005; Topp et al., 1980). The travel, or
transit, time is recorded and used to determine the apparent dielectric constant in
Equation 10 (Soilmoisture Equipment Corporation, 2005).
K* = K'+i K"+
(9)
y2 J
(10)
21
-------
Ka is the apparent dielectric constant, t is the transit time, c is the speed of light, L is
the length of the transmission line.
The dielectric constant can be used to determine moisture content of the host material
through the empirical Topp Equation, illustrated in Equation 11 (Topp et al., 1980).
Ka = 3.03 + 930v +146.O0v2 -16.101 Q1)
Ka is the apparent dielectric constant and 9v is the volumetric moisture content.
It is also a physical property that is a factor in ground penetrating radar (GPR)
transmission, as is conductivity. Electromagnetic wave propagation velocity and
reflection interfaces are strongly influenced by dielectric constant (Martinez and Byrnes,
2001). The relationship between the velocity of wave propagation and the dielectric
constant is described in Equation 12 (Martinez and Byrnes, 2001).
V is the velocity of wave propagation, c is the speed of light in a vacuum, and 8 is the
permittivity of the material.
Dielectric constant is a direct and indirect factor in numerous geophysical methods,
including GPR, which has been used successfully to map DNAPL (Brewster and Annan,
1994).
Water Quality Measurements
Dissolved Oxygen
Dissolved oxygen (DO) refers to the amount of oxygen that is dissolved in water.
The range of DO in natural water is between 0 to 10,200 |ig/L (Borden et al., 1995;
22
-------
Kreamer, D.K., personal communication, November 2009). DO concentrations in a
system affect chemical and biological reactions that depend upon available oxygen.
Changes in redox or other conditions may alter surface conduction, resulting in a
geoelectrical response (Werkema, 2008). Identifying changing DO in conjunction with
conductivity, resistivity, and phase shift measurements, will aid in determining its impact
on the geoelectrical response.
Understanding the change in subsurface dissolved oxygen as a result of surfactant
application is important for multi-pronged remediation efforts. If biodegradation is being
considered as a remediation process to follow SEAR, understanding how oxygen content
is changing is imperative.
pH measures the activity of hydrogen ions in a system, and ranges between 0 and 14.
pH is the cologarithm (i.e. colog) of the activity of dissolved hydrogen ions (Eq. 13).
Acidity increases with smaller numbers, and larger numbers are increasingly alkaline.
A measurement of 7 is considered neutral. pH values of natural waters typically range
between 6.2 and 8.0 (Hoyle, 1989; Kehew and Passero, 1990; Nicholson et al., 1983).
The pH of a system can affect chemical and biological behavior. Low pH, indicative
of an acidic system, or high pH, indicative of alkalinity, can affect which microorganisms
will be present in an environment, how much chemical weathering of solids will occur,
and the behavior of a contaminant plume in the subsurface, as well as the complex
conductivity response (Olhoeft, 1985).
pH
(13)
23
-------
Lower pH in the subsurface has been associated with a higher degree of etching on
the solid materials and grain surfaces (Atekwana et al., 2004; Sauck, 2000), which could
result in a change in conductivity mechanism. As discussed previously in the chapter,
this could result in a change in measured real or imaginary conductivity.
Specific Conductivity
Specific conductivity electrolytic conduction, or conduction by movement of ions
through pore fluid. Specific conductivity measures electrical conduction through a
medium that is under the influence of an applied electrical field. The range of specific
conductivity values of typical natural waters is between 40 and 400 |iS/cm (Williams et
al., 1993).
Specific conductivity is a component of real conductivity. As such, the ability to
compare any changes in specific conductivity with changes in the resistivity measured
with SIP is a powerful tool which can help us to understand the importance of the
different conduction mechanisms taking place in the subsurface.
Statistics
When reporting experimental results, it is important to be able to communicate the
relevant information in a meaningful way. Statistical methods are helpful in describing a
data set's overall character, relationships among the points in a data set, and relationships
to a predictive model for large data populations. Common statistical evaluations include
a data set's mean, standard deviation, and R2 value. The following section will describe
some of the statistical tools used in this project.
24
-------
Regression Analysis
Regression analysis is a method used to fit a mathematical model to a data set
(Anderson and Whitcomb, 2000). It is a way to produce a statistical model describing the
relationship between a dependent variable, or response, and one or more independent
variables, or factors (Kleinbaum et al., 1998). Regression analysis is useful for
characterizing relationships, finding a quantitative formula to predict the trend of a
response, or evaluating interactive effects of multiple factors on a response.
Proper experimental design and data analysis can result in statistical models, which
then require careful consideration due to the inherent noise in all data sets. Alternately, a
deterministic model, such as the equation to find a falling object's velocity on Earth,
lacks error. It is considered a perfect mathematical model because the response (velocity)
varies exactly as predicted as the model is derived analytically rather than empirically
(Kleinbaum et al., 1998).
Analysis of Variance
Analysis of variance, also referred to as ANOVA, is a statistical method that assesses
the significance of experimental results through evaluation of a data set's variance.
Variance is the measure of spread or variability in a data set (Anderson and Whitcomb,
2000). Many basic statistical parameters, including standard deviation (o), coefficient of
variance (C.V.), and the multiple correlation coefficient (R2) values, are based on
estimates of several components of variance. The calculations of these fundamental
statistical variables are either derived directly from variance, or indirectly through
parameters like the sum of squares (Kleinbaum et al., 1998).
25
-------
The term "analysis of variance" is derived from the method of determining the
ANOVA statistics. The total variability within a data set is partitioned into separate
components, which are used to calculate useful parameters like the sum of squares
(Montgomery, 1997). The sum of squares terms are included in the ANOVA. A sum of
squares is the sum of the squared distances of each data point from the data set mean
(Anderson and Whitcomb, 2000). It can be separated into several components, including
the total sum of squares (SStot), the sum of squares between treatments (SStrt), the sum
of squares due to error within treatments (SSe), and several others (Montgomery, 1997).
The R2 value is the multiple correlation coefficient. It ranges between 0 and 1 and
provides an estimate of the overall variation in the data set which is accounted for by a
proposed statistical model (Anderson and Whitcomb, 2000). More successful predictive
models will maximize the R2 values. It is calculated using sum of squares in Equation 14
(Montgomery, 1997).
VV - VV
ft2 — tot (\4\
VV
R2 is the multiple correlation coefficient, SStot is the total sum of squares, and SSe is
the sum of squares due to the error or residuals.
There are three important versions of the R2 parameter which should all be examined
to determine the relevancy of a variable to the data set and the proposed model. The first
version is the simple R2, discussed above. The adjusted R2 and predicted R2 values are
described below. Ideally, all three of the different R2 values discussed in this section
would be maximized and in close agreement. Values that differ greatly could indicate a
problem in the experimental design.
26
-------
R2 is affected by the number of independent variables, or terms, included in the
statistical model, and has a tendency to increase with the number of variables regardless
of whether all terms are significant (Kleinbaum et al., 1998). Because of this, models
with large R2 values may actually be poor predictors of a response (Montgomery, 1997).
The dependence of R2 on the number of independent factors has resulted in the
development of the adjusted R2 (R2adj)- This term is the multiple correlation coefficient
which is corrected for the number of model terms and points in the design. In general, if
irrelevant terms are added to a model, the R2adj value will decrease; the more R2 and R2adj
differ, the more likely it is that non-significant terms have been added (Montgomery,
1997).
A third version of the R2 value is predicted R2 (R2pred) (Eq. 15) (Montgomery, 1997).
This parameter describes the amount of variation in the predicted data set that cannot be
explained by the model, and makes use of the predicted residual sum of squares (PRESS).
PRESS is a measure of how well a statistical model fits each point in the design and is
determined by repeatedly fitting the model to each of the design points except for the one
that is being predicted. The difference between the predicted and actual value of each
point is squared and summed, resulting in the PRESS (Anderson and Whitcomb, 2000).
In short, it is the sum of squares of the PRESS residuals (Montgomery, 1997).
2 PRESS
pred gg (15)
PRESS is the predicted sum of squares and SSy is the sum of squares of the response.
27
-------
Experimental Design
Considered a way to develop and perform more effective experimentation, DOE
(design of experiments) has been used in manufacturing for several years. While in the
past, users were required to set up experiments to maximize or minimize responses based
on limiting factors, several software programs now exist that can be operated on ordinary
personal computers. Most of this software is able to not only aid in the set up of
experimental designs, but also to perform statistical analysis on the experimental data.
The experimental agenda for this project was designed with the help of Design Expert
7.0 (Stat-Ease, 2007), an experimental design statistical software used mainly in
manufacturing and industrial engineering to optimize performance through combinations
of factors (Anderson and Whitcomb, 2000). Experimental designers must be careful not
to use this software as a type of "black box" utility, however. In order for the software to
suggest a design that will maximize potential response, the experimenter must understand
the components and styles of experimental design.
A factor is a variable, ideally assumed to be independent of any other testable factors,
that is manipulated during an experiment to examine its effect on responses. A response
is a measurable product or effect that is thought to be affected by the experimental factors
(Anderson and Whitcomb, 2000). There are two common types of factors: categorical
and numerical. A categorical factor is one which has conditions that represent discrete
levels or options (Anderson and Whitcomb, 2000). For example in this experiment,
water type is either tap or deionized, with no other steps or possibilities in between
considered. A numerical factor is a quantitative variable that can be adjusted through a
28
-------
continuous range (Anderson and Whitcomb, 2000). Temperature and surfactant
concentration are examples of this type of factor.
There are several different styles of experiments, ranging from simple comparison to
the more complex response surface methods. The simplest form, the F-test, compares
two or more discrete levels of a single factor by evaluating the variance among the
treatments and comparing it to the variance among the individual repetitions within each
treatment. It is considered a one-factor design (Anderson and Whitcomb, 2000).
If the factor has little to no effect on a response, the F-ratio will be close to 1. The F-
ratio, or F-value, is a ratio used in the ANOVA (discussed below) that is derived from the
F-test. It is essentially the ratio of the difference in response between the treatments
compared to the experimental noise. As a factor's influence on a response increases, the
F-ratio will also increase, while decreasing the chance that the suggested correlation is
due to chance or noise. The p-value, derived from the F-ratio, is a parameter that sets a
quantitative value on the probability that the correlation is due to noise. A p-value less
than 0.05 (5%) indicates there may be a significant relationship between the tested factor
and response (Anderson and Whitcomb, 2000).
The factorial group of designs is more complex than simple comparisons. These
designs allow experimentation on multiple factors at multiple levels. The simplest of the
factorial group of designs is 2 factors with 2 discrete levels. One of the advantages of the
factorial designs is that they can require fewer experimental runs to produce statistically
valuable results than experimental designs that test only one factor at a time. As such, the
more factors and levels involved, the more advantageous it is to utilize a factorial design
(Anderson and Whitcomb, 2000).
29
-------
The nature of the factors can affect which experimental design type to implement. As
described above, factors can be categorical or numerical. An experimenter with only
categorical factors may find it best to perform a general factorial design. In this style of
design, all of the possible combinations of factors are run (Anderson and Whitcomb,
2000).
In addition to factor and response consideration, experimental designs must also
consider the effects of environmental changes that cannot be easily controlled. These
include slight temperature changes that will affect conductivity measurements, diurnal
effects, and instrumental drift. These variables can be accounted for through the use of
blocking. Blocking is a DOE technique that divides a suite of experiments into packages
that can be performed in a single time period. If a specific block has higher or lower
measured values, they can be adjusted for during statistical analysis (Anderson and
Whitcomb, 2000).
The project described in this thesis utilized a 2FI, or a two-factor interaction design.
This is a version of a general factorial. In this design, the two factors are surfactant
treatment and water type. Both are categorical. Traditional experimental design does not
identify or detect interactions between factors or that interaction's effects on the tested
responses. The 2FI design allows to evaluate the interactions of two factors at various
discrete levels and the responses associated with them (Anderson and Whitcomb, 2000).
The experiments were divided into three separate blocks of 10 experiments each. Each
block was performed over 2 days.
This chapter has detailed some of the relevant background information that will be
helpful in understanding the methods used in the project, as well as some of the reasons
30
-------
behind attempting this particular type of research. The overall objective of the research
was to gain additional insight into the geophysical and physico-chemical responses in a
quartz sand-water environment in the presence of surfactants and to form a simple
predictive model for each response. Changes in the measured parameters could indicate
the feasibility of using geophysical field methods to monitor surfactants used in
groundwater remediation of DNAPLs. The following chapter will detail the methods and
materials used in the research described in this paper.
31
-------
CHAPTER 3
MATERIALS AND METHODOLOGY
This chapter provides an explanation of the techniques and materials used in the
project documented in this thesis. The constituents and construction of the experimental
column are described, in addition to the methods used for the spectral induced
polarization (SIP), water quality, and time domain reflectometry measurements. The last
section of the chapter details the statistical analyses performed on the collected data,
including how the data was manipulated and evaluated for integrity.
Experimental Column
A simplistic analog of an aquifer environment was created using electrically inert
solid materials and an electrolyte. In this research, the term "analog aquifer" refers to a
simulated aquifer environment made with clean quartz sand acting as aquifer solid
material and the experimental solutions acting as pore fluid. It is considered analog
because it is not a true aquifer environment. The testing conditions in the proposed
experiments are surfactant-saturated, quartz sand environments. All experiments were
performed using Ottawa silica sand as the aquifer matrix material. According to U.S.
Silica Company (1997), sieve testing places 99% of the sand at 20-30 sieve size (0.600-
0.850mm diameter), and chemical analysis places the quartz content at Si02 99.8%. Of
the remaining 0.2%, 0.1% includes 0.02% Fe203, 0.06% AI2O3, 0.01%TiO2, and less than
0.01%) each of CaO, MgO, Na20, and K20. The remaining 0.1% was lost on ignition
through analysis (U.S.Silica Company, 1997). The sand conforms to American Society
for Testing and Materials C778, a standard specification.
32
-------
Surfactants were mixed into separate deionized and tap water solutions at
concentrations commonly used in the field, resulting in 10 experimental solutions (Table
1). The surfactant formulas chosen for this project have been successfully used in field
and laboratory studies (Londergan et al., 2001; Ramsburg and Pennell, 2001; Rothmel et
al., 1998) and include Aerosol MA-80-I (AMA-80-I), Dowfax 8390, and Steol CS-330.
All of these surfactants are non-ionic and displayed the highest, median, and lowest
response, respectively, in Werkema (2008). These responses are shown in Table 2. A
control with no surfactant was also tested for this thesis. Concentrations used were 8%
AMA-80-I, 5% Dowfax 8390, 0.5% Dowfax 8390, 0.025% Steol CS-330, concentrations
that have previously been used in field and lab-based studies (Londergan et al., 2001;
Ramsburg and Pennell, 2001; Rothmel et al., 1998). SIP experiments were also
conducted on DI solutions of 8% Dowfax 8390 and 8% Steol CS-330 in order to compare
the surfactant responses to one another.
Tap water was sourced from a single spigot in the U.S. EPA POS building, Room 21
on the U.S. EPA's Las Vegas campus. The water measured a specific conductivity value
of 1025 |iS/cm at the time that solutions were mixed. Deionized water was sourced from
the DI system located at the Quality Assurance Lab, also on the U.S. EPA's Las Vegas
campus. The DI system is monitored on a weekly basis by U.S. EPA contractors and is
rated to 18 MQ. All solutions were mixed in a 1 L Ehrlenmeyer flask and stored in
cubitainers. Surfactant was added to the flask first, followed by the water. The flask was
capped and swirled to mix the components. In the case of the 5% and 8% solutions,
water was added to the flask 300 to 400 mL at a time. Swirling followed each addition to
more efficiently and uniformly mix the components.
33
-------
Spectral Induced Polarization
For the SIP experiments, the analog aquifer was confined to a custom-made PVC pipe
apparatus engineered to hold the aquifer material (Figure 9). It is similar in design to
columns used by several other researchers (Slater and Glaser, 2003; Slater and Lesmes,
2002; Vanhala and Soininen, 1995). The column, built on site, is 18 cm in length, 3.5 cm
in diameter, and uses 4 silver-silver chloride (Ag-AgCl) electrodes, 2 current and 2
potential, to make the SIP measurements. Ag-AgCl electrodes were chosen because they
have been shown to produce minimal surface impedance and voltage drop over time
(Vanhala and Soininen, 1995). High surface impedance can result in large phase shifts,
masking the true response (Vanhala and Soininen, 1995). To coat electrodes, 14 gauge
fine silver (99.9%) wires, were cut and shaped, then soaked in bleach (NaCIO) overnight.
Current electrodes were coiled into a disk shape of slightly less than 3.5 cm diameter,
while potential electrodes were straight lengths of wire cut to 4.5 cm long (Figure 9).
The potential electrodes are housed outside of and at right angles to the main
experimental column within 0.8 cm diameter PVC pipe, and isolated from the solid
aquifer materials by a 150 micron nylon mesh. The purpose of the mesh is to allow the
electrolyte to submerge the potential electrodes, without allowing contact with the solid
materials in order to avoid polarization of the electrodes.
The PVC column was sealed with rubber stoppers (Figure 9). Two stoppers (3.5 cm
in diameter) were fitted in both ends of the cylinder and sealed into place with electrical
tape. Two smaller stoppers (0.8 cm in diameter) were inserted into the ports that were
designed to house the potential electrodes. All four stoppers described above were
configured with one electrode each. A stopcock was placed at each end of the
34
-------
experimental column to enable saturation and flow. Stopcocks were inserted into pre-
drilled holes in the large stoppers and glued into place. In order to inhibit movement of
aquifer material out of the column, a piece of 150 micron nylon mesh was glued over the
internal portion the stopcock pipe.
Prior to a SIP measurement, the column was prepared for measurements. The empty
column was weighed, followed by a zeroing of the balance. 286 g (±2.5 g) of Ottawa
silica sand was added to the column, weighed, followed by a re-zeroing of the balance.
62 g (±2.0 g) of solution was then injected into the column and the column was re-
weighed. Optimization of saturation was achieved by injecting the solution into the
bottom stopcock of the vertical column. Saturation was assumed when the solution
escaped at the top stopcock and the potential electrode chambers were filled with
solution. Consistent packing was ensured by using a tapping method while adding sand
to the column. The columns equilibrated for 15-30 minutes prior to SIP measurement
collection.
Systematic error tests were conducted over two days on five columns to check for
potential errors due to electrodes, column construction, or packing procedure. As a total
of eighteen columns were used during SIP testing, 27.8% of experimental columns were
tested for systematic error. Random columns were chosen and filled with equivalent
masses of sand and tap water, with random current and potential electrodes. Tap water
was pulled from the spigot on day one and was stored in a cubitainer for the remainder of
the systematic error tests. Resistance and phase were measured. It was determined that
an error of ±5.1% should be considered with the real conductivity data, and ±4.6% error
35
-------
should be included with the imaginary conductivity data. Data and the method for
calculating percent error from the systematic error tests can be found in Appendix A.
Electrical properties were measured with the SIPLab II®, a multi-electrode
acquisition system developed by Radic Research, Inc. (Radic Research, 2007) to measure
SIP, or more specifically, the impedance magnitude and phase shift of the materials in the
column through a spectrum of alternating current frequencies. The SIP equipment
generates and transmits a sinusoidal current which can sweep through a range of
frequencies from 1 mHz to 12 kHz. The SIPLab II® measures from a 4 electrode
configuration, in this case a Wenner array (Figure 10), and has the ability to apply current
and measure multiple electrode configurations in quick succession (Radic Research,
2007). Current was applied through the coiled current electrodes, and resistance
magnitude and phase shift were recorded between the potential electrodes at 18
logarithmically-spaced frequencies between 0.091 Hz and 12 kHz. The SIPLabll®
makes 32 measurements at each frequency. The recorded response is an average of these
measurements (Radic, T., personal communication, September 2009).
pH, Dissolved Oxygen, Specific Conductivity
After SIP measurements were completed, the saturated column was attached to a low
flow cell to take pH, DO, and specific conductivity measurements. The closed
configuration allowed DO measurements by avoiding degassing and mixing with ambient
atmospheric conditions. Flow was driven by an Ismatec (IDEX Corporation, 2007) low-
flow peristaltic pump through the circuit outlined in Figure 11. The tubing utilized has an
inner diameter of 1.6 mm, and the initial flow rate was set at 9.6 mL/min. Flow moved
36
-------
from a surfactant reservoir per each surfactant concentration, through the saturated
column, and into the low flow cell, which housed a Troll 9500®, until the flow cell was
completely filled. The Troll 9500®, produced by In-Situ, Inc. (In-Situ, 2008), is a multi-
parameter water quality monitoring system with pH, specific conductivity, optical
dissolved oxygen, and temperature probes. Temperature was not controlled during these
experiments, but was recorded with the water quality measurements in order to identify
any changes. It was not analyzed as a parameter of interest. Changes in environmental
conditions, particularly ambient temperature, can have large effects on dissolved oxygen
and specific conductivity values.
Systematic error tests were conducted on four columns in order to check for potential
errors due to column construction or packing procedure. As a total of fifteen columns
were used during water quality testing, 26.7% of experimental columns were tested for
systematic error. Random columns were chosen and filled with equivalent masses of
sand and tap water. Specific conductivity, pH, and dissolved oxygen were measured and
recorded. Based on the calculations located in Appendix A, errors of ±0.37% should be
applied to the pH parameter, 10.1% to the specific conductivity data, and 4.3% to the
dissolved oxygen data. Data and calculation methods used to determine the errors can be
found in Appendix A.
The optical DO probe does not utilize ion exchange or consume oxygen, thus
allowing accurate conductivity measurements both before and after DO measurements, as
the ion concentration in the electrolyte will not change. The flow cell allows more
accurate DO measurements by inhibiting degassing of the water and mixing with ambient
atmospheric gases, resulting in contamination.
37
-------
After initiation of circulation by pumping, the low flow cell filled within 70 minutes.
Pumping rate was then lowered to 3.0 mL/min, and DO, pH, and conductivity readings
were recorded at 0, 15, and 30 minutes. In systematic error tests, measurements were
made at 0, 15, 30, 45, and 60 minutes, but it was determined to be unnecessary to
continue readings past 30 minutes as readings appeared to change negligibly. Calculated
systematic errors, discussed above, can be found in Appendix A.
Time Domain Reflectometry
Time domain reflectometry measurements were performed in a 30.5cm diameter,
30cm tall PVC column. Each column was filled with 8.77 kg (±0.2 kg) of quartzitic
sand, and saturated with 1880 mL (±150 mL) surfactant solution from the bottom of the
column by gravity flow (Figure 12). The sand used in the TDR experiments is of the
same type as described in the Spectral Induced Polarization section above, as are the
surfactant solutions. The TDR instrument utilizes a square wave with a period of 1.34
|iS, correlating to a frequency of 746 kHz (van Calker, A., personal communication,
September 2009). Three readings were taken immediately upon saturation and averaged.
The dielectric constant was measured at a 10 ps sampling resolution, through a 10 ns
window. At 10 minutes, three more measurements were made and then averaged.
Systematic error tests were conducted on four columns in order to check for potential
errors due to column construction and packing procedure. As a total of five columns
were used during TDR testing, 80% of experimental columns were tested for systematic
error. Columns were chosen and filled with equivalent masses of sand and tap water.
Moisture content and dielectric constant were measured. An error of ±3.6% should be
38
-------
applied to the dielectric constant response. Data from the systematic error tests can be
found in Appendix A
Readings were taken using a SoilMoisture, Inc. MiniTrase (Soilmoisture Equipment
Corporation, 2005) time domain reflectometer and a 8 cm three-prong uncoated, buriable
waveguide (Figure 13) that was inserted vertically into the saturated sand up to its cable
attachment.
Statistical Analysis
After completion of the experiments, results were transferred from a written lab
notebook to the Stat-Ease program. Statistical analyses were performed separately on
each measured response. For each test, the ANOVA table was examined, with particular
attention paid to the p-value and three R2 values. The table contained information
including sum of squares, F-value, and p-value. Information was given for the two
factors, surfactant and water type, as well as for the model, residuals, and corrected total.
The mean, standard deviation, correlation variable, PRESS, and R2 values were also
included. In addition to the ANOVA parameters, a predictive model was presented by
the software. The predictive model offered an estimated value of a given parameter
based on the measured data set. After the ANOVA output was evaluated, the data was
examined graphically for normality and homoscedasticity of variance.
A series of plots using residual data were prepared in order to evaluate the robustness
of the data set. The Box-Cox plot was evaluated for the potential use of a data transform,
in addition to examining the maximum to minimum ratio of the data set. Typical analysis
requires that the data set is normally distributed and homoscedastic. Homoscedasticity is
39
-------
also referred to as homogeneity of variance, and implies that variance is constant across
the data range (Kleinbaum et al., 1998). If these two conditions are not fulfilled, it may
be necessary to apply a power transform over the data set. The Box-Cox plot is a tool
used in the Stat-Ease software that aids in determining whether a power transformation
would be helpful or necessary in the analysis of a set of data, along with suggesting
which one should be applied (Stat-Ease, 2007).
On a Box-Cox Plot, the x-axis is Lambda (X), while the y-axis is the residual sum of
squares. Lambda is the power by which a transform would minimize the residual sum of
squares. Ideally, the residual sum of squares is minimized, as a totally homoscedastic
data set would have a sum of squares equal to zero (Box and Cox, 1964). The lowest
point on the plotted curve gives the X value to use in the power transformation.
Transforming the data by the power X should create the most stable variance over the data
set (Stat-Ease, 2007). In the case that a transformation is applied to the data set, either
because of a maximum to minimum ratio larger than three or the Box-Cox plot indicates
the benefit of one, analyses are repeated with the new conditions.
To determine whether the normality assumption is valid, a plot of residuals versus
normal percent probability was evaluated. Ideally, residuals plot along a straight, 45
degree line (Stat-Ease, 2007). If normality was deemed valid, plots were next examined
for trends in the data that could be attributed to experimental design or changes in the
environment. For example, a plot of the residuals versus run number can help evaluate
the potential of instrument drift over the duration of all experiments or between blocks.
A series of influence plots, including Cook's D and leverage, were next evaluated for
evidence of any individual runs that were unduly influencing the data set statistics.
40
-------
Cook's Distance (Cook's D) is a statistical combination of the leverage and t-test
influence parameters (Cook, 1977). Like other influence tests, it describes how a single
point affects a model and serves as a criterion for exclusion of outlying data (Kleinbaum
et al., 1998). The Cook's D for a point is determined by measuring how much a
predictive model would change if the data point was removed. Large Cook's D values
are generally associated with high leverage values and large studentized residuals (t-test)
(Anderson and Whitcomb, 2000). A large relative Cook's D value may indicate an
outlier and should be examined further. A large Cook's D alone is not enough reason to
exclude a data point. In this project, if outliers were identified, analysis was started over.
If an outlier was found in one response, the data for that run was excluded from analyses
of all responses.
After confirming normality and other assumptions, plots of the interactions between
factor and response were evaluated. Each response had a single interaction graph with
the data set divided by water type into two plots. Surfactant treatment was plotted along
the x-axis and response on the y-axis.
The goal of this chapter was to provide the reader with the materials and methods
used in this project. In addition, the reader should now be aware of which statistical
analyses were used and how data was evaluated for robustness.
41
-------
CHAPTER 4
RESULTS
This chapter presents the results of the previously described experiments. The
chapter is separated into sections by measured parameter: real conductivity, imaginary
conductivity, pH, specific conductivity, dissolved oxygen, and dielectric constant. Each
parameter's section will include ANOVA and predictive model output, along with
normality, influence, and interaction plots. Through analysis, a predictive model was
proposed for each parameter using Stat-Ease, Inc. Design Expert v 7. The classical Sum
of Squares method was utilized for analysis and model development. The model is
presented as a final predictive value for each surfactant treatment in each parameter.
Each component is described and explained, with discussion to follow in the next
chapter. A response table (Table 3) lists, in run order, the unaltered responses of each
experimental run.
Data from 11.7 Hz readings was analyzed for the real and imaginary conductivities.
While the measured frequency range is between 0.091 Hz and 12 kHz, the frequencies
most often used in applications are between 0.1 and 10 kHz (Vanhala, 1997).
Frequencies outside of this range have a tendency to produce a large amount of noise that
can obscure the true response (Vanhala, 1997). After examining the real and imaginary
responses at several frequencies between 1 and 100 Hz, it was determined that 11.7 Hz
had the most stable values. For the raw data, please see Appendix B.
Runs 13 and 22 have been omitted from all analyses. In both runs, the imaginary
conductivity is negative, a physical impossibility, and correspond to 5% Dowfax, tap and
8% AMA 80-1, tap, respectively. Because of the omissions of these runs, 5% Dowfax,
42
-------
tap and 8% AMA 80-1, tap statistics were calculated based on two runs of each type
rather than three. In the cases where statistics were calculated on two runs, the raw data
was examined to determine whether the remaining runs were consistent.
Real Conductivity
The real conductivity response ranges from 12.06 [j,S/cm to 1910.83 [j,S/cm, with a
maximum to minimum ratio of 158.4. A ratio greater than 10, as seen in this response,
may indicate the potential benefit of a transformation (Stat-Ease, 2007). In addition, the
Box-Cox plot suggests a log transform (Figure 14), which was performed. Analysis of
both logio(real conductivity) and ln(real conductivity) produce the same statistical result.
The ANOVA for real conductivity is found in Table 4.
The mean of the untransformed data set is 561.93 [j,S/cm with a standard deviation of
672.50 [j,S/cm. The large standard deviation, relative to the mean value, is most likely
due to the large range of values in responses. The R2 value is 0.9997 with a predicted R2
value, based on the proposed model, of 0.9992. High quality models produce high R2
values in both categories. These two values are close to 1.0 and in close agreement,
indicating that the model may be a good predictor of real conductivity. The modeled
values for logio(real conductivity), as well as the untransformed real conductivity
response can be found in Tables 5 and 6. The averages and standard deviations of the
measured real conductivity values can be found in Table 6.
The data set was next examined graphically to confirm the required normality
assumption (Figure 15). The internally studentized residuals were plotted against the
normal percent probability. If the residuals lie in a generally straight line, close to 45
43
-------
degrees, normality can be assumed. If the data displays a pronounced "S" shape, the data
may not meet the normality assumption (Stat-Ease, 2007).
A series of plots of residuals versus predicted value, run number, surfactant treatment,
and surfactant treatment were evaluated for trends that could possibly be related to
experimental or systematic error and could result in exaggerated relationships in the
predictive model (Figures 16-19).
The transformed data set was next examined graphically for any design points with
potential undue influence over the predictive model. The leverage plot (Figure 20)
appears normal, with no runs showing leverage values of concern (none greater than 0.8).
Additionally, a plot of the t-test (Figure 21) shows that all of the experimental run values
fall within 95% confidence intervals.
The last influence plot to be evaluated was the Cook's Distance, a combination of the
t-test and leverage (please see Chapter 3 for full explanation) (Figure 22). The logio(real
conductivity) plot does not indicate any runs with large Cook's D values. The successful
evaluation of both normality and influence plots suggests that any outliers have been
previously removed from the analysis (i.e. runs 13 and 22 as noted above) and further
examination of individual runs is not necessary for the real conductivity response.
The real conductivity responses of the experimental treatments appear to be affected
by surfactant type, surfactant concentration, and water type. The results are plotted in
Figure 23. Overall, tap solutions, regardless of surfactant, had higher real conductivity
responses than the corresponding DI solutions. There is a non-linear relationship
between surfactant treatments and real conductivity while the difference in real
conductivity by water type appears to be smaller and typically constant, although the
44
-------
difference may be declining with concentration. The data points appear to be well-
constrained with small error bars and little overlap among treatments (Figure 24). An
error of ±5.1% should be considered with the real conductivity responses as a result of
system error assumed to relate to slight packing differences and sand to solution ratio.
Please see Appendix A for explanation of this error.
The plot of logio(real conductivity) of tap solutions shows a more moderate increase
in real conductivity than what appears in the DI solutions (Figure 24). This apparent
discordance may be a result of the log transform of the data. If the untransformed data is
plotted (Figure 25), the relationship between surfactant and real conductivity is similar
between water types with tap water showing a higher conductance. The real conductivity
response does not appear to show frequency dependence (Figure 26) over the measured
frequency spectrum. This suggests that polarization is unlikely in this environment, as a
frequency-dependent change in real conductivity would likely be associated with a
corresponding change in imaginary conductivity as a result of polarization of the fluid-
filled medium.
Values at the extreme upper end of the frequency range appear to show a slight drop
in real conductivity. This is consistent through the surfactant treatments and water types
and is most likely a result of instrument noise. Measurements at frequencies above
10kHz are often affected by considerable noise likely due to interference by instrument
wiring (Vanhala, 1997).
45
-------
Imaginary Conductivity
Imaginary conductivity response ranges from 0.0113 to 0.0412 |iS/cm. The maximum
to minimum ratio is 3.65. A ratio greater than 10 may indicate the necessity of a data
transformation (Stat-Ease, 2007). In addition, the Box-Cox plot does not indicate the use
of a power law (Figure 27). The ANOVA for imaginary conductivity is found in Table 7.
The mean of the data is 0.0214 [j,S/cm with a standard deviation of 0.0067 [j,S/cm.
The R2 value is 0.530, but the predicted R2 value, based on the proposed model, is
-0.4086. As stated previously, high quality models produce R2 values approaching 1.0 in
both categories. A negative predicted R2 value as seen here indicates that the mean of the
data set may be a better predictor of imaginary conductivity than the proposed model
(Stat-Ease, 2007). The modeled values of the imaginary conductivity response, along
with the averages and standard deviations of the measured values are found in Table 8.
The data set was next examined graphically to confirm the normality assumption
(Figure 28). The internally studentized residuals were plotted against the normal percent
probability. As stated in the previous section, if the plot lies in a generally straight line,
close to 45 degrees, normality can be assumed. If the data displays a pronounced "S"
shape, the data may not meet the normality assumption (Stat-Ease, 2007). In this case, an
"S" has not been clearly identified in the data, although a higher number of data points
would help clarify the normality assumption.
A series of plots of residuals versus predicted value, run number, surfactant treatment,
and water type were evaluated for trends possibly related to experimental or systematic
error (Figures 29-32) and that could result in exaggerated relationships in the predictive
46
-------
model. Run 19, corresponding to 8% AMA 80-1, tap, falls outside of the confidence
interval in Figure 29 and is discussed further in the following paragraphs.
The data set was next examined graphically for any design points with potential
undue influence over the predictive model. The leverage plot (Figure 33) appears
normal, with no runs showing leverage values of concern. Additionally, a plot of the t-
test (Figure 34) shows that 29 of the 30 experimental run values fall within 95%
confidence intervals. Run#19, corresponding to 8% AMA 80-1, DI, falls below the
confidence interval. It should be carefully evaluated as a potential outlier as done below.
The last influence plot to be evaluated was the Cook's Distance, a combination of the
t-test and leverage (please see Chapter 3 for full explanation) (Figure 35). The plot does
not indicate any runs with problematic Cook's D values. While Run #19 plots higher
than the other runs, it is not sufficiently high to omit from analyses. The generally
accepted threshold for omission is a Cook's D value approaching 1 or greater (Stat-Ease,
2007). Additionally, there is no record of any data collection problems associated with
Run #19 that would warrant removal from analyses. The evaluation of both normality
and influence plots suggests that any outliers (i.e. runs 13 and 22) have been removed
from the analysis and further examination of individual runs is not necessary in this
response.
The imaginary conductivity responses of the experimental treatments do not appear to
be significantly affected by either surfactant or water type (Figure 36). There is overlap
in response over the treatments as a whole, with large error bars attached to every
condition (Figure 37). Furthermore the values are very small suggesting there is little to
no polarization or imaginary conductance.
47
-------
An error of ±4.6% should be considered with the imaginary conductivity responses as
a result of system error assumed to relate to slight packing differences and sand to
solution ratio. Please see Appendix A for explanation of this error.
Within a specific surfactant type and concentration, the controls, 0.025% Steol, and
5% Dowfax have almost complete overlap with no clear difference between the tap and
DI treatments. The 0.5% Dowfax treatments display the most defined gap between the
tap and DI samples. The tap solutions and associated error bars are completely separate
from the associated DI solutions. The lowest 0.5% Dowfax, tap response is 8.88E-
3|iS/cm larger than the largest 0.5% Dowfax, DI response.
The 8%> AMA 80-1 treatments have the opposite relationship with water type. The DI
responses were higher than the tap responses. However, the associated error bars are
much closer, in comparison to the 0.5% Dowfax, tap treatments. In addition, the
measured values are spread apart, with a range of 2.70E-2 |iS/cm in the DI measurements
and a range of 6.73E-3 |iS/cm in the tap measurements. Note also that the larger DI
measurement is well outside of its associated upper error bar.
The overall imaginary conductivity response of the system over a range of
frequencies is a non-linear increase with increasing frequency (Figure 38). This suggests
the imaginary conductivity response is frequency dependent; however the surfactant
treatments do not show a significant deviation from the control. With the exception of
the 8%> AMA 80-1 treatments, imaginary conductivity does not vary systematically from
surfactant to surfactant or between water types at any given frequency. While the data set
varies almost 3 orders of magnitude throughout the frequency range, the experimental
solutions containing surfactant do not vary substantially from the control solutions.
48
-------
The 8% AMA 80-1 treatments showed anomalously high imaginary conductivities at
low frequencies relative to the control. The very low and very high conductivities
showed some scattered data points. Within a more moderate frequency range, from 0.366
to 187.5 Hz (Figure 39), the largest spread between surfactant treatments is 0.034 |iS/cm
at 5.86 Hz between 8% AMA, tap and 8% AMA, DI.
The 0.5% Dowfax, DI solution produced the largest range of imaginary conductivity
between the frequencies 0.366 Hz to 187.5 Hz, while the smallest range of imaginary
conductivity corresponded to 0.025% Steol, DI. Within this frequency range, the lowest
recorded response was 4.86E-3 |iS/cm from the 8% AMA 80-1, tap solution at 5.86 Hz.
The highest response corresponded to the 0.5% Dowfax, DI solution at 187.5 Hz.
pH
pH response ranges from 6.16 to 9.44 with a maximum to minimum ratio of 1.53. A
ratio greater than 10 may indicate the necessity of a transformation (Stat-Ease, 2007).
Because this ratio is not greater than 10, and the Box-Cox plot does not indicate the
necessity of a transform (Figure 40), one was not performed. The ANOVA for the pH
response is found in Table 9. It is recognized that the analysis of the pH values do not
necessarily reflect the analysis of the true activity of the hydrogen ion. Separate analyses
would need to be performed after calculating the activity in order to compare the two
types of data.
The mean of the data set is 8.06 with a standard deviation of 0.74. The R2 value is
0.9897, with a predicted R2 value, based on the proposed model, is 0.9686. High quality
models produce high R2 values in both categories (Stat-Ease, 2007). These two values
49
-------
are sufficiently high and in close agreement, indicating that the model may be a good
predictor of pH. Modeled pH values, as well as the averages and standard deviations of
the measured values, are found in Table 10.
The data set was next examined graphically to confirm the normality assumption
(Figure 41). As in the previous parameters, the internally studentized residuals were
plotted against the normal percent probability. The plot was generally linear, indicating
normality in the data set, although a small subset of the data showed minor variation.
Additionally, a series of plots of residuals versus run number, water type, and surfactant
treatment were evaluated for trends possibly related to experimental or systematic error
(Figures 42-45).
The data set was evaluated graphically for any design points with potential undue
influence over the predictive model. The leverage plot (Figure 46) appears normal, with
no runs showing leverage values of concern. Additionally, a plot of the t-test (Figure 47)
shows that all 30 of the experimental run values fall within the 95% confidence interval.
The last influence plot to be evaluated was the Cook's Distance, a combination of the
t-test and leverage (see Chapter 3 for full explanation). The pH plot does not indicate any
runs with large Cook's D values (Figure 48).
The successful evaluation of both normality and influence plots suggests that any
outliers have been previously removed from the analysis and further examination of
individual runs is not necessary.
The pH responses of the experimental treatments appear to be affected by surfactant
and water type, as well as the interaction of the two factors (Figures 49-51). In the low
concentration treatments, the pH values of the DI solutions are appreciably higher than
50
-------
the tap solutions. The 8% AMA 80-1 treatments are the exceptions, with the tap solution
measuring a higher pH than the DI solution. The pH of each solution is fairly well-
constrained. The largest range within a single treatment is a pH difference of 0.51 in the
5% Dowfax, DI solution.
Within the DI solutions, the control, 0.025% Steol, and 0.5% Dowfax solutions are
similar, with some range overlap. 5% Dowfax shows a slight decrease in pH in
comparison to the lower concentration solutions, and 8% AMA 80-1 shows a sharp
decrease in pH. Within the tap solutions, there is response overlap in all treatments
except 5%> Dowfax. This solution showed the highest average pH response at 8.33 with a
standard deviation of 0.08.
An error of ±0.37%> should be considered with the pH responses as a result of system
error assumed to relate to instrument error. Please see Appendix A for explanation of this
error.
Specific Conductivity
Specific conductivity response ranges from 18.38 to 8963.62 [j,S/cm with a maximum
to minimum ratio of 487.68. A ratio greater than 10 may indicate a positive response to a
transformation (Stat-Ease, 2007). In this case, the Box-Cox plot indicated the potential of
using a square root transform (Figure 52), which was performed. The transform is
referred to as sqrt(specific conductivity) in the associated tables and plots. The ANOVA
for the specific conductivity response is found in Table 11.
The mean of the entire data set is 1568.95[j,S/cm with a standard deviation of 3052.58
[j,S/cm. The large standard deviation, relative to the mean value is most likely due to the
51
-------
large range of values in responses. The R2 value is 0.9999, with a predicted R2 value,
based on the proposed model, is 0.9998. As discussed previously, high quality models
produce R2 values approaching 1.0 in both categories. These two values are sufficiently
high and in close agreement, indicating that the model may be a good predictor of
specific conductivity. The modeled values for sqrt(specific conductivity) are found in
Table 12. The untransformed modeled values, as well as the averages and standard
deviations of the measured specific conductivity values are located in Table 13.
The data set was next examined graphically to confirm the normality assumption
(Figure 53). The internally studentized residuals were plotted against the normal percent
probability, as discussed in the previous sections. A series of plots of residuals versus
predicted, run number, surfactant treatment, and water type were evaluated for trends
possibly related to experimental or systematic error (Figures 54-57). A wider range of
residuals appears in low predicted values, DI solutions, and the control groups.
The transformed data set was evaluated graphically for any design points with
potential undue influence over the predictive model. The leverage plot (Figure 58)
appears normal, with no runs showing leverage values of concern. Additionally, a plot of
the t-test (Figure 59) shows that all of the experimental run values fall within 95%
confidence intervals. The last influence plot to be evaluated was the Cook's Distance
(please see Chapter 3 for full explanation) (Figure 60). The sqrt(specific conductivity)
plot does not indicate any runs with large Cook's D values (greater than 0.95).
The evaluation of both normality and influence plots suggests that any outliers have
been removed from the analysis previously and further examination of individual runs is
not necessary for the specific conductivity response.
52
-------
The specific conductivity responses of the experimental treatments appear to be
affected by both surfactant and water type (Figure 61). There is a non-linear positive
correlation between surfactant, concentration, water type and specific conductivity
(Figure 62).
Overall, tap solutions had higher specific conductivity than the corresponding DI
solutions. The plot of square root of specific conductivity shows similar increases in both
tap and DI solutions as surfactant concentration increases (Figure 62). A plot of the
untransformed data show similar trends (Figure 63, 64).
Each treatment is well-constrained with small error bars and very little overlap.
Specific conductivity appears to be more strongly influenced by surfactant treatment and
concentration than by water type. The difference in specific conductivity between the tap
and DI solutions of a specific surfactant treatment averages 907 [j,S/cm, with a standard
deviation of 180. The difference appears to decline with increasing surfactant
concentration. While this is far from a constant difference, it is smaller than the averaged
and standard deviations of the surfactant treatments relative to each other. The DI
solutions measured on average 2208 [j,S/cm with a standard deviation of 3565. The tap
solutions measured on average 3081 [j.S/cm with a standard deviation of 3387.
An error of ±10.1% should be considered with the specific conductivity responses as
a result of systematic error assumed to relate to instrument error and slight variations in
solution temperature. Please see Appendix A for explanation of this error.
53
-------
Dissolved Oxygen
Dissolved oxygen response ranged from 4375.5 to 8644 |ig/L with a maximum to
minimum ratio of 1.98. A ratio greater than 10 may indicate the necessity of a transform,
but a power transform generally has little to no effect on ratios less than 3 (Stat-Ease,
2007). As the dissolved oxygen ratio is low, no transform was performed. In addition,
the Box-Cox plot does not recommend the use of a power law (Figure 65). The ANOVA
for the DO response is found in Table 14.
The mean of the data set is 7394.88 |ig/L with a standard deviation of 832.46. The R2
value is 0.6710, with a predicted R2 value, based on the proposed model, of 0.0432. High
quality models produce high R2 values in both categories. The values for dissolved
oxygen are neither maximized or in close agreement. The proposed model is unlikely to
be a good predictor for dissolved oxygen. Modeled values are found in Table 15, along
with the averages and standard deviations of the measured DO values.
The data set was examined graphically to confirm the normality assumption (Figure
66). The internally studentized residuals were plotted against the normal percent
probability, and an "S" shape was interpreted. Because of this, the assumption of
normality may not be met in this data set, although the same analyses were completed for
dissolved oxygen as for the other parameters.
A series of plots of residuals versus predicted values, run number, surfactant
treatment, and water type were evaluated for trends possibly related to experimental or
systematic error that could result in poor relationship prediction by the model (Figures
67-70). The 0.025% Steol, tap treatment appeared to show the most range in of residuals
throughout the plots. The transformed data set was evaluated graphically for any design
54
-------
points with potential undue influence over the predictive model. The leverage plot
(Figure 71) appears normal, with no runs showing leverage values of concern.
Additionally, a plot of the t-test (Figure 72) shows that 29 of the 30 experimental run
values fall within 95% confidence intervals. Run #9, 0.025% Steol, tap, lies well outside,
while Run 15, 0.025% Steol, tap solution, lies just inside the lower boundary.
The last influence plot to be evaluated was the Cook's Distance, a combination of the
t-test and leverage (please see Chapter 3 for full explanation) (Figure 73). The dissolved
oxygen plot does not indicate any runs with very large Cook's D values. While Run 9
and 15 are higher than most, it is not sufficient to omit the data points from the analysis;
Nothing unusual was noted during the experimental runtime, including temperature
fluctuation or substantial amounts of air entering the flow cell. In addition, measurements
of other parameters, taken simultaneously using the same equipment, do not reflect the
same outlier potential. The successful evaluation of both normality and influence plots
suggests that any outliers have been previously removed from the analysis and further
examination of individual runs is not necessary.
The dissolved oxygen responses do not appear to be significantly affected by either
surfactant or water type (Figure 74, 75). With the exception of the tap water control,
there is overlap in the measured responses across all experimental treatments. The lowest
measured tap control responses are 222 [j,g/L higher than any other measured responses.
The largest range is in the 0.025% Steol, tap treatment, of 3274 (J,g/L. The other
treatments average a range almost one magnitude smaller, at 330.6 [j,g/L with a standard
deviation of 157.9 (Figure 75).
55
-------
With the exception of 0.025% Steol, tap treatment, there is overlap across the
modeled responses. The 0.025% Steol, tap treatment is modeled much lower than the
other treatments. Despite the very large range in its measured responses, the error bar
attached to the 0.025% Steol, tap treatment is the same size as the other treatments.
In the associated bar graph (Figure 76), there is overlap in response over the
treatments as a whole, which is consistent with the previous plot. Furthermore, the error
bar associated with 0.025% Steol, tap in this plot is substantially larger than the error bars
associated with the other experimental treatments. The large error bar here is most likely
due to Run 9, as discussed above. While most of the treatments appear to be well-
constrained, there is overlap in all treatments except the tap control.
An error of ±4.3% should be considered with the DO responses as a result of system
error assumed to relate to instrument error and slight temperature variation. Please see
Appendix A for explanation of this error.
Dielectric Constant
Dielectric constant response ranges from 21.03 to 25.07, with a maximum to
minimum ratio of 1.19. Ratios greater than 10 may indicate the necessity of a data
transform. Because this ratio is low and the Box-Cox plot does not suggest a transform
(Figure 77), one was not performed on the data set. The ANOVA for the specific
conductivity response is found in Table 16.
The mean of the data set is 23.5 with a standard deviation of 1.10. The R2 value is
0.4618, with a predicted R2 value, based on the proposed model, of -0.5449. High quality
models produce high R2 values in both categories. A negative predicted R2 value as seen
56
-------
here indicates that the mean of the data set may be a better predictor of dielectric constant
than the proposed model. Modeled values are found in Table 17, along with the averages
and standard deviations of the measured dielectric constants.
As with the previous parameters, the data set was examined graphically to confirm
the normality assumption with a plot of the internally studentized residuals plotted
against the normal percent probability (Figure 78). Identification of an "S" shape in the
plot may suggest that the data set does not meet the normality requirements (Stat-Ease,
2007).
A series of plots of residuals versus predicted value, run number, surfactant treatment,
and water type were evaluated for trends possibly related to experimental or systematic
error that could result in significant errors in predictive models (Figures 79-82).
The data set was evaluated graphically for any design points with potential undue
influence over the predictive model. The leverage plot (Figure 83) appears normal, with
no runs showing leverage values greater than 0.8. Additionally, a plot of the t-test
(Figure 84) shows that all the experimental run values fall within 95% confidence
intervals.
The last influence plot to be evaluated was the Cook's Distance, a combination of the
t-test and leverage (please see Chapter 3 for full explanation) (Figure 85). The dielectric
constant plot does not indicate any runs with large Cook's D values.
The evaluation of both normality and influence plots suggests that any outliers have
been removed from the program previously and further examination of individual runs is
not necessary for the dielectric conductivity response.
57
-------
The dielectric constant responses of the experimental treatments do not appear to be
statistically significant effects as a results of either surfactant or water type (Figure 86).
All dielectric responses overlap among the treatments. None of the treatments are well-
constrained, with large error bars in both measured and modeled data (Figure 87).
In the associated bar graph (Figure 88), there is overlap in response over the
treatments as a whole, consistent with the previous plots (Figure 86, 87). Again, none of
the treatments are well-constrained. There do not appear to be any trends in dielectric
constant based on water type, surfactant type, or concentration.
An error of ±3.6% should be considered with the dielectric constant responses as a
result of system error assumed to relate to instrument error, slight packing differences,
and the sand to solution ratio. Please see Appendix A for explanation of this error.
Results Summary
Table 18 outlines the overall responses of each experimental treatment. Real and
specific conductivities, along with pH, produced models with the highest R2 and R2pred
values, suggesting that these three parameters have the most predictable response in a
quartz sand-water environment of the parameters measured. Imaginary conductivity and
dissolved oxygen have negative R2pred values, indicating that those parameters may be
better predicted by the mean of the data set than by the proposed models.
58
-------
CHAPTER 5
DISCUSSION
In the previous chapter, the measured responses were described, along with the
analyses used to determine the robustness of data. This chapter will discuss the results in
more detail and attempt to place the parameter responses in a broader context while
proposing possible explanations for each parameter's behavior.
Geoelectrical Measurements
There is a clear correlation between surfactant and real conductivity, and to a lesser
extent, water type and real conductivity. Real conductivity response appears as a
logarithmic increase with increasing surfactant concentration, which suggests that
surfactant concentration is the key to real conductivity response.
After the conclusion of the initial 30-run experiment plan, real and imaginary
conductivity values were measured for 8% DI solutions of Steol CS-330 and Dowfax
8390, in addition to the original 8% Aerosol MA 80-1 experiments in order to determine
if the conductivity could be linked to either the surfactant's chemical make-up or the
actual solution concentration. These experiments were performed identically to the SIP
experiments described in Chapter 3. If concentration is the controlling factor of real
conductivity response, all three of the 8% solutions should show similar real conductivity
measurements. The real conductivity measurements at 8% were compared to the real
conductivity value of a DI control. Because the Steol and Dowfax 8% runs took place
after the conclusion of the experimental runs, they were compared to a different DI
control than the original 8% AMA runs (Table 19). Comparisons were made by
59
-------
calculating a percent difference (Eqn 16) between the DI control and the experimental
solution.
PercentDifference = —— — * 100 (16)
°'di
Percent Difference is the calculated percent difference between the DI control and the
experimental solution, &DI is the real conductivity value of the measured DI control, and
cr'exp is the real conductivity of the experimental surfactant treatment.
The 8% Steol solution averaged a real conductivity value of 497.5 ±14.1 [j,S/cm,
which corresponds to a 1234% increase from the DI control value. The 8% Dowfax
treatment averaged a real conductivity value of 714.2 ±8.8 |iS/cm, which corresponds to
a 1816 % increase from the DI baseline. These two solutions are somewhat comparable,
suggesting that real conductivity is affected more by surfactant concentration than
surfactant brand or molecular make-up. However, there is still a difference of 582%
between the two surfactants that is most likely explained by differences in chemical
composition and structure. The 8% AMA 80-1 solution averaged 1793.6 ±58.35 |iS/cm,
corresponding to a 13945%) increase from the DI control. This is significantly larger than
the values of the other two surfactants, thus there is likely a different cause for the
conductivity measurement than concentration alone.
Additionally, the specific conductivity of a 0.5% AMA 80-1, DI solution was
measured in an effort to clarify the role of concentration on the geoelectrical responses.
The responses were compared to the 0.5% Dowfax, DI solution response recorded during
the initial experiments. The 0.5% Dowfax solution averaged a specific conductivity
value of 123.75 ±4.08 |iS/cm, which corresponds to a 1337%) increase from the DI
60
-------
control. The 0.5% AMA 80-1 solution, however, averaged 199.02 ±0,1 1 |iS/cm,
corresponding to a 9506% increase from the DI control. The substantially larger percent
increase of the 0.5% AMA 80-1 solution compared to the Dowfax solution is in close
agreement with the findings from the 8% solution experiments described above.
Concentration cannot account for the entire difference in the conductivity responses
among surfactant treatments.
The three surfactant formulas, Steol CS-330, Dowfax 8390, and Aerosol MA 80-1 are
all anionic, suggesting that this property cannot be cited as a reason for differences in
conductivity. The molecular formula of Steol, primarily sodium laureth sulfate (Stepan
Company, 2005) , is C^Ife^FLiO^C^S" (Karapanagioti et al., 2005), and its structure
can be found in Figure 89. The molecular formula of Dowfax 8390, or alkyldiphenyl
oxide disulfonate (Dow Chemical Company, 2009), is C28H40O7S22" (Karapanagioti et
al., 2005), and it's structure can be viewed in Figure 90. AMA 80-1, or dihexyl
sulfosuccinate, has a molecular formula of Cu^yCbNaS (Cytec Industries, 1994). This
particular surfactant contains the alcohol isopropanol (isopropyl alcohol), and its
structure is found in Figure 91.
It seems apparent that some surfactant formulations act as stronger electrolytes than
others when in solution. There are several potential mechanisms to explain this effect.
They include differences in dissociation constant, as well as differences in the number of
dissociable ions on an individual molecule. The use of cosolvents and other additives
could also affect differences in conductivities among surfactant formulations.
It is possible that the increase in pore fluid conductivity is related to the number of
easily dissociable ions in a surfactant molecule. The dissociated ions would increase the
61
-------
real and specific conductivity by increasing the total dissolved solids and salts in the
electrolyte, and thus the number of dissociating ions would control the degree of
conductivity increase. If one assumes that the number of sodium ions corresponds to the
number of dissociable ions on a surfactant molecule, Dowfax has two, while Aerosol MA
80-1 and Steol have only one. Therefore it is unlikely that the number of sodium ions in a
surfactant is the controlling factor on conductivity.
A somewhat related possibility is tied to the dissociation constant of each surfactant.
A dissociation constant describes a compounds ability to break apart into smaller
components. Due to complex chemical properties, some compounds dissociate more
easily in polar solvents than others. While Aerosol MA 80-1 may have fewer dissociable
ions, its dissociation constant may be higher, resulting in more complete dissociation
when in solution. This could result in higher electrolytic conductivity. Additionally,
Aerosol MA 80-1 includes isopropanol, an alcohol, in its formulation, which may affect
its geoelectrical properties. However, research has found that alcohols have low
conductivity relative to the specific conductivity responses of the surfactant solutions
(Prego et al., 2000). This suggests that the isopropanol in the Aerosol MA 80-1 solutions
does not account for the significantly higher conductivity responses in comparison to the
surfactant solutions that do not contain the alcohol.
Water type appears to affect real conductivity as a semi-constant, with an average of
175.8 |iS/cm (standard deviation 44.7 |iS/cm) separating the tap and DI measurements of
a specific surfactant treatment. The semi-constant gap appears to increase with
increasing surfactant concentration. While the standard deviation initially seems large, in
comparison to the difference in response between the surfactant treatments it is not.
62
-------
This semi-constant separation is likely a result of the differences in starting
conductivity of the two water types, rather than an effect of interactions with either
surfactant or aquifer solids. The starting specific conductivity of the DI water was
measured at 3.8 |iS/cm, while the tap water measured 1025 |iS/cm, which is a substantial
difference. This hypothesis, that intial differences in water type are important, is
supported by the specific conductivity response, which also indicates a semi-constant gap
between the correlated tap and DI measurements.
In addition, the semi-constant state of the difference between the DI and tap solutions
is likely a result of the size of the role played by the water type in real conductivity.
When the surfactant concentration is low, the measured conductivity is due to the water
type. As the surfactant concentration increases, it is likely the result of a change in the
dominant conductivity source at different concentrations of surfactant. At low
concentrations, the conductivity of the water dominates the response, while at high
concentrations, the conductivity of the surfactant is dominant.
Imaginary conductivity does not appear to have the same relationships with either
surfactant treatment or water type as seen in real and specific conductivities. The
difference between the control and surfactant treatments is minimal, with little to no
significant correlation. This finding implies that the main electrical conduction
mechanism in the tested environment is electrolytic, and there is little to no surface
conduction that results from the presence of surfactants in a saturated quartzitic sand
environment. Because specific conductivity is a single component of real conductivity, it
is expected that any change in conduction mechanism, via polarization or increased
surface conductivity, would alter the real conductivity response, while the specific
63
-------
conductivity remained unaffected. Because the measured real and specific conductivity
responses in this thesis follow the same trend throughout the surfactant treatments, it
suggests that changes in the real conductivity are related to changes in the specific
conductivity. This leads to the conclusion that electrolytic conductivity is the main
conduction mechanism.
While real conductivity shows no significant value change across the measured
frequency range, imaginary conductivity values display an overall increase with
increasing frequency. This frequency effect does not appear to be dependent upon
surfactant treatment or water type, however. Both controls and all surfactant treatments
display the same general trend through the frequency spectrum. In addition, all
treatments lie within one order of magnitude from one another and have fairly large
standard deviations, indicating there is overlap among treatments.
There is a clear correlation of surfactant treatment to real conductivity (Table 6) and
specific conductivity (Table 13) while there is no clear correlation of imaginary
conductivity (Table 8) to the presence of surfactant in a quartz sand-water environment.
The behaviors of the real, specific, and imaginary conductivities in the presence of
surfactant in quartz sand-water environment suggests that the geoelectrical conduction
mechanism is primarily electrolytic and a function of pore fluid chemistry in this
particular set of conditions. This supports the findings of Werkema (2008) that there is
an increase in pore fluid conductivity in relation to surfactant presence (Figure 92).
64
-------
Water Quality Measurements
The water quality measurements of pH, specific conductivity, and dissolved oxygen
(DO) display varying responses to surfactant presence in a quartz sand-water
environment. The pH and specific conductivity parameters appear to repond in a
statistically significant manner to both surfactant treatment and water type. Dissolved
oxygen does not show the same correlation, and statistical analysis did not meet the
required normality assumption. Because of this, statistical significance could not be
assessed in this parameter and will therefore not be discussed in depth.
pH
The pH is affected by the water type, as well as the interaction between water type
and surfactant treatment (Table 21). The response can be broken into two parts: the
response of tap solutions versus DI solutions and the response of the high concentration
solutions (5%, 8%) versus the low concentration solutions (control, 0.025%, 0.5%).
With the exception of the 8% AMA 80-1 solution, the DI solutions measured higher
pH values than the tap solutions. This is likely a result of the pH of the water that was
mixed with the surfactant, rather than a comment on the surfactants themselves. This is
concluded due to the higher pH values measured in the DI control in relative to the tap
water control. The gap between the tap and DI solutions at low surfactant concentrations
averages 1.57 with a standard deviation of 0.25. The larger concentration solutions do
not appear to have the same relationship with each other. While pH of the tap solutions
appear to be moderated or perhaps buffered by the tap water itself, the high concentration
DI solutions, particularly 8% AMA 80-1, show a precipitous decrease in pH. The reason
for this response may be related to the molecular structure of the surfactant, including the
65
-------
presence of the isopropyl alcohol. Some ions may be more readily dissociated in DI
water than in tap water, driving a decrease in pH. It may be reasonable to hypothesize
that solutions of Steol and Dowfax at 8% concentration may measure similar decreased
pH values due to similar surfactant structures. Alternately, if dissociation constants are
more of a controlling factor, then one may expect variation in pH values among the
different surfactant formulations. As noted in the Geoelectrical Measurements section,
AMA 80-1 contains isopropanol, which may also explain the decrease in pH associated
with that surfactant.
Specific Conductivity
As discussed in the Geoelectrical Measurements section earlier in this chapter,
specific conductivity response suggests a statistically significant link exists between
surfactant treatment and specific conductivity, as well as water type and specific
conductivity.
The general relationships seem to mirror that of real conductivity, indicating that
ionic conductivity is likely the primary conduction mechanism in the saturated sand
analog aquifer. The DI solutions display overall lower specific conductivity
measurements than the associated tap solutions. The difference between the two is a
semi-constant averaging 873 [j,S/cm with a standard deviation of 184. Low concentration
treatments have substantially lower specific conductivity values than the higher
concentration treatments. Potential mechanisms for this behavior are discussed above in
the Geoelectrical Measurements section.
A simple comparison of the findings of this research to the findings of Werkema
(2008) yields the plot in Figure 85. This comparison clearly shows that the trends of
66
-------
specific conductivity organized by surfactant type and concentration are similar in the
two works. The slight differences between the two may be related to the starting specific
conductivity value of the DI water used in solution. There is also the potential that the
sand matrix used in this research contributed dissolved solids and ions to the electrolyte,
increasing the specific conductivities in the higher concentration solutions. Investigating
what effects the matrix has on the geoelectrical response was a main goal of the thesis
research.
Dissolved Oxygen
Dissolved oxygen is a difficult parameter to measure and can be affected by
temperature, flow velocity, or outside air leaking into the system. The raw data suggests
that dissolved oxygen response does not change significantly with surfactant presence.
Although the 0.025% Steol, tap solutions appear to show a significantly lower response
than the other surfactant treatments as well as its associated DI solution, further testing
could not replicate the low numbers. In addition, statistical analysis suggests that the
normality assumption is not valid, leaving any proposed model in an uncertain state.
Dielectric Constant
Dielectric constant response to surfactants in a quartz sand-water environment shows
little response. One suggestion to explain this is insufficient instrument sensitivity. The
sensitivity for the MiniTrase is ±2% moisture content, equivalent to a dielectric constant
of 3.27, using a standard waveguide (Soilmoisture Equipment Corporation, 2005). It
should be noted that the experiments performed in this project used a shorter waveguide
than is standard. A standard waveguide is 15 cm. The experiments in this research
67
-------
utilized an 8 cm waveguide. It is expected that this would increase error. The expected
error overshadows the relatively small response differences among surfactant treatments
in the time domain reflectometry experiments. Additionally, the surfactant molecules
may be too large to "twist," a behavior necessary to the relaxation phenomenon on which
time domain reflectometry response depends (Endres, A.L., personal communication,
December 2008).
In addition, the typical dielectric constant values for saturated sand are between 20
and 30 (Kirsch, 2006). All of the values measured in this project fall in that category,
suggesting that dielectric constant may be more strongly impacted by the matrix materials
and moisture content than the solution itself in the experimental quartzitic sand
environment.
While dielectric response of surfactant-quartz sand-water may not be significant, this
does not rule out the potential of GPR as a surfactant monitoring method, as the method
is also affected by the conductivity of the surveyed area. As demonstrated in this project,
the conductivity measurements are affected by the surfactant solutions.
68
-------
CHAPTER 6
CONCLUDING REMARKS
The research presented in this thesis has provided further information to the scientific
community, as well as indicating potential directions for future work. Increased real and
specific conductivities associated with surfactants used in SEAR support the work of
Werkema (2008), and indicate that the geoelectrical responses in quartz sand-water
environments may be useful in monitoring subsurface surfactants with geophysical
methods. Resistivity surveys in particular show potential as the real and specific
conductivities can show a strong response to surfactant treatments.
The positive correlation between real and specific conductivities suggests that
electrolytic conduction is the primary electrical conduction mechanism in quartz sand-
water environments. A lack of significant imaginary conductivity response supports this
suggestion, and rules out substantial conduction via surface or electronic conduction.
The pH response also appears to be affected by surfactant presence in a quartz sand-
water environment. High surfactant concentrations appear to decrease the pH value of
the environment. If this response is scaled to a field environment, there is a potential to
negatively affect subsurface organisms, including bacteria and microbes that are actively
aiding in bioremediation.
The dissolved oxygen and dielectric constant parameters do not appear to be
significantly affected by the presence of surfactants in a quartz sand-water environment.
The possible reasons for this are discussed in Chapter 5.
Future work should continue to increase both the scale of experimentation, as well as
the complexity. Studies will need to incorporate heterogeneous solid materials, including
69
-------
clays. The presence of clay in the subsurface is likely to increase the imaginary
conductivity response, as well as providing sorption sites for surfactant. The surfactants
may also interact with clay particles, resulting in changes to the geoelectrical responses.
The introduction of clays to the experimental environment will likely increase the
imaginary conductivity component of the geoelectrical response, possibly masking any
surfactant-related response.
Increasing the complexity of the experimental electrolyte will also be important for
future work. SEAR is only used in environments containing contaminants. It will be
important to include potential contaminants, like tetrachloroethylene, in the saturating
solutions in order to observe any interactions between surfactants and DNAPL.
Chlorinated solvents such as tetrachloroethylene (PCE) and trichloroethylene (TCE) have
been associated with decreased conductivity in field studies (Chambers et al., 2004) and
would therefore be expected to buffer the increased conductivity responses shown with
surfactant presence. Additionally, DNAPL contaminants have low dielectric constants,
generally below 10 (Ajo-Franklin et al., 2006). The interactions of DNAPL and
surfactant may result in measurable changes to dielectric constant.
The geoelectrical response should also be investigated for changes related to temporal
variations. The experiments presented in this research concentrate on readings made
within an hour of saturation. Field applications require a substantially larger time scale,
stretching beyond a full year and up to several years. Dissolved oxygen, while not
responding to the surfactants over the time scale used in this research, may react
differently over a longer period of time.
70
-------
This research also utilized a fully-saturated environment. However, responses to
surfactants within a more complex saturation profile should also be investigated in an
effort to bring the complexity of the environment closer to the scale of a field application.
It is expected that decreasing saturation will lower both the conductivity and dielectric
constant of an environment.
In conclusion, geophysical, and particularly geoelectrical, methods have the potential
to monitor surfactants in the subsurface. A substantial amount of future work must
increase the scale and complexity of the experimental conditions in order to determine
the true feasibility. Ultimately, the ability to monitor surfactants in the subsurface could
result in more efficient and effective groundwater remediation, which will be beneficial
to all living organisms.
71
-------
EXHIBITS
TABLES AND FIGURES
Table 1. Experimental Treatments.
Surfactant
Concentration
Water Type
None
-
DI
None
-
Tap
Steol CS-330
0.025%
DI
Steol CS-331
0.025%
Tap
Dowfax 8390
0.5%
DI
Dowfax 8390
0.5%
Tap
Dowfax 8390
5%
DI
Dowfax 8390
5%
Tap
Aerosol MA 80-1
8%
DI
Aerosol MA 80-1
8%
Tap
Table 2. Specific Conductivity results from Werkema 2008.
Specific Conductivity
(jiS/cm)
Control
3.93
Steol CS-330
16.35
0.5% Dowfax 8390
405.74
5% Dowfax 8390
2465.86
8% Aerosol MA 80-1
7232.65
72
-------
Table 3. Run order and response of each experimental treatment at 11.7Hz. Runs 13 and 22 are lined out as they were not included
in analyses. Real, imaginary, and specific conductivity are reported in [j,S/cm. Units for dissolved oxygen (DO) are |ig/L.
-j
Run
Block
Sui'fa ct ant
Water
Real
Imaginary
Sp Cond
DO
Dielectric
pH
1
1
0.5 Dowfax
DI
123.07
1.71E-02
385.4
7420.5
22.3
9.44
2
1
5 Dowfax
Tap
925.945
3.61E-02
3048.44
7859
22.73
8.37
3
1
5 Dowfax
DI
733.025
2.82E-02
2223
7453
22.2
9.12
4
1
None
Tap
324.395
2.56E-02
1019.95
8507
23.86
7.65
5
1
.025 Steol
DI
23.57
3.38E-02
32.92
7383
23.1
9.39
6
1
8 AMA 80-1
DI
2630.775
5.15E-02
8352.18
7406
21.03
6.36
7
1
8 AMA 80-1
Tap
2865.44
2.15E-02
8963.62
7313
21.5
7.76
8
1
0.5 Dowfax
Tap
409.755
4.15E-02
1254.36
7817.5
23.13
7.85
9
1
.025 Steol
Tap
346.285
2.87E-02
1053.65
7649.5
23.83
7.71
10
1
None
DI
19.285
2.53E-02
24.48
7158.5
23.96
9.34
11
2
None
DI
18.275
3.48E-02
18.38
7676
23.33
9.095
12
2
8 AMA 80-1
Tap
2895.2
3.17E-02
8924.695
7565
22.26
7.8
+3-
—
5 Dowfax
Tap
1033.475
-6.00E-02
3136.585
7616.5
14
2
.025 Steol
DI
22.4
2.28E-02
35.665
7683
24.37
9.17
15 2
2
.025 Steol
Tap
353.21
3.33E-02
1069.095
4375.5
21.97
7.32
16
2
0.5 Dowfax
DI
129.985
2.80E-02
364.48
7141.5
24.17
9.14
17
2
5 Dowfax
DI
707.755
3.69E-02
2250.5
7293.5
23.3
8.78
18
2
None
Tap
333.77
2.30E-02
1048.77
8081
22.5
7.625
19
2
8 AMA 80-1
DI
2649.57
2.17E-02
8369.16
7470.5
23.13
6.33
20
2
0.5 Dowfax
Tap
425.545
4.43E-02
1254.86
7555.5
24.17
8.025
21
3
None
DI
19.92
3.67E-02
34.6
7560.5
21.9
9.26
2S
8 AMA 80-1
Tap
2786.67
-1.20E-01
9038.19
7567.5
2^43-
TtSS
3
-------
23
3
.025 Steol
Tap
354.245
2.79E-02
1079.77
5179.5
24.23
7.37
24
3
.025 Steol
DI
21.835
3.51E-02
32.46
7442.5
25.07
9.28
25
3
None
Tap
332.89
3.19E-02
1076.01
8644
21.57
7.6
26
3
8 AMA 80-1
DI
2790.92
6.25E-02
8371.99
7348
22.13
6.16
27
3
0.5 Dowfax
DI
119.55
2.39E-02
363.89
7190
23.9
9.19
28
3
5 Dowfax
Tap
982.845
3.38E-02
3129.63
7503
22.57
8.38
29
3
5 Dowfax
DI
763.405
2.32E-02
2272.47
7408.5
23.27
8.61
30
3
0.5 Dowfax
Tap
399.545
4.77E-02
1279.93
7425
24.67
8.07
^1
-------
Table 4. ANOVA of Real Conductivity; Classical Sum of Squares, Type II. Units are
[j,S/cm.
Sum of Squares
F-factor
p-value
Surfactant
10.23
11222.87
<0001
Water Type
3.07
13482.86
<0001
Model
14.14
6898.72
<0001
Mean
561.93
C.V.
0.66
Standard Deviation
672.50
R2
0.9997
Maximum
1910.83
R adj
0.9996
Minimum
12.06
R2
av ored
0.9992
Table 5. Modeled values of logio (real conductivity) values
DI
Tap
Control
1.11
2.34
0.025% Steol
1.18
2.37
0.5% Dowfax
1.92
2.44
5% Dowfax
2.69
2.80
8% AMA 80-1
3.25
3.28
75
-------
Table 6. Modeled and average measured real conductivity values for DI and tap
solutions. Modeled values do not include standard deviations. Units are [j,S/cm.
DI
Tap
Modeled
Mean
Std
Mean
Std
±5.1%
Dev
±5.1%
Dev
Control
12.88
-
218.78
-
0.025% Steol
15.14
-
234.42
-
0.5% Dowfax
83.18
-
275.42
-
5% Dowfax
489.78
-
630.96
-
8% AMA 80-1
1778.3
-
1905.5
-
Measured
Control
12.77
0.55
220.23
3.45
0.025% Steol
15.07
0.59
234.17
2.89
0.5% Dowfax
82.80
3.54
274.41
8.73
5% Dowfax
489.82
18.58
653.84
35.86
8% AMA 80-1
1793.6
58.36
1899.4
37.39
Table 7. ANOVA of Imaginary Conductivity; Classical Sum of Squares, Type II.
Units are in |iS/m
Sum of Squares
F-factor
p-value
Surfactant
1.241E-4
0.91
0.4796
Water Type
4.855E-6
0.14
0.7102
Model
6.115E-4
2.00
0.1081
Mean
0.0214
c.v.
26.97%
Standard Deviation
0.0067
R2
0.5297
Maximum
0.0113
R adj
0.2652
Minimum
0.0412
av ored
-0.4086
76
-------
Table 8. Modeled and average measured imaginary conductivity values for DI and tap
solutions. Standard deviations are included for measured values. The predictive model
did not produce standard deviations for the modeled values. Units are in [j,S/cm.
DI
Tap
Modeled
Mean
±4.6%
Std Dev
Mean
±4.6%
Std Dev
Control
0.02148
-
0.01786
-
0.025% Steol
0.02035
-
0.01997
-
0.5% Dowfax
0.01533
-
0.02965
-
5% Dowfax
0.01960
-
0.02279
-
8% AMA 80-1
0.03004
-
0.01868
-
Measured
Control
0.0215
0.0041
0.0179
0.0031
0.025% Steol
0.0204
0.0045
0.0200
0.0019
0.5% Dowfax
0.0153
0.0037
0.0296
0.0021
5% Dowfax
0.0196
0.0046
0.0241
-
8% AMA 80-1
0.0300
0.0140
0.0177
0.0048
Table 9. ANOVA of pH; Classical Sum of Squares, Type II. pH units are used.
Sum of Squares
F-factor
p-value
Surfactant
12.21
185.04
<0001
Water Type
5.01
303.69
<0001
Model
25.43
171.34
<0001
Mean
8.25
C.V.
1.56%
Standard Deviation
0.13
R2
0.9897
Maximum
9.44
R adj
0.9840
Minimum
6.16
n2
av ored
0.9686
77
-------
Table 10. Modeled and average measured pH values for DI and tap solutions.
Measured values also include standard deviations. The predictive model did not produce
standard deviations for the modeled values. pH units are used.
DI
Tap
Modeled
Mean
±0.37%
Std Dev
Mean
±0.37%
Std Dev
Control
9.23
-
7.63
-
0.025% Steol
9.28
-
7.47
-
0.5% Dowfax
9.26
-
7.98
-
5% Dowfax
8.84
-
8.35
-
8% AMA 80-1
6.28
-
7.76
-
Measured
Control
9.34
0.01
7.63
0.03
0.025% Steol
9.39
0.03
7.47
0.18
0.5% Dowfax
9.44
0.02
7.99
0.11
5% Dowfax
9.13
0.01
8.33
0.07
8% AMA 80-1
6.54
0.3
7.81
0.06
Table 11. ANOVA of Specific Conductivity; Classical Sum of Squares, Type II. Data
is presented in [j,S/cm.
Sum of Squares
F-factor
p-value
Surfactant
21568.38
48626.39
<.0001
Water Type
2065.97
18631.12
<.0001
Model
23256.34
23303.08
<.0001
Mean
1568.95
C.V.
0.84%
Standard Deviation
3052.58
R2
0.9999
Maximum
8963.62
R adj
0.9999
Minimum
18.38
1}2
av pred
0.9998
78
-------
Table 12. Modeled values of Sqrt (Specific Conductivity).
DI
Tap
Control
5.039
32
38
0.025% Steol
5.802
32
67
0.5% Dowfax
19.27
35
54
5% Dowfax
47.42
55
52
8% AMA 80-1
91.46
94
69
Table 13. Modeled and average measured specific conductivity values. Standard
deviations are listed with the associated measurement. Units are [j,S/cm.
DI
Tap
Modeled
Mean
±10.1%
Std Dev
Mean
±10.1%
Std Dev
Control
25.39
-
1048.14
-
0.025% Steol
33.66
-
1067.46
-
0.5% Dowfax
371.33
-
1263.09
-
5% Dowfax
2248.66
-
3082.47
-
8% AMA 80-1
8364.93
-
8966.2
-
Measured
Control
23.87
6.75
1048.18
24.08
0.025% Steol
32.79
2.06
1066.37
11.36
0.5% Dowfax
367.99
10.46
1262
13.35
5% Dowfax
2248.8
21.5
3102.63
42.32
8% AMA 80-1
8363.52
8.64
8971.09
49.17
79
-------
Table 14. ANOVA of Dissolved Oxygen; Classical Sum of Squares, Type II. Data is
presented in (J,g/L.
Sum of Squares
F-factor
p-value
Surfactant
5.39E6
3.77
0.0242
Water Type
29854.07
0.083
0.7763
Model
1.17E7
3.63
0.0121
Mean
7375.34
C.V.
8.11%
Standard Deviation
597.95
R2
0.6710
Maximum
8644
R adj
0.4859
Minimum
4375.5
n2
av pred
0.0432
Table 15.
solutions,
in (J,g/L.
DI
Tap
Modeled
Mean
±4.3%
Std
Dev
Mean
±4.3%
Std Dev
Control
7465
-
8410.7
-
0.025% Steol
7502.8
-
5734.8
-
0.5% Dowfax
7250.7
-
7599.3
-
5% Dowfax
7385
-
7607.8
-
8% AMA 80-1
7408.2
-
7402.2
-
Measured
Control
7468.22
228.49
8430.22
257.02
0.025% Steol
7495.67
116.92
5708.56
1517.22
0.5% Dowfax
7290.22
135.5
7633.78
170.63
5% Dowfax
7401.13
78.52
7692.33
161.4
8% AMA 80-1
7442.33
80.18
7508
129.04
Modeled and average measured dissolved oxygen values for DI and tap
Measured values include associated standard deviations. Values are presented
80
-------
Table 16. ANOVA of Dielectric Constant; Classical Sum of Squares, Type II.
Dielectric constant is dimensionless.
Sum of Squares
F-factor
p-value
Surfactant
11.52
3.02
0.0495
Water Type
0.33
0.35
0.5650
Model
13.10
1.53
0.2210
Mean
23.08
C.V.
4.23%
Standard Deviation
0.98
R2
0.4618
Maximum
25.07
R adj
0.1590
Minimum
21.03
R2
av pred
-0.5449
Table 17. Modeled and average measured dielectric constant values of DI and tap
solutions. Standard deviations are shown with their associated measured value. The data
is dimensionless.
DI
Tap
Modeled
Mean
±3.6%
Std Dev
Mean
±3.6%
Std Dev
Control
23.06
-
22.64
-
0.025% Steol
24.18
-
23.34
-
0.5% Dowfax
23.46
-
23.99
-
5% Dowfax
22.92
-
22.74
-
8% AMA 80-1
22.1
-
21.93
-
Measured
Control
23.2
1.045579
23.0
0.818176
0.025% Steol
24.4
0.828358
23.6
1.239808
0.5% Dowfax
23.8
0.96962
24.5
0.49405
5% Dowfax
23.2
0.787588
23.4
0.584103
8% AMA 80-1
22.6
1.206789
22.6
0.889812
81
-------
Table 18. Results summary of the mean measured responses over all experimental treatments. Real, imaginary, and specific
conductivity are reported in [j,S/cm. Units for dissolved oxygen (DO) are (J,g/L.
Water
Real Cond
±5.1%
Imaginary Cond
±4.6%
pH
±0.37%
Specific Cond
±10.1%
Dissolved Oxygen
±4.3%
Dielectric
±3.6%
Control
DI
12.77
0.0215
9.23
25.39
7465
23.06
Tap
220.23
0.0179
7.63
1048.14
8410.7
22.64
0.025% Steol
DI
Tap
15.07
234.17
0.0204
0.0200
9.28
7.47
33.66
1067.46
7502.8
5734.8
24.18
23.34
0.5% Dowfax
DI
Tap
82.80
274.41
0.0153
0.0296
9.26
7.98
371.33
1263.09
7250.7
7599.3
23.46
23.99
5% Dowfax
DI
Tap
489.82
653.84
0.0196
0.0241
8.84
8.35
2248.66
3082.47
7385
7607.8
22.92
22.74
8% AMA 80-1
DI
Tap
1793.6
1899.4
0.0300
0.0177
6.28
7.76
8364.93
8966.2
7408.2
7402.2
22.1
21.93
R2
_
0.9997
0.5297
0.9897
0.9999
0.6710
0.4618
R2
av pred
_
0.9992
-0.4086
0.9686
0.9998
0.0432
-0.5449
-------
Table 19. Summarized Response of 8% Surfactant Solutions and Percent Change from
Control. Control 1 is the DI control measured during the main 30 run experimental
program. Control 2 is a second DI control that was measured to coincide with the 8%
Steol and 8% Dowfax measurements. Percent change calculations were made using the
associated DI control. Means and standard deviations are in |iS/cm. % Change is in
percent.
Mean
Std Dev
% Change
Control 1
12.77
0.55
N/A
AMA 80-1
1793.6
58.35
13945
Control 2
37.27
1.4
N/A
Steol
497.5
14.1
1234
Dowfax
714.2
8.8
1816
Table 20. Summarized specific conductivity response of 0.5% surfactant solutions.
Control 1 is the DI control measured during the main 30 run experimental program.
Control 2 is a second DI control that was measured to coincide with the 0.5% AMA 80-1
measurements. Percent change calculations were made using the associated DI control.
Mean and standard deviation are reported in |iS/cm. % Change is reported in percent.
Mean
Std Dev
% Change
Control 1
8.61
2.73
NA
0.5% Dowfax
123.75
4.08
1337.86
Control 2
2.07
1.06
NA
0.5% AMA 80-1
199.02
0.11
9506.72
83
-------
DNAPL RELEASE
VADOSE ZONE
GROUNDWATER
FLOW
' DISSOLVED:
,, DNAPL POOLING
'-trrnTWrnz
FRACTURED
CLAY
J _DNAPL IN
r FRACTURE
GROUNDWATER
^ FLOW
LOWER
AQUIFER
\ DISSOLVED
PLUME
Figure 1. Schematic of a typical DNAPL release. Free phase DNAPL moves through
vadose zone, past the water table and through the saturated zone. DNAPL can flow
through fractures to contaminate lower strata, as well as pooling up-hydraulic-gradient.
A dissolved plume is pictured in the vadose zone, upper aquifer, and lower aquifer (After
Kueper and McWhorter, 1991).
Water Phase
Surfactant
Monomer
Interface
DNAPL Phase
Figure 2. Surfactants accumulating at water-DNAPL interface. Monomers amass at
the interface between the aqueous and NAPL phases. The hydrophilic head locates to the
water phase while the hydrophobic tail is in the NAPL phase (After Lowe et al., 1999).
84
-------
Carbon chain = hydrophobic tail
\
c c c c c c o
/\/\/\/\/\ /\ II
C C C C C C O—S—0'
Na+
Functional group = hydrophilic head
Figure 3. Surfactant monomer. The carbon chain acts as the hydrophobic group while
the functional group acts as the hydrophilic group. A schemati c of the monomer is
shown to the left of the arrow (After Lowe et al., 1999).
Surfactant
Monomer
NAPL Phase
Surfactant
Micelle
Figure 4. Surfactant micelle. Surfactant micelles form when the concentration of
surfactant added reaches the critical micelle concentration (CMC). Surfactant monomers
cluster together to form structures with hydrophobic interiors and hydrophilic exteriors.
NAPL contaminant molecules can collect in the interiors while the micelle itself is
soluble in the aqueous phase. This process effectively increases the solubility of the
contaminant (After Lowe et al., 1999).
85
-------
V\feste
Storage
Injection
\V\fells
Treatment
and
Separation
Brtractioir
\V\fe lis
Free-phase.
DNAPL
Figure 5. Schematic of surfactant-enhanced aquifer remediation. Free-phase DNAPL
is removed from the site using a traditional pump-and-treat method. Surfactant solution
is then injected into the subsurface via injection wells. The solution moves through the
contaminant plume (outlined in red dashed line), and the solubilized or mobilized
contaminant is extracted through a series of extraction wells. The extract is sent for
treatment and separation of the surfactant from the rest of the solution for continued use.
The brown shaded areas with dashed lines are lenses of low permeability material. Free-
phase DNAPL is shown in solid red. After Battelle and Duke Engineering Services,
2002.
Figure 6. Schematic of ionic conduction, which is a mechanism of electrical
conduction in which ions move through the pore spaces between grains via the pore fluid.
Grains are represented by orange spherical shapes, pore fluid is in blue.
86
-------
Figure 7. Schematic of an Electrical Double Layer (EDL). The EDL commonly occurs
around clay grains. A fixed layer of charged ions is adhered to the grain surface, while a
diffuse layer of charged ions is located adjacent to the grain surface, in the pore fluid.
The concentration of charged ions in the diffuse layer decreases with distance from the
grain surface. Grains are represented by orange spherical shapes, pore fluid is in blue.
Figure 8. Schematic of polarization of ions at the fluid-grain interface. Inducing an
electrical current at some frequency can cause polarization of ions in some materials.
The ions within the electrical double layer (EDL) and pore fluid segregate into positive
and negative groups on opposite sides of the fluid-grain interface, and slowly reintegrate
with the removal of the electrical current. Grains are represented by orange spherical
shapes, pore fluid is in blue.
87
-------
Solution
Output
6 cm
6 cm
6 cm
18 cm
Current
Electrode
Current
Electrode
Potential
Electrode
Cross-sectional View
Solution
input
1.75 cm
Figure 9. Diagram of experimental PVC apparatus. Diagram shows the layout of the
PVC apparatus, including locations of electrodes and input/output. The column is 18 cm
long and 6 cm is the spacing between electrodes. The column radius is 1.75 cm.
88
-------
X
Figure 10. Schematic of Wenner array. Four electrodes are separated by spacing "a".
The two current electrodes are located on the outside, with 2 potential electrodes located
between them. The recorded measurement represents conditions at location X.
Figure 11. Schematic of flow system. Surfactant solutions will flow in a closed loop
between a peristaltic pump, the PVC apparatus, and the flow cell. A T-valve exists in the
line between the pump and flow cell in order to add more surfactant solution from an
Erlenmeyer flask if necessary. If additional solution is not required to fill the flow cell,
the valve is closed. In the flow cell, the Troll 9500 will make DO, conductivity, and pH
measurements and relay them to a computer. Figure is not to scale.
89
-------
Column
MiniTrase
Figure 12. Schematic of time domain reflectometry laboratory set-up. The column is
filled with sand, after which solution flowed from the reservoir into the column by
gravity feed. After saturation was complete, measurements were taken using the
MiniTrase.
Figure 13. Photo of TDR waveguide. The waveguide is a buriable, 8cm long model.
There is a I cm between each prong.
90
-------
Design-Expert® Software
Log 10(Real Conductivity)
Lambda
Current = 0
Best = 0.13
Low C.I. = -0.1
High C.I. = 0.36
Recommend transform:
Log
(Lambda = 0)
Box-Cox Plot for Power Transforms
Lambda
Figure 14. Box-Cox Plot of Real Conductivity. This plot indicates that a log transform
of the real conductivity may be beneficial to minimizing and stabilizing the data
residuals.
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"13.28559
Normal Plot of Residuals
Log1
t
-0.84
internally Studentized Residuals
Figure 15. Scatter plot of residuals versus the normal percent probability of logio (real
conductivity). To verify the normality assumption, this plot should show a close fit of the
residuals to the red straight line. An indication of poor normality would be an "S" shape.
This plot appears to verify the normality assumption.
91
-------
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"13.28559
Log1
t
Residuals vs. Predicted
Predicted
Figure 16. Scatter plot of internally studentized residuals versus the predicted real
conductivity. This plot should show random scatter, indicating that the variance is
constant over the predicted range. There do not appear to be any trends in the residuals.
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"|3.28559
Residuals vs. Run
Log1
t
Run Number
Figure 17. Scatter plot of internally studentized real conductivity residuals versus run
number. Residual values below ±3.00 indicate that the proposed model of real
conductivity is fairly good. Random scatter indicates that the variance is constant over
all runs with no trends between residual and run number. All of the runs lie within the
confidence interval, and there do not appear to be any trends in the residuals based on run
order.
92
-------
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"13.28559
Log1
t
Residuals vs. surf
A:surf
Figure 18. Plot of real conductivity residuals by surfactant type and concentration.
This plot should show fairly consistent residual range across the 5 surfactant treatments.
The overall fit of the data is good.
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"|3.28559
Residuals vs. water type
Log1
t
B :water type
Figure 19. Plot of real conductivity residuals by water type. This plot should show
fairly consistent residual range between the two water types. The overall fit of the data is
good.
93
-------
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"13.28559
Log1
t
Leverage vs. Run
n ¦ n n n n n
I 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 I
13 17 21 25 29
Run Number
Figure 20. Leverage versus Run Number of Logio (Real Conductivity). Leverage
values at or above 2 times the leverage average may unduly influence at least one model
parameter. The logio (real conductivity) plot does not appear to show any points with
exceptional leverage.
Design-Expert® Software
Log 10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"13.28559
Log1
£
Externally Studentized Residuals
i 1 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i
1 5 9 13 17 21 25 29
Run Number
Figure 21. Externally Studentized Residuals versus Run Number of Logio (Real
Conductivity). This plot is used to indicate whether data falls inside of the 95%
confidence interval (t-test). All runs lie inside the confidence interval.
94
-------
Design-Expert® Software
Log10(Real Conductivity)
Color points by value of
Loa10(Real Conductivity):
"|3.28559
Cook's Distance
Log1
i
Run Number
Figure 22. Cook's Distance of Logio (Real Conductivity). It is a measure of how much
the estimated parameter, in this case logio (real conductivity), would change if a
particular run was omitted, and can be used to identify potential outliers. This plot does
not appear to identify any potential outliers.
3000
£
*
1500
500
2
Block
• DI Control
~ Tap Control
• 0.025% Steol, DI
~ 0.025% Steol, Tap
• 0.5% Dowfax, DI
~ 0.5% Dowfax, Tap
• 5% Dowfax, DI
~ 5% Dowfax, Tap
• 8% AMA 80-1, DI
~ 8% AMA 80-1, Tap
Figure 23. Plot of real conductivity results by block. The plot indicates that real
conductivity values for individual surfactant treatments are consistent through all three
runs.
95
-------
B: water type
.025 steol
0.5 dowfiax
5 dowfiax
A: surf
Figure 24. Plot of modeled and measured data of logio (real conductivity). The green
triangles and error bars connected with a dotted line represents the modeled real
conductivity in tap water solutions. The green circles represent the measured data. The
red square and error bars connected with a dashed line represent the modeled real
conductivity in DI solutions. The red circles represent the measured data.
Control
0.025% Steol 0.5%Dowfax 5%Dowfax 8%AMi80-I
Surfa ct ant T reatment
Figure 25. Plot of Real Conductivity versus Surfactant Treatment. Data in blue
represents the untransformed real conductivity values of DI solutions. The data in orange
represent the tap solutions.
96
-------
2501
2001
S
0
1
£. 1S01
• P*
¦c
o
3
efi
1001
501
•yss
g
i-i-i-i-ij-14-i-i I > I *r^i
0.01 0.1 1 10 130 1000 10000 100000
Frequency (Hz)
-DT Control
-Tap Control
¦0.025°/6Steol,DI
0.025%Steol,Tap
-0.5%Dowfax,DI
-0.5%Dowfax,Tap
-5%Dowfax,DI
- 5% Dowfsx, T ap
¦8% AM A 80-1,1)1
¦8% AM A 80-1, Tap
Figure 26. Real conductivity responses of each surfactant treatment over the range of
measured frequencies (12000-0.091Hz). Response does not appear to be greatly affected
by frequency over the measured frequencies, although there does appear to be a slight
drop in value at the uppermost end of the frequency spectrum. This is most likely a result
of instrument noise.
Design-Expert® Software
Imaginary Conductivity
Lambda
Current = 1
Best = 0.01
LowC.I. = -1.55
High C.I. = 1.45
Recommend transform:
None
(Lambda = 1)
Box-Cox Plot for Power Transforms
Lambda
Figure 27. Box-Cox Plot of Imaginary Conductivity. This plot indicates a transform of
the imaginary conductivity response is unlikely to aid in minimizing and stabilizing the
data residuals.
97
-------
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Normal Plot of Residuals
-1.92
internally Studentized Residuals
Figure 28. Scatter plot of residuals versus the normal percent probability of imaginary
conductivity. To verify the normality assumption, this plot should show a close fit of the
residuals to the red straight line. An indication of poor normality would be an "S" shape.
This plot appears to verify the normality assumption.
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Residuals vs. Predicted
—i 1—
0.014 0.019
1 1
0.028 0.032
Predicted
Figure 29. Scatter plot of internally studentized residuals versus the predicted
imaginary conductivity. This plot should show random scatter, indicating that the
variance is constant over the predicted range. There do not appear to be any trends in the
residuals although Run #19 (8AMA, DI) is outside of the confidence interval.
98
-------
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Residuals vs. Run
Run Number
Figure 30. Scatter plot of internally studentized residuals versus run number. Residual
values below ±3.00 indicate that the proposed model of imaginary conductivity is good.
Random scatter indicates that variance is constant over all runs with no apparent trends.
All but one of the runs lies within the confidence interval, and there are no apparent
trends in the residuals. Run #19 is outside the confidence interval.
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
~ 0.0415
0.01141
Residuals vs. surf
A: surf
Figure 31. Plot of residuals by surfactant type and concentration. This plot should
show fairly consistent range across the 5 surfactant treatments. The overall fit is good,
although there appears to be a slight megaphone shape towards the higher surfactant
concentrations.
99
-------
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Residuals vs. water type
~
1.43 —
~
~
B
1
~
0
0.13 —
~
u
S
~
y
~
1.70
3.27 —
B :water type
Figure 32. Plot of residuals by water type. This plot should show fairly consistent
residual range between the 2 water types. DI (type 1) displays a larger range of residuals
than tap (type 2). However, if the most negative residual, Run #19 is removed, the 2
water types show much greater consistency.
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
~ 0.0415
0.01141
Leverage vs. Run
Run Number
Figure 33. Leverage versus Run Number of Imaginary Conductivity. Leverage values
at or above 2 times the leverage average may unduly influence at least one model
parameter. The imaginary conductivity plot does not appear to show any points with
exceptional leverage.
100
-------
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Externally Studentized Residuals
~ ~ ~ 1=1
I 1 1 1 I 1 1 1 I 1 1 1 I
I 1 1 1 I
Run Number
Figure 34. Externally Studentized Residuals versus Run Number of Imaginary
Conductivity. This plot is used to indicate whether data falls inside of the 95%
confidence interval (t-test). All runs except one lie inside the confidence interval. Run
19, 8% AMA, tap, lies outside. It should be carefully evaluated for outlier potential.
Design-Expert® Software
Imaginary Conductivity
Color points by value of
Imaginary Conductivity:
DO. 0415
0.01141
Cook's Distance
Run Number
Figure 35. Cook's Distance of Imaginary Conductivity. It is a measure of how much
the estimated parameter, in this case imaginary conductivity, would change if a particular
run was omitted, and can be used to identify potential outliers. This plot, as in the t-test,
identifies Run 19 as being a potential outlier. The Cook's D is not sufficiently high to
omit from analyses.
101
-------
i
I
~
~
• DI Control
~ Tap Control
• 0.025% Steol, DI
~ 0.025% Steol, Tap
• 0.5% Dowfax, DI
~ 0.5% Dowfax, Tap
• 5% Dowfax, DI
~ 5% Dowfax, Tap
• 8% AM A 80-1, DI
~ 8% AM A 80-1, Tap
2
Block
Figure 36. Plot of imaginary conductivity results by block. The plot indicates that
imaginary conductivity values for individual surfactant treatments are inconsistent
through all three runs.
B: water type
-005
-005
-006
-006
-006
.025 steol
0.5 dowfax
5 dowfax
8 am a- 08- i
none
Figure 37. Plot of modeled and measured data of imaginary conductivity. The green
triangles and error bars connected with a dotted line represents the modeled imaginary
conductivity in tap water solutions. The green circles represent the measured data. The
red square and error bars connected with a dashed line represent the modeled imaginary
conductivity in DI solutions. The red circles represent the measured data.
102
-------
10
DI Control
Tap Control
0.025%Steol,DI
0.5%Dowfax,DI
0.1
5%Dowfax,DI
8%AMA 80-1. DI
8%AMA 80-1, Tap
0.001
0.01
0.1
1000
10000 100000
Figure 38. Imaginary conductivity responses of each surfactant treatment over the
range of measured frequencies (0.091 Hz-12 kHz).
¦DI Control
Tap Control
0.025°oSteol,DI
0.025% Steol, Tap
0.59'oDowfax, DI
8°oAMA 80-1, DI
8° .1 AM A 80-1. Tap
0.001
0.1 1 10 100 1000
Frequency (Hz)
Figure 39. Imaginary conductivity responses of each surfactant treatment over a
limited frequency range (0.366-187.5 Hz).
103
-------
Design-Expert® Software
pH
Lambda
Current = 1
Best = 0.34
Low C.I. = -2.25
High C.I. = 3.46
Recommend transform:
None
(Lambda = 1)
Box-Cox Plot for Power Transforms
CO
CO
Lambda
Figure 40. Box-Cox Plot of pH. This plot indicates a transform of the pH is unlikely to
aid in minimizing and stabilizing the data residuals.
Design-Expert® Software
pH
Color points by value of
pH:
9.44
6.16
Normal Plot of Residuals
Internally Studentized Residuals
Figure 41. Scatter plot of residuals versus the normal percent probability of pH. To
verify the normality assumption, this plot should show a close fit of the residuals to the
red straight line. An indication of poor normality would be an "S" shape. This plot
appears to verify the normality assumption.
104
-------
Design-Expert® Software
pH
Color points by value of
pH:
B
9.44
6.16
to
>»
Residuals vs. Predicted
~
~ ~
~~r~
8.59
~
~
&
~
Predicted
Figure 42. Scatter plot of internally studentized residuals versus the predicted pH. This
plot should show random scatter, indicating that the variance is constant over the
predicted range. There do not appear to be any trends in the residuals.
Design-Expert® Software
pH
Color points by value of
p_H:
9.44
6.16
Residuals vs. Run
CD
>»
Run Number
Figure 43. Scatter plot of internally studentized residuals versus run number. Residual
values below ±3.00 indicate that the proposed model of pH is fairly good. Random scatter
indicates that the variance is constant over all runs with no trends between residual and
run number. All of the runs lie within the confidence interval, and there do not appear to
be any trends in the residuals based on run order.
105
-------
Design-Expert® Software
PH
Color points by value of
pH:
D9.44
6.16
Residuals vs. surf
A: surf
Figure 44. Plot of residuals by surfactant type and concentration. This plot should
show fairly consistent residual range across the 5 surfactant treatments. The Dowfax
treatments (3 and 4) have slightly larger ranges, but still fall within the confidence
interval.
Design-Expert® Software
PH
Color points by value of
pH:
9.44
Residuals vs. water type
E
6.16
B:water type
Figure 45. Plot of residuals by water type. This plot should show fairly consistent
residual range between the two water types. The overall fit of the data is good, with
fairly consistent spreads and all data points within the confidence interval. Note that the
measured pH values of the DI solutions (water type 1) read both the highest and lowest,
while the tap solutions (water type 2) appear to lie in the middle values.
106
-------
Design-Ex pert® Software
pH
Color points by value of
pH:
I"'
|_J 6.16
Leverage vs. Run
Run Number
Figure 46. Leverage versus Run Number of pH. Leverage values at or above 2 times
the leverage average may unduly influence at least one model parameter. The pH plot
does not appear to show any points with exceptional leverage.
Design-Expert® Software
PH
Color points by value of
pH:
~ 9.44
6.16
Externally Studentized Residuals
Run Number
Figure 47. Externally Studentized Residuals versus Run Number of pH. This plot is
used to indicate whether data falls inside of the 95% confidence interval (t-test). All runs
lie inside the confidence interval.
107
-------
Design-Experl® Software
pH
Color points by value of
pH:
9.44
Cook's Distance
D
6.16
o
o
O
Run Number
Figure 48. Cook's Distance of pH. It is a measure of how much the estimated
parameter, in this case imaginary conductivity, would change if a particular run was
omitted, and can be used to identify potential outliers. The plot does not indicate any
potential outliers.
a
e.
14
12
10
8
6
4
2
0
I
~
2
Block
• DI Control
• Tap Control
• 0.025° b Steol, DI
0.025% Steol, Tap
• 0.5% Dowfax, DI
• 0.5% Dowfax, Tap
• 5% Dowfax, DI
4 5% Dowfax, Tap
• 8% AMA 80-1. DI
• 8°h AMA 80-1. Tap
Figure 49. Plot of pH results by block. The plot indicates that pH values for individual
surfactant treatments are consistent through all three runs, although there is significant
overlap among treatments.
108
-------
Interaction
B: water type
1 1 T
,025steol 0.5 dowfax 5 dowfax
Figure 50. Plot of modeled and measured data pH. The green triangles and error bars
connected with a dotted line represents the modeled pH in tap water solutions. The green
circles represent the measured data. The red square and error bars connected with a
dashed line represent the modeled pH in DI solutions. The red circles represent the
measured data.
14
12
10
DI
Tap
Control 0.025% Steol 0.5% Dowfax 5% Dowfax 8% AMA 80-1
Sur fa eta nt Tr ea tment
Figure 51. Bar graph of measured pH responses. The bar represents the median value,
while the upper error bar is the treatment's maximum, and the lower error bar is the
minimum.
109
-------
Design-Expert® Software
Sqrt(cond)
Lambda
Current = 0.5
Best = 0.73
LowC.I. = 0.52
High C.I. = 0.95
Recommend transform:
Square root
(Lambda = 0.5)
Box-Cox Plot for Power Transforms
CO
CO
Lambda
Figure 52. Box-Cox Plot of Specific Conductivity. This plot indicates that a square
root transform of the specific conductivity may be beneficial to minimizing and
stabilizing the data residuals.
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~ 94.6764
4.28719
Normal Plot of Residuals
Internally Studentized Residuals
Figure 53. Scatter plot of residuals versus the normal percent probability of sqrt
(specific conductivity). To verify the normality assumption, this plot should show a close
fit of the residuals to the red straight line. An indication of poor normality would be an
"S" shape. This plot appears to verify the normality assumption.
110
-------
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
¦ 94.6764
14.28719
Residuals vs. Predicted
Predicted
Figure 54. Scatter plot of internally studentized residuals versus the predicted sqrt
(specific conductivity). This plot should show random scatter, indicating that the
variance is constant over the predicted range. There is a megaphone shape to the plot,
with the largest range of residuals appearing in the lowest values predicted conductivity.
This may be a function of normal variation in the DI water used in the experiments.
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~ 94.6764
4.28719
Residuals vs. Run
Run Number
Figure 55. Scatter plot of internally studentized residuals versus run number. Residual
values below ±3.00 indicate that the proposed model of sqrt (specific conductivity) is
fairly good. Random scatter indicates that the variance is constant over all runs with no
trends between residual and run number. All of the runs lie within the confidence
interval, and there do not appear to be any trends in the residuals based on run order.
Ill
-------
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~ 94.6764
4.28719
Residuals vs. surf
A:surf
Figure 56. Plot of residuals by surfactant type and concentration. This plot should
show fairly consistent residual range across the 5 surfactant treatments although the
control groups (no surfactant) appear to have larger residual ranges.
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~ 94.6764
4.28719
Residuals vs. water type
B:water type
Figure 57. Plot of residuals by water type. This plot should show fairly consistent
residual range between the two water types. DI (type 1) has a much larger range of
residuals than the tap treatments (type 2).
112
-------
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~\ 94.6764
0
4.28719
Leverage vs. Run
!~~~~¦~¦ ~~
Run Number
Figure 58. Leverage versus Run Number of Sqrt (Specific Conductivity). Leverage
values at or above 2 times the leverage average may unduly influence at least one model
parameter. The sqrt (specific conductivity) plot does not appear to show any points with
exceptional leverage.
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
¦ 94.6764
14.28719
Externally Studentized Residuals
Run Number
Figure 59. Externally Studentized Residuals versus Run Number of Sqrt (Specific
Conductivity). This plot is used to indicate whether data falls inside of the 95%
confidence interval (t-test). All runs lie inside the confidence interval.
113
-------
Design-Expert® Software
Sqrt(cond)
Color points by value of
Sqrt(cond):
~ 94.6764
4.28719
Cook's Distance
Run Number
Figure 60. Cook's Distance of Sqrt (Specific Conductivity). It is a measure of how
much the estimated parameter, in this case sqrt (specific conductivity), would change if a
particular run was omitted, and can be used to identify potential outliers. This plot does
not appear to identify any potential outliers.
10000
£
9000
u
8000
a.
7000
6000
3
1
5000
o
U
4000
u
5
3000
Q.
2000
5C
1000
0
2
Block
• DI Control
• Tap Control
• 0.025% Steol, DI
0.025% Steol. Tap
• 0.5% Dowfax, DI
• 0.5% Dowfax, Tap
• 5% Dowfax, DI
4 5% Dowfax, Tap
• 8°h AMA 80-1. DI
• 8°h AMA 80-1. Tap
Figure 61. Plot of specific conductivity by block. The plot indicates that specific
conductivity values for individual surfactant treatments are consistent through all three
runs.
114
-------
B: water type
Figure 62. Plot of modeled and measured data of sqrt (specific conductivity). The
green triangles and error bars connected with a dotted line represents the modeled
specific conductivity in tap water solutions. The green circles represent the measured
data. The red square and error bars connected with a dashed line represent the modeled
specific conductivity in DI solutions. The red circles represent the measured data.
v
x
3
tr*
>
z
o
U
u
a
*C
D
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
¦DI
Tap
Control
t
r »
0.025% Steol 0.5%Dowfax 5%Dowfax 8%AMA80-I
Sur fa eta nt Tr en tment
Figure 63. Plot of Specific Conductivity versus Surfactant Treatment. Data in blue
represents the untransformed specific conductivity values of DI solutions. The data in
orange represent the tap solutions.
115
-------
V
x
3
>,
P
z
o
U
u
a
*C
D
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
a
Control 0.025% Steol 0.5% Dowfax 5% Dowfax 8% AMA 80-1
Sur fa eta nt Tr ea tment
Figure 64. Bar graph of specific conductivity. The bar represents the median value,
while the upper error bar is the treatment's maximum, and the lower error bar is the
minimum.
Design-Expert® Software
do
Lambda
Current = 1
Best = 3
Low C.I. =
High C.I. =
Recommend transform:
None
(Lambda = 1)
Box-Cox Plot for Power Transforms
Lambda
Figure 65. Box-Cox Plot of Dissolved Oxygen. This plot indicates a transform of the
dissolved oxygen is unlikely to aid in minimizing and stabilizing the data residuals.
116
-------
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
Normal Plot of Residuals
internally Studentized Residuals
Figure 66. Scatter plot of residuals versus the normal percent probability of dissolved
oxygen. To verify the normality assumption, this plot should show a close fit of the
residuals to the red straight line. An indication of poor normality would be an "S" shape.
This plot does not appear to verify the normality assumption.
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
Residuals vs. Predicted
CO
>
~ft
~
ft
**3
¦a
~~
Predicted
Figure 67. Scatter plot of internally studentized residuals versus the predicted dissolved
oxygen. This plot should show random scatter, indicating that the variance is constant
over the predicted range. While the high predicted values appear to have randomly
scatter residuals, the low predicted values have a much larger range. This may
correspond to the anomalously low measured dissolved oxygen values in the 0.025%
Steol, tap experimental treatment.
117
-------
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
Residuals vs. Run
[
f0—^
^ i/\n K
fv \ / b
Run Number
Figure 68. Scatter plot of internally studentized residuals versus run number. Residual
values below ±3.00 indicate that the proposed model of dissolved is fair. Random scatter
indicates that variance is constant over all runs with no apparent trends. Run #9 lies
outside of the confidence interval, which corresponds to a 0.025% Steol, tap treatment
DO value that is much higher than other similar treatments.
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
Residuals vs. surf
A: surf
Figure 69. Plot of residuals by surfactant type and concentration. This plot should
show fairly consistent range across the 5 surfactant treatments. With the exception of
Run #9 and Run #15, both corresponding to 0.025% Steol, the overall fit of the data is
good.
118
-------
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
Residuals vs. water type
3.67 —
~
w
ro
"a
2.00 —
if)
O)
(Z
"O
0)
N
B
Q
C
a)
0.33 —
H
CO
B
1
>>
~
~
to
c
~
~
0)
c
¦
B:water type
Figure 70. Plot of residuals by water type. This plot should show fairly consistent
residual range across the two water types. With the exception of Run #9 and Run #15,
both in tap water (type 2), the overall fit of the data is good.
D e s ig n-Expe rt® S of twa re
do
Color points by value of
do:
~ 8644
4375.5
Leverage vs. Run
~ ~~~C ~¦~~~~~~ ~~
I 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 I 1 1 1 I 1
1 5 9 13 17 21 25 29
Run Number
Figure 11. Leverage versus Run Number in Dissolved Oxygen. Leverage values at or
above 2 times the leverage average may unduly influence at least one model parameter.
The dissolved oxygen plot does not appear to show any points with exceptional leverage.
119
-------
Design-Expert® Software
do
Color points by value of
do:
~ 8644
4375.5
"O 5.75 —
Externally Studentized Residuals
Ja ~
I 1 1 1 I
I 1 1 1 I
1 I 1 1 1 I 1
21 25
Run Number
Figure 72. Externally Studentized Residuals versus Run Number of Dissolved Oxygen.
This plot is used to indicate whether data falls inside of the 95% confidence interval (t-
test). Run 9, corresponding to 0.025% Steol, tap, is well outside of the confidence
interval.
Design-Expert® Software
do
Color points by value of
do:
¦ 8644
14375.5
Cook's Distance
nnpn
Run Number
Figure 73. Cook's Distance of Dissolved Oxygen. It is a measure of how much the
estimated parameter, in this case dissolved oxygen, would change if a particular run was
omitted, and can be used to identify potential outliers. This plot, as in the t-test, identifies
Run 9 as being a potential outlier. The Cook's D is not sufficiently high to omit from
analyses. Run 9 corresponds to 0.025% Steol, tap.
120
-------
10000
9000
8000
J
j* 7000
<«
bX
j? 5000
o
13
r
T
i
n "
6000
4000
3000
2000
1000
i
~
I
2
Block
• DI Control
• Tap Control
• 0.025% Steol, DI
0.0.25% Steol Tap
• 0.5% Dowfax, DI
• 0.5% Dowfax Tap
• 5% Dowfax, DI
4 5% Dowfax, Tap
• 8° h AMA 80-1. DI
• 8°h AMA 80-1. Tap
Figure 74. Plot of dissolved oxygen by block. Plot indicates that there is significant
overlap among treatments. The tap control is consistently higher than the other
treatments over all three runs, while the 0.025% Steol, tap treatment is substantially lower
in blocks 2 and 3 than in block 1.
11 ilci aouui i
B: water type
I
Figure 75. Plot of the modeled and measured dissolved oxygen. The green triangles
and error bars connected with a dotted line represent the modeled dissolved oxygen in tap
water solutions. The green circles represent the measured data. The red square and error
bars connected with a dashed line represent the modeled dissolved oxygen in DI
solutions. The red circles represent the measured data.
121
-------
10000
9000
DI
Tap
Control 0.025% Steol 0.5% Dowfax 5% Dowfax 8% AMA 80-1
Sur fa eta nt Tr ea tment
Figure 76. Bar graph of measured dissolved oxygen responses. The bar represents the
median value, while the upper error bar is the treatment's maximum, and the lower error
bar is the minimum.
Design-Expert® Software
dielectric
Lambda
Current = 1
Best = 2.4
Low C.I. = -8.97
High C.I. = 13.76
Recommend transform:
None
(Lambda = 1)
Box-Cox Plot for Power Transforms
-------
Design-Expert® Software
dielectric
Color points by value of
dielectric:
D 25.07
21.03
Normal Plot of Residuals
Internally Studentized Residuals
Figure 78. Scatter plot of residuals versus the normal percent probability of dielectric
constant. To verify the normality assumption, this plot should show a close fit of the
residuals to the red straight line. An indication of poor normality would be an "S" shape.
This plot appears to verify the normality assumption.
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Residuals vs. Predicted
Dn
I I I I I
21.66 22.33 23.01 23.68 24.35
Predicted
Figure 79. Scatter plot of internally studentized residuals versus the predicted dielectric
constant. This plot should show random scatter, indicating that the variance is constant
over the predicted range. There do not appear to be any trends in the residuals.
123
-------
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Residuals vs. Run
Run Number
Figure 80. Scatter plot of internally studentized residuals versus run number. Residual
values below ±3.00 indicate that the proposed model of dissolved is fairly good. Random
scatter indicates that the variance is constant over all runs with no trends between residual
and run number. All of the runs lie within the confidence interval, and there do not
appear to be any trends in the residuals based on run order.
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Residuals vs. surf
CO
A:surf
Figure 81. Plot of residuals by surfactant type and concentration. This plot should
show fairly consistent range across the 5 surfactants. The overall fit is good, although the
5.0% Dowfax treatment has a much smaller range than the other surfactants.
124
-------
Design-Expert® Software
dielectric
Color points by value of
dielectric:
D
25.07
21.03
Residuals vs. water type
B:water type
Figure 82. Plot of residuals by water type. This plot should show fairly consistent
range in residuals between the two water types. Type 1 is DI; Type 2 is tap water. The
overall fit of the data is good.
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Leverage vs. Run
~~~~~~~~ ~¦~~~ ~~
Run Number
Figure 83. Leverage versus Run Number in Dielectric Constant. Leverage values at or
above 2 times the leverage average may unduly influence at least one model parameter.
The dielectric constant plot does not appear to show any points with exceptional leverage.
125
-------
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Externally Studentized Residuals
Run Number
Figure 84. Externally Studentized Residuals versus Run Number in Dielectric
Constant. This plot is used to indicate whether data falls inside of the 95% confidence
interval (t-test). All data points lie within the confidence interval.
Design-Expert® Software
dielectric
Color points by value of
dielectric:
~ 25.07
21.03
Cook's Distance
~ nan
Run Number
Figure 85. Cook's Distance of Dielectric Constant. It is a measure of how much the
estimated parameter, in this case dielectric constant, would change if a particular run was
omitted, and can be used to identify potential outliers. This plot does not appear to
identify any potential outliers.
126
-------
26
25
24 I
E
«
!/3
E
0
•a 23'
~
~
22
21
!
~
t
~ •
20
2
Block
~ DI Control
~ Tap Control
~ 0.025° o Steol, DI
~ 0.025° b Steol. Tap
~ 0.5° o Dowfax DI
~ 0.5° o Dowfax Tap
5° o Dowfax DI
~ 5° o Dowfax Tap
~ 8° b AMA 80-1. DI
~ 8° b AMA 80-1. Tap
Figure 86. Plot of dielectric constant values by block and surfactant treatment. The
plot indicates that dielectric constant value in an individual surfactant treatment is
inconsistent across runs. There is also substantial overlap among all treatments.
ii ilci aouui i
B: water type
A: surf
Figure 87. Plot of the modeled and measured dielectric constant. The green triangles
and error bars connected with a dotted line represent the modeled dielectric constant in
tap water solutions. The green circles represent the measured data. The red square and
error bars connected with a dashed line represent the modeled dielectric constant in DI
solutions. The red circles represent the measured data.
127
-------
26
Control 0.025% Steol 0.5% Dowfax 5% Dowfax 8% AMA 80-1
Sur fa eta nt Tr ea tment
Figure 88. Bar graph of measured dielectric constant values. The bar represents the
median value, while the upper error bar is the treatment's maximum, and the lower error
bar is the minimum.
0
I.
Na
,' T
on
o
"0
Figure 89. Molecular structure of Steol CS-330 (sodium laureth sulfate). Key Centre
for Polymer Colloids, University of Sydney, Australia.
128
-------
Figure 90. Molecular structure of Dowfax 8390. www, chemicalregister.com
o o"
XJc
Na
0
Figure 91. Molecular structure of Aerosol MA 80-1. Key Centre for Polymer Colloids.
University of Sydney, Australia.
129
-------
i Werkema 2008
i Magili Thesis
•c
3 4000
£
a
O 3000
O
s
v
24 2000
Surfactant Treatment
Figure 92. Comparison plot of specific conductivity results in Werkema 2008 and
Magill thesis. This graph indicates that specific conductivity trend with respect to
surfactant treatment is consistent in both investigations.
130
-------
APPENDIX A
SYSTEMATIC ERROR RESULTS
Spectral Induced Polarization
Five randomly chosen 18 cm PVC columns were filled with 283.0 ±1.6 g (±0.5%)
sand and saturated with 64.4 ±0.72 g (±1.1%) DI water. As eighteen columns were used
to perform all of the SIP tests, 27.8% of the columns were tested for systematic error.
Spectral induced polarization (SIP) measurements were made over 2 days. The data
collected are located in Table 21. All measurements are in |iS/cm. Three columns,
numbers 9, 2, and 18, were tested on November 10, 2008 and have been examined
together. The other two columns, numbers 3 and 16, were tested on November 11, 2008
and so have been examined separately from the first 3. This was done in order to limit
error due to daily environmental (i.e. laboratory temperature fluctuations, etc.) changes
and instrument drift.
Averages and standard deviations of the data are located in Table 22. Two separate
averages were calculated to eliminate the daily variability. Columns 9, 2, and 18 were
averaged together separately from columns 3 and 16. There is variation in both real and
imaginary conductivity in the very high end of the frequency spectrum. The variability
lessens at 187 Hz and lower frequencies. This is important as the range of interest is
between 93.75 and 0.366 Hz. The real conductivity variation at 11.7 Hz, the frequency
analyzed in this research, is ±5.1% from the average. This was determined by calculating
the percent error of the measured real conductivity values of each column relative to the
average real conductivity value (Eqn 17). This was done separately for each day's
measurements.
131
-------
_! I
PE = ^L< ^*100 (17)
avg
This equation describes the method for calculating the percent error. PE is the
percent error, a'mg is the average real conductivity, and cr'act is the actual real
conductivity. In addition, this equation was modified to calculate the errors associated
with imaginary conductivity, dielectric constant, and the water quality measurements.
The largest percent error was then chosen to represent the mean system error. In this
case, the mean system error is defined as error attributable to physical differences
between columns, packing, water to sand ratios, and instrument drift and error. The
imaginary conductivity variation at the same frequency is ±4.6% from the average. The
error for imaginary conductivity was calculated in the same manner as real conductivity.
Plotting the real conductivity by frequency for the columns separated per day (Figures
93, 94) shows that all columns have the same general trend, with stable values through
the low and middle sections of the frequency range. All columns show a real
conductivity drop between 750 and 1500 Hz and continue to fall through the highest
frequencies. Similarly, the imaginary conductivity plotted by column number and
separated by day (Figure 95, 96) appears to follow a similar trend which lies within the
same value range regardless of column number. Columns 9, 2, and 18 show a slightly
different shape through the frequency range than columns 3 and 16, but all columns'
imaginary response over the measured frequency range are similar in shape to others
tested on the same day. This analysis suggest the differences between columns due to
the packing method and column preparation results in a mean system error of 5.1% for
132
-------
real conductivity and 4.6% for imaginary conductivity. This systematic error has not
been added to the measured data, but is included as a heading in tables.
The amount of sand and solution added to each column is located in Table 23. Also
in this table is the sand to water ratio of each column. This ratio has been plotted against
the average measured real conductivity (Figure 97) and imaginary conductivity (Figure
98) to identify any related trends. There do not appear to be any clear relationships
between the measured real and imaginary conductivity values and the sand to water ratio
within the range of ratios.
pH, Specific Conductivity, Dissolved Oxygen
Four randomly chosen 18cm PVC columns were filled with 283.0 ±1.6 g (±0.5%)
sand and saturated with 64.4 ±0.72 g (±1.1%) tap water. In total, fifteen different
columns were used for water quality measurements. As such, 26.7% of the columns were
tested for systematic error. pH, specific conductivity, and dissolved oxygen
measurements were made using the In-Situ, Inc.'s Troll 9500 multi-parameter water
quality monitoring instrument. The data, collected over 3 days, are located in Table 24.
Over the measured time ranges, temperature, pH, conductivity, and saturated RDO
(Rugged Dissolved Oxygen) were averaged within each different column to yield a
percent change in each parameter. Each parameter's data range and the largest calculated
percent change follow. Temperature overall ranges from 22.60 to 26.68°C. Within each
column, the largest percent difference in temperature was ±11.8%. pH measurements
ranged between 7.53 and 8.13, with a maximum percent difference of ±0.37%>. Specific
conductivity measurements ranged between 128.4 and 534.6 |iS/cm, with a maximum
133
-------
percent difference of ±10.1%. DO measurements ranged from 6619 to 7794 |ig/L, with a
maximum percent difference of ±4.3%.
The combined calculations (Table 25), display differences among columns in the
conductivity, DO, and saturated DO parameters. These differences are most likely due to
laboratory temperature or other environmental differences in the laboratory as the
experiments were run on 3 separate days. Columns 2 and 3 were run on the same day,
followed by column 16 the next day, and column 18 the day after. Additional specific
conductivity measurements were made using an Accumet 4-electrode specific
conductivity probe as a quality check for the Troll 9500. The readings made with this
instrument corroborated the differences in conductivity among the columns, suggesting
that the specific conductivity value of the tap water was not consistent for the entirety of
the systematic error tests.
Dielectric Constant
Systematic error tests were performed on the TDR apparatus and four columns. A
sand mass of 8535 ±218 g (±2.5%) was loaded into each column. The column was
saturated with 1863 ±100 mL (±5.4%) of DI water by gravity feed infiltration through the
bottom of the column (Figure 12). Six dielectric constant measurements were made on
each column in two groups of three. The first set was collected immediately upon
saturation, with the second set following ten minutes later. All systematic error testing of
time domain reflectometry was performed on November 8, 2009. In total, five columns
were used in the TDR tests, resulting in systematic error testing of 80% of the columns.
134
-------
The measured dielectric constant values are located in Table 25. The values are
separated by column, and from there into readings. Overall, the first readings appear to
be slightly higher than the second set of readings across all four columns, although there
is overlap between the two sets. In addition, standard deviations of readings, both within
a column and among the columns, are small.
The amount of sand and solution added to each column is located in Table 26. Also
in this table is the sand to water ratio of each column. This ratio has been plotted against
the average dielectric constant measured to identify any related trends (Figure 99). There
do not appear to be any clear relationship between the measured dielectric constants and
the sand to water ratio, although the ratio range measured is small.
Based on the dielectric constants measured during the systematic error tests, the
physical differences between columns should not contribute substantially to the water
quality responses. However, the percent error in the sand to water ratio is ±5.5%, while
the maximum percent error of measured dielectric values is ±3.6%. It is possible that the
ratio, related to packing error, could account for the majority of error in the dielectric
constant measurements.
135
-------
Table 21.
Systematic error tests for SIP response. Data is reported in |iS/cm.
G\
Date of test
11/10/2009
11/10/2009
11/10/2009
11/11/2009
11/11/2009
Column #9
Column #2
Column #18
Column #16
Column #3
Frequency
(Hz)
Real
Imaginary
Real
Imaginary
Real
Imaginary
Real
Imaginary
Real
Imaginary
12000
3.82
3.820
13.35
3.405
1.21
4.367
-16.33
2.976
6.71
2.225
6000
14.46
1.905
15.75
1.731
-9.18
2.212
-8.46
1.459
-12.9
1.166
3000
-15.70
0.967
-16.72
0.896
-5.60
1.130
-13.20
0.791
-4.60
0.559
1500
-2.94
0.497
-0.57
0.470
-9.06
0.580
7.63
0.403
8.70
0.319
750
9.70
0.265
-1485.9
-3198.47
6.37
0.295
15.96
0.215
15.25
0.154
375
14.28
0.142
14.99
0.138
11.97
0.167
18.64
0.111
16.24
0.113
187.5
15.57
0.082
16.38
0.064
14.24
0.081
19.12
0.081
17.06
0.062
93.75
16.07
0.042
16.61
0.041
14.58
0.058
19.35
0.060
17.16
0.049
46.875
16.16
0.028
16.70
0.028
14.86
0.028
19.46
0.046
17.27
0.034
23.4375
16.18
0.021
16.73
0.019
14.91
0.019
19.50
0.040
17.28
0.031
11.71875
16.19
0.018
16.73
0.021
14.90
0.021
19.54
0.030
17.28
0.031
5.859375
16.20
0.015
16.75
0.015
14.93
0.016
19.55
0.028
17.30
0.025
2.929687
16.20
0.014
16.75
0.015
14.96
0.013
19.58
0.024
17.33
0.019
1.464844
16.20
0.013
16.78
0.012
15.01
0.012
19.62
0.019
17.33
0.019
0.732422
16.21
0.011
16.81
0.015
15.10
0.013
19.70
0.014
17.43
0.013
0.366211
16.22
0.012
16.88
0.013
15.26
0.011
19.82
0.013
17.52
0.009
0.183105
16.25
0.011
16.98
0.013
15.51
0.012
20.05
0.011
17.67
0.009
0.091553
16.30
0.014
17.15
0.013
15.85
0.015
20.42
0.008
17.95
0.006
-------
Table 22. Averages and standard deviations of SIP systematic error tests. Calculations
combine readings from the five columns. Readings from columns were initially averaged
according to day of experiment, and then those averages were used to calculate an overall
average.
Real (|iS/cm)
Imaginary (jiS/cm)
Frequency
(Hz)
Average
(±5.1%)
Stdev
Average
(±4.6%)
Stdev
12000
5.26
33.23
10.076
2.448
6000
-0.21
41.89
5.084
1.207
3000
-33.49
17.08
2.607
0.638
1500
2.26
22.35
1.361
0.295
750
-863.20
2009.48
-1918.52
4291.510
375
45.66
7.39
0.403
0.070
187.5
49.42
5.44
0.222
0.030
93.75
50.26
5.23
0.150
0.026
46.875
50.67
5.07
0.099
0.023
23.4375
50.76
5.06
0.079
0.028
11.71875
50.78
5.13
0.072
0.017
5.859375
50.84
5.11
0.060
0.018
2.929687
50.89
5.11
0.051
0.014
1.464844
50.97
5.12
0.046
0.011
0.732422
51.15
5.14
0.040
0.005
0.366211
51.42
5.15
0.034
0.005
0.183105
51.87
5.22
0.034
0.004
0.091553
52.59
5.41
0.034
0.011
Table 23. Experimental conditions of SIP systematic error tests.
Column
Sand (g)
Water (mL)
Sand:Water
% Difference
9
282.7
63.7
4.44
-0.97
2
284.6
65.1
4.37
0.54
18
282.4
63.8
4.43
-0.70
16
282
64.2
4.39
0.07
3
283.1
65.1
4.35
1.06
137
-------
Table 24. Systematic error tests for water quality responses. Percent difference column indicates the largest percent difference from
the calculated average of each column. The column labeled "Accumet" refers to conductivity values measured with an Accumet 4-
electrode specific conductivity probe.
Omin
15min
30min
45 min
60min
90min
Average
Std
Dev
Accumet
%
Difference
Column #2
T(°C)
26.22
26.24
26.34
26.47
26.32
0.11
-0.58
11/16/2009
pH
8.00
8.03
8.04
8.03
8.03
0.02
0.31
-------
Table 25. Measured dielectric constant values from TDR systematic error tests,
readings were made on November 8, 2008.
Column
1st Rdg
2nd Rdg
Average
Std Dev
Percent
Error
1
24.3
24.1
24.1
23.7
23.6
23.6
23.9
0.30
-2.26
2
23.3
23.0
23.3
23.0
22.9
23.2
23.1
0.17
1.09
3
23.7
23.5
23.7
23.0
22.9
22.9
23.3
0.39
0.37
4
23.4
23.4
23.4
23.4
22.8
22.7
23.2
0.34
0.80
Average
Std Dev
23.6
0.37
23.1
0.33
23.4
0.36
0.30
0.09
Table 26. Experimental conditions of TDR systematic error tests.
Column
Sand (g)
Water (mL)
Sand:Water
Percent Error
1
8493
1960
4.33
5.41
2
8753.6
1820
4.81
-4.99
3
8447.6
1900
4.45
2.95
4
8334.8
1760
4.74
-3.37
139
-------
20
18
a 1G
8
i» 14
i*
'€
o
U
B
0>
12
10
5
6
4
2
0
0.1
I t n«*
—I 1 1
10 100 1000
Frequency (Hz)
10000
—~— Column 9
-¦—Column 2
Column 18
100000
Figure 93. Plot of real conductivities measured during systematic error tests on
columns 9, 2, and 18. The columns are listed in the order they were tested, and values
show a stable real conductivity reading in the low and medium ranges of the frequency
spectrum.
25.00
Column 3
Column 16
20.00
15.00
10.G0
5.00
0.00
0.1
1
10
100
1000
10000 100000
Frequency (Hz)
Figure 94. Plot of real conductivities measured during systematic error tests on
columns 9, 2, and 18. The columns are listed in the order they were tested, and values
show a stable real conductivity reading in the low and medium ranges of the frequency
spectrum.
140
-------
04
=
0.01
0.01
¦ Column 9
¦ Column 2
Column 18
100
Frequency (Hz)
10000
Figure 95. Plot of imaginary conductivities measured during systematic error tests of
columns 9, 2, and 18. The columns are listed in the order they were tested. Overall,
measurements show a similar trend and value range among different columns, with
imaginary conductivity increasing with frequency.
Col u m n 3
o
Column 16
0.01
0.001
0.01
1
100
10000
Frequency (Hz)
Figure 96. Plot of imaginary conductivities measured during systematic error tests of
columns 3 and 16. The columns are listed in the order they were tested. Overall,
measurements show a similar trend and value range among different columns, with
imaginary conductivity increasing with frequency.
141
-------
rn
0)
ce.
10
0 -I 1 1 1 1 1 ,
4.34 4.36 4.38 4.40 4.42 4.44 4.46
Sarid/Water
Figure 97. Plot of sand to water ratio and associated real conductivity values. There is
a decrease in real conductivity relative with increasing sand. The largest percent
difference in real conductivity values in these systematic error tests is 1.06, associated
with a water to sand ratio of 4.35.
0.1
F
0.09
O
0.08
=L
>»
0.07
>
0.06
T3
0.05
c
o
0.04
u
>•
&—
0 03
fO
'oxi
0.02
ro
h
0.01
0
4.34 4.36 4.38 4.40 4.42 4.44 4.46
Sarid/Water
Figure 98. Plot of sand to water ratio and measured imaginary conductivity. There
does not appear to be a systematic relationship of the sand to water ratio and imaginary
conductivity within this ratio range.
142
-------
Sand/Water
Figure 99. Plot of sand to water ratio and measured dielectric constant. There may be a
systematic relationship of the sand to water ratio and dielectric constant within this ratio
range. The plot shows a slightly higher dielectric constant response at higher sand to
water ratios. It should be noted that the range of the response is only 0.8.
143
-------
APPENDIX B
RAW DATA
This appendix contains the complete data sets used for each measured parameter.
The first 10 tables outline the data from the SIP measurements. They are separated by
experimental treatment, with the 5 DI solutions first, followed by the 5 tap solutions. The
next tables are the raw data for pH, specific conductivity, dissolved oxygen, and
dielectric constant.
Data that is missing from a table, either due to recording error or simple absence of
data, is denoted by a hyphen (-). Data in parentheses, as seen in specific conductivity,
denotes values measured using a 4-electrode conductivity probe. This data was used as a
quality control check and was not included in analysis.
144
-------
Complete SIP data set for DI control solutions.
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
10
12000
49658.46
265.4281
-14.8008
-0.25832
258.3178
-7.7354
3.2081
10
6000
48487.93
259.1715
-7.17718
-0.12526
125.2633
8.0552
1.6069
10
3000
48344.26
258.4036
-3.667
-0.064
64.00006
-11.1598
0.8250
10
1500
48304.24
258.1897
-2.00926
-0.03507
35.06753
-5.4810
0.4526
10
750
48321.65
258.2827
-1.00801
-0.01759
17.59287
6.8858
0.2270
10
375
48338.47
258.3726
-0.5668
-0.00989
9.892325
10.8838
0.1276
10
187.5
48303.97
258.1882
-0.31384
-0.00548
5.477415
12.2799
0.0707
10
93.75
48372.34
258.5536
-0.18872
-0.00329
3.293748
12.6633
0.0425
1
10
46.875
48347.53
258.421
-0.13451
-0.00235
2.347516
12.7823
0.0303
10
23.4375
48354.05
258.4559
-0.06641
-0.00116
1.159036
12.8687
0.0149
10
11.71875
48366.52
258.5225
-0.07484
-0.00131
1.306183
12.8577
0.0168
10
5.859375
48301.68
258.176
-0.06503
-0.00113
1.134899
12.8838
0.0147
10
2.929687
48207.25
257.6712
-0.07879
-0.00138
1.375035
12.8963
0.0178
10
1.464844
48022.24
256.6824
-0.06475
-0.00113
1.130064
12.9590
0.0147
10
0.732422
47688.52
254.8986
-0.06274
-0.00109
1.094914
13.0514
0.0143
10
0.366211
47146.3
252.0004
-0.06183
-0.00108
1.079084
13.2022
0.0143
10
0.183105
46385.42
247.9334
-0.05489
-0.00096
0.95796
13.4242
0.0129
10
0.091553
45461.08
242.9927
-0.06412
-0.00112
1.119104
13.6896
0.0154
2
11
12000
51352.95
274.4852
-21.3072
-0.37187
371.8737
-9.4122
4.4126
11
6000
50950.16
272.3323
-10.62
-0.18535
185.3508
-4.4897
2.2557
11
3000
50795.93
271.5079
-5.27639
-0.09209
92.08883
6.5630
1.1290
11
1500
50733.07
271.1719
-2.66319
-0.04648
46.48067
-10.9123
0.5711
11
750
50781.27
271.4296
-1.37071
-0.02392
23.92302
2.4408
0.2938
-------
11
375
50905.14
272.0916
11
187.5
50897.29
272.0497
11
93.75
50857.02
271.8344
11
46.875
50854.4
271.8204
11
23.4375
50906.4
272.0984
11
11.71875
50882.3
271.9696
11
5.859375
50840.03
271.7436
11
2.929687
50732.78
271.1704
11
1.464844
50552.96
270.2092
11
0.732422
50209.05
268.371
11
0.366211
49659.02
265.431
11
0.183105
48843.39
261.0715
11
0.091553
47843.18
255.7252
21
12000
47493.33
253.8553
21
6000
-39382
-210.5
21
3000
46634.88
249.2668
21
1500
46703.33
249.6327
21
750
46696.76
249.5975
21
375
46732.95
249.791
21
187.5
46731.63
249.7839
21
93.75
46662.72
249.4156
21
46.875
46686.72
249.5439
21
23.4375
46683.27
249.5254
21
11.71875
46702.68
249.6292
21
5.859375
46668.66
249.4474
21
2.929687
46593.88
249.0477
-0.75728
-0.01322
13.21672
8.9028
0.1619
-0.39078
-0.00682
6.820196
11.3290
0.0836
-0.25405
-0.00443
4.433917
11.8688
0.0544
-0.13973
-0.00244
2.43869
12.1435
0.0299
-0.09341
-0.00163
1.630267
12.1971
0.0200
-0.1084
-0.00189
1.891923
12.1843
0.0232
-0.07489
-0.00131
1.30702
12.2321
0.0160
-0.07138
-0.00125
1.245813
12.2611
0.0153
-0.07136
-0.00125
1.245359
12.3047
0.0154
-0.06
-0.00105
1.047163
12.3983
0.0130
-0.0678
-0.00118
1.183348
12.5293
0.0149
-0.06339
-0.00111
1.106363
12.7423
0.0141
-0.05432
-0.00095
0.947995
13.0156
0.0124
-17.2814
-0.30161
301.6125
0.0348
3.9007
-9.05279
-0.158
157.9984
14.7523
-2.4916
-4.36289
-0.07615
76.14543
-4.5792
1.0173
-2.21759
-0.0387
38.70363
-8.0469
0.5167
-1.21333
-0.02118
21.17616
4.6729
0.2828
-0.67331
-0.01175
11.75135
10.4322
0.1568
-0.44541
-0.00777
7.773723
12.0429
0.1037
-0.2098
-0.00366
3.66157
13.0715
0.0489
-0.15359
-0.00268
2.680589
13.2005
0.0358
-0.08943
-0.00156
1.560804
13.3053
0.0209
-0.10478
-0.00183
1.828673
13.2799
0.0244
-0.06852
-0.0012
1.195862
13.3315
0.0160
-0.0645
-0.00113
1.125736
13.3565
0.0151
-------
21
1.464844
46426.5
248.153
-0.06313
-0.0011
1.101825
13.4058
0.0148
21
0.732422
46138.3
246.6125
-0.05835
-0.00102
1.018313
13.4935
0.0138
21
0.366211
45678.17
244.1531
-0.04551
-0.00079
0.794216
13.6385
0.0108
21
0.183105
45055.39
240.8243
-0.0714
-0.00125
1.246197
13.8061
0.0172
21
0.091553
44198.46
236.2439
-0.04737
-0.00083
0.826714
14.0939
0.0117
Complete SIP data set for 0.025% Steol CS-330, DI solutions.
-j
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
5
12000
40025.37094
213.9384984
-14.409207
-0.2514839
251.4838898
-4.1865
3.8771
5
6000
39512.46114
211.1969585
-7.155721
-0.1248888
124.8887986
10.1467
1.9660
5
3000
39354.18901
210.3509825
-3.599902
-0.0628291
62.82908961
-14.2112
0.9950
5
1500
39458.50659
210.9085675
-1.885441
-0.0329066
32.90660177
-4.8912
0.5200
5
750
39490.49173
211.0795304
-0.97691
-0.01705
17.05001023
8.8369
0.2692
5
375
39575.59904
211.5344351
-0.551163
-0.0096194
9.619447839
13.4244
0.1516
5
187.5
39496.10156
211.1095153
-0.315311
-0.0055031
5.503122883
15.0112
0.0869
5
93.75
39532.90441
211.3062292
-0.169556
-0.0029593
2.959260868
15.5487
0.0467
1
5
46.875
39532.50879
211.3041146
-0.169891
-0.0029651
2.965107623
15.5479
0.0468
5
23.4375
39551.81675
211.407317
-0.080707
-0.0014086
1.408579271
15.7160
0.0222
5
11.71875
39553.76182
211.4177136
-0.081727
-0.0014264
1.426381331
15.7139
0.0225
5
5.859375
39573.42152
211.5227961
-0.069368
-0.0012107
1.210679704
15.7208
0.0191
5
2.929687
39578.70989
211.5510628
-0.05988
-0.0010451
1.04508564
15.7284
0.0165
5
1.464844
39574.6653
211.5294442
-0.061555
-0.0010743
1.074319415
15.7284
0.0169
5
0.732422
39544.60923
211.3687923
-0.060108
-0.0010491
1.049064924
15.7417
0.0165
5
0.366211
39480.17793
211.0244024
-0.057148
-0.0009974
0.997404044
15.7702
0.0158
5
0.183105
39350.22844
210.329813
-0.05586
-0.0009749
0.97492458
15.8234
0.0155
5
0.091553
39134.41607
209.1762803
-0.049038
-0.0008559
0.855860214
15.9164
0.0136
-------
14
12000
42295.96973
226.0750129
-13.963157
-0.243699
243.6989791
2.5527
3.5577
14
6000
41390.1402
221.2332887
-6.926636
-0.1208906
120.8905781
12.0541
1.8170
14
3000
41516.70495
221.9097864
-3.524894
-0.06152
61.51997498
-13.9311
0.9235
14
1500
41558.15759
222.1313537
-1.875861
-0.0327394
32.73940203
-4.5072
0.4912
14
750
41592.6629
222.3157871
-0.981581
-0.0171315
17.13153319
8.3321
0.2569
14
375
41731.67887
223.0588375
-0.612613
-0.0106919
10.69193469
12.2262
0.1598
14
187.5
41561.75053
222.1505582
-0.349233
-0.0060952
6.095163549
14.0991
0.0915
14
93.75
41667.64944
222.7165956
-0.160879
-0.0028078
2.807821187
14.7734
0.0420
2
14
46.875
41583.34325
222.2659729
-0.00865
-0.000151
0.15096845
14.9965
0.0023
14
23.4375
41695.61893
222.8660945
-0.059276
-0.0010345
1.034544028
14.9304
0.0155
14
11.71875
41692.22471
222.8479522
-0.058254
-0.0010167
1.016707062
14.9325
0.0152
14
5.859375
41719.76719
222.9951687
-0.066144
-0.0011544
1.154411232
14.9153
0.0173
14
2.929687
36085.70536
192.8807013
-64.758329
-1.1302271
1130.227116
-6.0180
15.6316
14
1.464844
41723.47561
223.0149905
-0.052105
-0.0009094
0.909388565
14.9264
0.0136
14
0.732422
41690.16248
222.8369294
-0.053594
-0.0009354
0.935376082
14.9371
0.0140
14
0.366211
41613.50243
222.4271759
-0.048958
-0.0008545
0.854463974
14.9682
0.0128
14
0.183105
41465.62819
221.6367774
-0.054085
-0.0009439
0.943945505
15.0176
0.0142
14
0.091553
41207.46254
220.2568634
-0.050813
-0.0008868
0.886839289
15.1143
0.0134
3
24
12000
43380.31157
231.8708984
-17.651809
-0.308077
308.0770225
5.2394
4.3591
24
6000
42837.22658
228.9680699
-8.743363
-0.1525979
152.5979144
-11.3070
2.2129
24
3000
42660.03825
228.0209854
-4.318074
-0.0753633
75.36334552
-5.6161
1.1007
24
1500
42655.70969
227.997849
-2.286889
-0.0399131
39.91307372
-9.5972
0.5834
24
750
42741.99103
228.459029
-1.231307
-0.02149
21.49000107
4.8587
0.3135
24
375
42661.89977
228.0309354
-0.594477
-0.0103754
10.37540708
12.1101
0.1517
24
187.5
42530.78736
227.3301301
-0.250154
-0.0043659
4.365937762
14.2066
0.0640
24
93.75
42664.53447
228.0450181
-0.171615
-0.0029952
2.995196595
14.4023
0.0438
24
46.875
42699.97128
228.2344304
-0.139544
-0.0024355
2.435461432
14.4629
0.0356
24
23.4375
42685.42096
228.1566579
-0.11106
-0.0019383
1.93833018
14.5198
0.0283
-------
24
11.71875
42660.5607
228.023778
-0.091541
-0.0015977
1.597665073
14.5572
0.0234
24
5.859375
42616.60626
227.7888383
-0.080384
-0.0014029
1.402941952
14.5862
0.0205
24
2.929687
42556.73658
227.4688306
-0.063954
-0.0011162
1.116189162
14.6241
0.0164
24
1.464844
42393.9679
226.5988202
-0.066656
-0.0011633
1.163347168
14.6776
0.0171
24
0.732422
42102.20038
225.0393017
-0.050355
-0.0008788
0.878845815
14.7935
0.0130
24
0.366211
41644.54234
222.5930865
-0.060403
-0.0010542
1.054213559
14.9477
0.0158
24
0.183105
40932.58287
218.7876117
-0.048342
-0.0008437
0.843712926
15.2177
0.0129
24
0.091553
40052.03726
214.0810318
-0.053137
-0.0009274
0.927400061
15.5485
0.0144
Complete SIP data set for 0.5% Dowfax 8390, DI solutions.
vo
No.
Run Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
1
1 12000
7639.887112
40.83574838
-2.527112
-0.0441057
44.10568574
-66.6959
3.5991
1 6000
7598.780192
40.61602893
-1.270714
-0.0221778
22.17777144
24.2596
1.8200
1 3000
7593.487768
40.58774054
-0.671071
-0.0117122
11.71220216
64.3180
0.9619
1 1500
7595.313781
40.59750071
-0.358215
-0.0062519
6.251926395
76.8951
0.5133
1 750
7594.166339
40.59136755
-0.186461
-0.0032543
3.254303833
80.6958
0.2672
1 375
7600.310673
40.62420946
-0.110626
-0.0019308
1.930755578
81.5513
0.1584
1 187.5
7597.242229
40.6078084
-0.066879
-0.0011672
1.167239187
81.9025
0.0958
1 93.75
7599.004791
40.61722942
-0.034494
-0.000602
0.602023782
82.0182
0.0494
1 46.875
7599.92798
40.62216394
-0.022426
-0.0003914
0.391400978
82.0364
0.0321
1 23.4375
7600.000953
40.62255398
-0.019469
-0.0003398
0.339792457
82.0407
0.0279
1 11.71875
7600.716536
40.62637883
-0.007965
-0.000139
0.139013145
82.0459
0.0114
1 5.859375
7600.154687
40.6233757
-0.006991
-0.000122
0.122013923
82.0526
0.0100
1 2.929687
7599.057451
40.6175109
-0.006233
-0.0001088
0.108784549
82.0648
0.0089
1 1.464844
7596.722834
40.6050322
-0.004038
-7.048E-05
0.070475214
82.0910
0.0058
1 0.732422
7592.48449
40.58237795
-0.003875
-6.763E-05
0.067630375
82.1368
0.0056
-------
1
0.366211
7586.201407
40.54879441
-0.002846
-4.967E-05
0.049671238
82.2052
0.0041
1
0.183105
7577.113796
40.50022048
-0.003186
-5.561E-05
0.055605258
82.3037
0.0046
1
0.091553
7567.905627
40.45100215
-0.006446
-0.0001125
0.112502038
82.4025
0.0093
16
12000
7245.246363
38.72636508
-2.278996
-0.0397753
39.77531719
-55.9883
3.4227
16
6000
7209.772325
38.53675378
-1.184086
-0.0206659
20.66585296
32.6220
1.7874
16
3000
7191.698624
38.44014855
-0.596765
-0.0104153
10.41533955
71.7269
0.9031
16
1500
7191.563219
38.4394248
-0.329726
-0.0057547
5.754707878
82.0452
0.4990
16
750
7194.124106
38.45311292
-0.180955
-0.0031582
3.158207615
85.2703
0.2738
16
375
7193.653104
38.45059538
-0.100113
-0.0017473
1.747272189
86.2573
0.1515
16
187.5
7197.012059
38.46854924
-0.054896
-0.0009581
0.958099888
86.5203
0.0830
16
93.75
7194.445609
38.45483138
-0.03382
-0.0005903
0.59026046
86.6322
0.0512
9
16
46.875
7196.784047
38.4673305
-0.02244
-0.0003916
0.39164532
86.6318
0.0339
Z
16
23.4375
7195.538424
38.46067255
-0.01484
-0.000259
0.25900252
86.6591
0.0224
16
11.71875
7196.041859
38.46336345
-0.012333
-0.0002152
0.215247849
86.6560
0.0187
16
5.859375
7193.84383
38.45161483
-0.009917
-0.0001731
0.173081401
86.6848
0.0150
16
2.929687
7191.110605
38.43700554
-0.00766
-0.0001337
0.13368998
86.7194
0.0116
16
1.464844
7186.677331
38.41330937
-0.004866
-8.493E-05
0.084926298
86.7745
0.0074
16
0.732422
7179.598933
38.37547482
-0.00539
-9.407E-05
0.09407167
86.8598
0.0082
16
0.366211
7169.646107
38.32227625
-0.008459
-0.0001476
0.147634927
86.9785
0.0128
16
0.183105
7158.790716
38.2642534
-0.007877
-0.0001375
0.137477281
87.1108
0.0120
16
0.091553
7150.022028
38.21738413
-0.010649
-0.0001859
0.185856997
87.2154
0.0162
3
27
12000
7869.620204
42.06368835
-2.566074
-0.0447857
44.78568952
-66.4794
3.5479
27
6000
7826.122546
41.83119023
-1.301528
-0.0227156
22.71556818
21.1984
1.8099
27
3000
7819.781752
41.79729823
-0.676446
-0.011806
11.80601204
62.1892
0.9415
27
1500
7822.816577
41.81351959
-0.37284
-0.0065072
6.50717652
74.2421
0.5187
27
750
7823.031702
41.81466944
-0.197268
-0.0034429
3.442918404
78.1708
0.2745
27
375
7822.077785
41.80957069
-0.105717
-0.0018451
1.845078801
79.2815
0.1471
27
187.5
7824.95386
41.82494352
-0.065102
-0.0011362
1.136225206
79.5284
0.0906
-------
27
93.75
7825.974256
41.83039761
-0.038052
-0.0006641
0.664121556
79.6292
0.0529
27
46.875
7825.474257
41.82772508
-0.023718
-0.000414
0.413950254
79.6695
0.0330
27
23.4375
7824.235303
41.82110278
-0.022137
-0.0003864
0.386357061
79.6850
0.0308
27
11.71875
7824.103166
41.8203965
-0.011444
-0.0001997
0.199732132
79.7007
0.0159
27
5.859375
7822.016861
41.80924505
-0.008322
-0.0001452
0.145243866
79.7244
0.0116
27
2.929687
7818.430979
41.79007825
-0.007962
-0.000139
0.138960786
79.7612
0.0111
27
1.464844
7812.394718
41.757814
-0.006716
-0.0001172
0.117214348
79.8236
0.0094
27
0.732422
7802.664996
41.70580793
-0.006869
-0.0001199
0.119884657
79.9230
0.0096
27
0.366211
7788.470199
41.62993571
-0.006879
-0.0001201
0.120059187
80.0687
0.0096
27
0.183105
7771.760949
41.54062357
-0.007214
-0.0001259
0.125905942
80.2406
0.0101
27
0.091553
7755.67774
41.45465765
-0.008754
-0.0001528
0.152783562
80.4061
0.0123
Complete SIP data set for 5% Dowfax 8390, DI solutions
No.
Run
Freq./Hz
Resistance
Resistivity
Phase/deg
Phase (rad)
phase (mrad)
Real Cond
Imag Cond
1
3
12000
1277.216526
6.826814575
-0.37181
-0.0064892
6.48919993
454.9078
3.1685
3
6000
1276.079507
6.820737126
-0.188641
-0.0032924
3.292351373
480.0361
1.6090
3
3000
1275.840051
6.819457217
-0.100985
-0.0017625
1.762491205
486.3072
0.8615
3
1500
1275.865237
6.819591838
-0.05376
-0.0009383
0.93827328
488.0816
0.4586
3
750
1275.948316
6.820035901
-0.029763
-0.0005195
0.519453639
488.5395
0.2539
3
375
1275.961355
6.820105595
-0.017509
-0.0003056
0.305584577
488.6761
0.1494
3
187.5
1276.021566
6.820427427
-0.009775
-0.0001706
0.170603075
488.7046
0.0834
3
93.75
1276.031672
6.820481445
-0.005805
-0.0001013
0.101314665
488.7158
0.0495
3
46.875
1276.070468
6.820688812
-0.003601
-6.285E-05
0.062848253
488.7060
0.0307
3
23.4375
1276.096632
6.820828661
-0.002955
-5.157E-05
0.051573615
488.6970
0.0252
3
11.71875
1276.133174
6.82102398
-0.002203
-3.845E-05
0.038448959
488.6840
0.0188
3
5.859375
1276.18041
6.82127646
-0.001704
-2.974E-05
0.029739912
488.6664
0.0145
-------
3
2.929687
1276.239026
6.821589767
-0.001117
-1.95E-05
0.019495001
488.6443
0.0095
3
1.464844
1276.390924
6.822401673
-0.000751
-1.311E-05
0.013107203
488.5864
0.0064
3
0.732422
1276.726925
6.824197622
-0.000838
-1.463E-05
0.014625614
488.4577
0.0071
3
0.366211
1277.513063
6.828399587
-0.000295
-5.149E-06
0.005148635
488.1573
0.0025
3
0.183105
1279.386551
6.838413515
-0.002368
-4.133E-05
0.041328704
487.4411
0.0201
3
0.091553
1283.349346
6.859594941
-0.002114
-3.69E-05
0.036895642
485.9363
0.0179
17
12000
1324.042905
7.077104953
-0.409527
-0.0071475
7.147474731
432.0549
3.3664
17
6000
1322.310958
7.067847571
-0.204056
-0.0035614
3.561389368
461.8345
1.6796
17
3000
1321.891026
7.065603004
-0.107826
-0.0018819
1.881887178
469.0293
0.8878
17
1500
1321.813652
7.065189434
-0.056207
-0.000981
0.980980771
471.0517
0.4628
17
750
1321.778925
7.065003816
-0.02888
-0.000504
0.50404264
471.6124
0.2378
17
375
1321.794604
7.065087621
-0.016698
-0.0002914
0.291430194
471.7378
0.1375
17
187.5
1321.761812
7.064912346
-0.010321
-0.0001801
0.180132413
471.7901
0.0850
17
93.75
1321.76937
7.064952744
-0.007146
-0.0001247
0.124719138
471.8005
0.0588
2
17
46.875
1321.761412
7.064910208
-0.003631
-6.337E-05
0.063371843
471.8123
0.0299
17
23.4375
1321.735364
7.064770979
-0.003116
-5.438E-05
0.054383548
471.8224
0.0257
17
11.71875
1321.695825
7.06455964
-0.002983
-5.206E-05
0.052062299
471.8367
0.0246
17
5.859375
1321.611924
7.064111184
-0.002563
-4.473E-05
0.044732039
471.8672
0.0211
17
2.929687
1321.468147
7.063342685
-0.002518
-4.395E-05
0.043946654
471.9186
0.0207
17
1.464844
1321.241795
7.062132817
-0.002374
-4.143E-05
0.041433422
471.9996
0.0196
17
0.732422
1320.944588
7.060544224
-0.002664
-4.649E-05
0.046494792
472.1055
0.0220
17
0.366211
1320.726802
7.059380143
-0.001338
-2.335E-05
0.023352114
472.1846
0.0110
17
0.183105
1320.968624
7.060672699
-0.002482
-4.332E-05
0.043318346
472.0971
0.0205
17
0.091553
1322.054488
7.06647672
-0.001139
-1.988E-05
0.019878967
471.7105
0.0094
3
29
12000
1226.394078
6.55516492
-0.352068
-0.0061446
6.144642804
477.3140
3.1246
29
6000
1225.567459
6.550746582
-0.182915
-0.0031924
3.192415495
500.3591
1.6245
29
3000
1225.377404
6.549730724
-0.095464
-0.0016661
1.666133192
506.6095
0.8479
29
1500
1225.385083
6.549771769
-0.048506
-0.0008466
0.846575218
508.3250
0.4308
-------
29
750
1225.475288
6.550253921
-0.027441
-0.0004789
0.478927773
508.6945
0.2437
29
375
1225.495706
6.550363057
-0.016857
-0.0002942
0.294205221
508.8053
0.1497
29
187.5
1225.388675
6.549790968
-0.008845
-0.0001544
0.154371785
508.9022
0.0786
29
93.75
1225.459014
6.550166935
-0.005694
-9.938E-05
0.099377382
508.8846
0.0506
29
46.875
1225.409279
6.549901098
-0.00349
-6.091E-05
0.06091097
508.9104
0.0310
29
23.4375
1225.40384
6.549872026
-0.003129
-5.461E-05
0.054610437
508.9133
0.0278
29
11.71875
1225.354008
6.549605671
-0.001739
-3.035E-05
0.030350767
508.9357
0.0154
29
5.859375
1225.265013
6.549129986
-0.000979
-1.709E-05
0.017086487
508.9732
0.0087
29
2.929687
1225.117716
6.548342673
-0.000558
-9.739E-06
0.009738774
509.0346
0.0050
29
1.464844
1224.896958
6.547162705
-0.000759
-1.325E-05
0.013246827
509.1262
0.0067
29
0.732422
1224.542734
6.545269352
-0.000655
-1.143E-05
0.011431715
509.2736
0.0058
29
0.366211
1224.160636
6.543227011
-0.001331
-2.323E-05
0.023229943
509.4322
0.0118
29
0.183105
1224.050536
6.542638519
-0.001638
-2.859E-05
0.028588014
509.4778
0.0146
29
0.091553
1224.520134
6.545148554
-0.001959
-3.419E-05
0.034190427
509.2821
0.0174
Complete SIP data for 8% Aerosol MA 80-1, DI solutions
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
1
6
12000
356.508799
1.905565279
-0.167806
-0.0029287
2.928718118
1724.6913
5.1231
6
6000
355.780798
1.901674061
-0.071864
-0.0012542
1.254242392
1748.3173
2.1985
6
3000
355.603374
1.900725716
-0.034661
-0.0006049
0.604938433
1752.6628
1.0609
6
1500
355.564279
1.900516751
-0.01852
-0.0003232
0.32322956
1753.6082
0.5669
6
750
355.557285
1.900479367
-0.009591
-0.0001674
0.167391723
1753.8628
0.2936
6
375
355.559099
1.900489063
-0.00555
-9.686E-05
0.09686415
1753.9075
0.1699
6
187.5
355.558811
1.900487524
-0.003594
-6.273E-05
0.062726082
1753.9246
0.1100
6
93.75
355.566908
1.900530803
-0.002473
-4.316E-05
0.043161269
1753.8906
0.0757
-------
6
46.875
355.565387
1.900522673
-0.00189
-3.299E-05
0.03298617
1753.9003
0.0579
6
23.4375
355.570177
1.900548276
-0.001404
-2.45E-05
0.024504012
1753.8781
0.0430
6
11.71875
355.575713
1.900577866
-0.001117
-1.95E-05
0.019495001
1753.8515
0.0342
6
5.859375
355.575713
1.900577866
-0.001117
-1.95E-05
0.019495001
1753.8515
0.0342
6
2.929687
355.589228
1.900650105
-0.000401
-6.999E-06
0.006998653
1753.7857
0.0123
6
1.464844
355.618273
1.900805353
0.000018
3.142E-07
-0.00031415
1753.6426
-0.0006
6
0.732422
355.680751
1.901139302
-0.00013
-2.269E-06
0.00226889
1753.3346
0.0040
6
0.366211
355.889993
1.902257716
0.000584
1.019E-05
-0.01019255
1752.3035
-0.0179
6
0.183105
356.281362
1.904349611
-0.012766
-0.0002228
0.222804998
1750.2362
0.3900
6
0.091553
357.108893
1.908772824
-0.00708
-0.0001236
0.12356724
1746.2789
0.2158
19
12000
353.163203
1.887682826
-0.054936
-0.0009588
0.958798008
1763.1694
1.6931
19
6000
353.083449
1.887256535
-0.027639
-0.0004824
0.482383467
1765.5577
0.8520
19
3000
353.070709
1.887188439
-0.015102
-0.0002636
0.263575206
1766.0946
0.4656
19
1500
353.064384
1.887154632
-0.009142
-0.0001596
0.159555326
1766.2538
0.2818
19
750
353.075299
1.887212973
-0.004937
-8.617E-05
0.086165461
1766.2515
0.1522
19
375
353.063744
1.887151211
-0.001972
-3.442E-05
0.034417316
1766.3274
0.0608
19
187.5
353.068926
1.887178909
-0.001784
-3.114E-05
0.031136152
1766.3021
0.0550
19
93.75
353.061247
1.887137864
-0.000685
-1.196E-05
0.011955305
1766.3429
0.0211
2
19
46.875
353.057169
1.887116067
-0.000718
-1.253E-05
0.012531254
1766.3633
0.0221
19
23.4375
353.052882
1.887093153
-0.000687
-1.199E-05
0.011990211
1766.3847
0.0212
19
11.71875
353.05356
1.887096777
-0.000468
-8.168E-06
0.008168004
1766.3816
0.0144
19
5.859375
353.040662
1.887027836
-0.000381
-6.65E-06
0.006649593
1766.4462
0.0117
19
2.929687
353.033696
1.886990602
0.000462
8.063E-06
-0.00806329
1766.4810
-0.0142
19
1.464844
353.013131
1.886880681
-0.000576
-1.005E-05
0.010052928
1766.5838
0.0178
19
0.732422
353.01351
1.886882706
-0.000619
-1.08E-05
0.010803407
1766.5818
0.0191
19
0.366211
353.053517
1.887096547
-0.001398
-2.44E-05
0.024399294
1766.3803
0.0431
19
0.183105
353.198671
1.887872405
-0.006567
-0.0001146
0.114613851
1765.6180
0.2024
19
0.091553
353.969055
1.891990163
-0.00061
-1.065E-05
0.01064633
1761.8129
0.0188
-------
26
12000
336.064882
1.796291066
-0.155709
-0.0027176
2.717589177
1833.2252
5.0430
26
6000
335.398887
1.792731274
-0.066851
-0.0011668
1.166750503
1855.2070
2.1694
26
3000
335.246355
1.79191598
-0.032872
-0.0005737
0.573715016
1859.2013
1.0672
26
1500
335.204659
1.791693112
-0.017879
-0.000312
0.312042187
1860.1403
0.5805
26
750
335.186377
1.791595393
-0.009525
-0.0001662
0.166239825
1860.4547
0.3093
26
375
335.187803
1.791603015
-0.005592
-9.76E-05
0.097597176
1860.5021
0.1816
26
187.5
335.18681
1.791597707
-0.002949
-5.147E-05
0.051468897
1860.5286
0.0958
26
93.75
335.180092
1.791561799
-0.002192
-3.826E-05
0.038256976
1860.5695
0.0712
26
46.875
335.176803
1.791544219
-0.001317
-2.299E-05
0.022985601
1860.5907
0.0428
26
23.4375
335.17256
1.79152154
-0.001278
-2.23E-05
0.022304934
1860.6143
0.0415
26
11.71875
335.17256
1.79152154
-0.001278
-2.23E-05
0.022304934
1860.6143
0.0415
3
26
5.859375
335.114413
1.79121074
-0.000774
-1.351E-05
0.013508622
1860.9381
0.0251
26
2.929687
335.101705
1.791142815
-0.000333
-5.812E-06
0.005811849
1861.0091
0.0108
26
1.464844
335.092477
1.791093491
0.000431
7.522E-06
-0.00752224
1861.0603
-0.0140
26
0.732422
335.105188
1.791161432
0.000211
3.683E-06
-0.00368258
1860.9899
-0.0069
26
0.366211
335.14925
1.791396946
-0.000689
-1.203E-05
0.012025117
1860.7448
0.0224
26
0.183105
335.411573
1.792799082
0.008598
0.0001501
-0.15006089
1859.2212
-0.2790
26
0.091553
335.943615
1.795642885
-0.001571
-2.742E-05
0.027418663
1856.3431
0.0509
Complete SIP data set for Tap control solutions
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
4
12000
2891.281478
15.45410831
-0.787551
-0.013745128
13.7451276
152.1888
2.9646
4
6000
2885.834894
15.42499593
-0.401679
-0.007010504
7.010503587
198.8992
1.5150
4
3000
2884.380968
15.41722459
-0.203781
-0.00355659
3.556589793
211.7347
0.7690
4
1500
2884.05171
15.41546468
-0.105947
-0.001849093
1.849092991
215.0206
0.3998
4
750
2884.023099
15.41531175
-0.0565
-0.000986095
0.9860945
215.8902
0.2132
-------
4
375
2884.504822
15.4178866
-0.030651
-0.000534952
0.534951903
216.0976
0.1157
4
187.5
2883.744756
15.41382399
-0.022232
-0.000388015
0.388015096
216.2027
0.0839
1
4
93.75
2884.046601
15.41543737
-0.011616
-0.000202734
0.202734048
216.2189
0.0438
4
46.875
2883.790764
15.41406991
-0.008437
-0.000147251
0.147250961
216.2449
0.0318
4
23.4375
2883.789697
15.4140642
-0.007327
-0.000127878
0.127878131
216.2469
0.0277
4
11.71875
2883.632612
15.41322457
-0.004509
-7.86956E-05
0.078695577
216.2623
0.0170
4
5.859375
2883.211367
15.41097299
-0.003882
-6.77525E-05
0.067752546
216.2945
0.0147
4
2.929687
2882.491212
15.40712371
-0.003141
-5.48199E-05
0.054819873
216.3491
0.0119
4
1.464844
2881.162912
15.40002385
-0.003028
-5.28477E-05
0.052847684
216.4489
0.0114
4
0.732422
2879.089924
15.38894358
-0.002608
-4.55174E-05
0.045517424
216.6050
0.0099
4
0.366211
2876.121659
15.37307799
-0.002523
-4.40339E-05
0.044033919
216.8286
0.0095
4
0.183105
2872.665056
15.35460219
-0.004328
-7.55366E-05
0.075536584
217.0881
0.0164
4
0.091553
2869.805054
15.33931528
-0.00365
-6.37035E-05
0.06370345
217.3051
0.0138
18
12000
2818.497459
15.06507248
-1.028465
-0.0179498
17.94979965
114.2011
3.9714
18
6000
2805.843888
14.99743823
-0.491952
-0.008586038
8.586038256
195.9029
1.9083
18
3000
2803.577778
14.9853257
-0.253529
-0.004424842
4.424841637
215.3292
0.9843
18
1500
2803.011664
14.98229978
-0.131879
-0.002301684
2.301684187
220.5528
0.5121
18
750
2802.941067
14.98192244
-0.066046
-0.001152701
1.152700838
222.0053
0.2565
18
375
2803.151233
14.98304579
-0.038469
-0.000671399
0.671399457
222.3091
0.1494
18
187.5
2803.106071
14.9828044
-0.02119
-0.000369829
0.36982907
222.4273
0.0823
18
93.75
2803.018155
14.98233448
-0.012266
-0.000214078
0.214078498
222.4675
0.0476
18
46.875
2803.023205
14.98236147
-0.008167
-0.000142539
0.142538651
222.4764
0.0317
Z
18
23.4375
2802.894605
14.9816741
-0.006075
-0.000106027
0.106026975
222.4899
0.0236
18
11.71875
2802.633687
14.98027947
-0.003939
-6.87474E-05
0.068747367
222.5130
0.0153
18
5.859375
2802.218497
14.97806025
-0.003464
-6.04572E-05
0.060457192
222.5464
0.0135
18
2.929687
2801.514931
14.97429964
-0.003023
-5.27604E-05
0.052760419
222.6026
0.0117
18
1.464844
2800.248378
14.96752982
-0.001947
-3.3981E-05
0.033980991
222.7039
0.0076
18
0.732422
2798.312344
14.95718158
-0.001918
-3.34749E-05
0.033474854
222.8580
0.0075
18
0.366211
2795.533743
14.94232976
-0.002733
-4.7699E-05
0.047699049
223.0791
0.0106
18
0.183105
2792.2921
14.92500294
-0.002457
-4.2882E-05
0.042882021
223.3382
0.0096
18
0.091553
2789.013481
14.90747848
-0.002646
-4.61806E-05
0.046180638
223.6006
0.0103
25
12000
2817.90434
15.06190221
-0.792374
-0.013829303
13.82930342
155.3936
3.0605
25
6000
2811.736926
15.02893694
-0.400048
-0.006982038
6.982037744
204.2820
1.5486
-------
25
3000
2811.060373
15.02532071
-0.204634
-0.003571477
3.571477202
217.2190
0.7923
25
1500
2810.275908
15.02112769
-0.110657
-0.001931297
1.931296621
220.5524
0.4286
25
750
2810.564929
15.02267253
-0.060171
-0.001050164
1.050164463
221.4853
0.2330
25
375
2810.472113
15.02217642
-0.031918
-0.000557065
0.557064854
221.7811
0.1236
25
187.5
2810.46535
15.02214027
-0.021971
-0.00038346
0.383459863
221.8411
0.0851
3
25
93.75
2810.416105
15.02187706
-0.01302
-0.000227238
0.22723806
221.8798
0.0504
25
46.875
2810.382048
15.02169502
-0.008581
-0.000149764
0.149764193
221.8931
0.0332
25
23.4375
2810.176915
15.02059857
-0.004566
-7.96904E-05
0.079690398
221.9152
0.0177
25
11.71875
2810.011646
15.01971519
-0.005489
-9.57995E-05
0.095799517
221.9272
0.0213
25
5.859375
2809.716769
15.01813905
-0.006486
-0.0001132
0.113200158
221.9492
0.0251
25
2.929687
2808.899291
15.01376958
-0.002566
-4.47844E-05
0.044784398
222.0177
0.0099
25
1.464844
2807.754771
15.00765203
-0.002266
-3.95485E-05
0.039548498
222.1083
0.0088
25
0.732422
2805.704289
14.99669206
-0.00241
-4.20617E-05
0.04206173
222.2706
0.0093
25
0.366211
2802.553308
14.97984984
-0.00265
-4.62505E-05
0.04625045
222.5204
0.0103
25
0.183105
2740.032874
14.6456736
-1.119485
-0.019538372
19.53837171
99.2662
4.4466
25
0.091553
2751.038291
14.70449835
-1.623632
-0.028337249
28.3372493
-11.9716
6.4229
Complete SIP data set for 0.025% Steol CS-330, tap solutions
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
1
9
12000
2707.81901
14.47348817
-0.721084
-0.012585079
12.58507905
172.9808
2.8983
9
6000
2702.46101
14.44484928
-0.360809
-0.006297199
6.297199477
215.9043
1.4531
9
3000
2701.330853
14.43880851
-0.183392
-0.003200741
3.200740576
226.9880
0.7389
9
1500
2701.042567
14.4372676
-0.099261
-0.001732402
1.732402233
229.7475
0.4000
9
750
2701.06799
14.43740348
-0.053679
-0.00093686
0.936859587
230.5492
0.2163
9
375
2701.526808
14.4398559
-0.033797
-0.000589859
0.589859041
230.7107
0.1362
9
187.5
2701.300037
14.43864379
-0.018388
-0.000320926
0.320925764
230.8229
0.0741
9
93.75
2701.480769
14.43960982
-0.014119
-0.000246419
0.246418907
230.8235
0.0569
9
46.875
2701.581336
14.44014736
-0.007543
-0.000131648
0.131647979
230.8313
0.0304
9
23.4375
2701.436865
14.43937515
-0.005174
-9.03018E-05
0.090301822
230.8472
0.0208
9
11.71875
2701.319492
14.43874778
-0.004747
-8.28494E-05
0.082849391
230.8577
0.0191
-------
9
5.859375
2701.225297
14.4382443
-0.00384
-6.70195E-05
0.06701952
230.8666
0.0155
9
2.929687
2700.872543
14.43635881
-0.002986
-5.21147E-05
0.052114658
230.8975
0.0120
9
1.464844
2700.22942
14.43292127
-0.002478
-4.32485E-05
0.043248534
230.9528
0.0100
9
0.732422
2699.284581
14.42787103
-0.002884
-5.03345E-05
0.050334452
231.0334
0.0116
9
0.366211
2698.036805
14.42120158
-0.002811
-4.90604E-05
0.049060383
231.1403
0.0113
9
0.183105
2697.145899
14.41643962
-0.002769
-4.83274E-05
0.048327357
231.2166
0.0112
9
0.091553
2697.051287
14.41593392
-0.004225
-7.37389E-05
0.073738925
231.2236
0.0171
15
12000
2655.757132
14.19521368
-1.061747
-0.01853067
18.53067039
114.4394
4.3511
15
6000
2649.169457
14.16000208
-0.533681
-0.009314334
9.314334493
202.6695
2.1926
15
3000
2647.893262
14.15318072
-0.277776
-0.004848025
4.848024528
226.4904
1.1418
15
1500
2647.894767
14.15318877
-0.142543
-0.002487803
2.487802979
233.1295
0.5859
15
750
2647.952273
14.15349614
-0.079725
-0.00139144
1.391440425
234.7650
0.3277
15
375
2648.253775
14.15510769
-0.040642
-0.000709325
0.709324826
235.2918
0.1670
15
187.5
2648.364879
14.15570155
-0.026692
-0.000465855
0.465855476
235.3925
0.1097
15
93.75
2648.359396
14.15567224
-0.014737
-0.000257205
0.257204861
235.4513
0.0606
9
15
46.875
2648.325511
14.15549112
-0.009547
-0.000166624
0.166623791
235.4691
0.0392
Z
15
23.4375
2648.449935
14.15615618
-0.006196
-0.000108139
0.108138788
235.4643
0.0255
15
11.71875
2648.359349
14.15567199
-0.005397
-9.41938E-05
0.094193841
235.4734
0.0222
15
5.859375
2648.294999
14.15532804
-0.00373
-6.50997E-05
0.06509969
235.4810
0.0153
15
2.929687
2648.12426
14.15441542
-0.003142
-5.48373E-05
0.054837326
235.4966
0.0129
15
1.464844
2647.780794
14.15257957
-0.003224
-5.62685E-05
0.056268472
235.5271
0.0133
15
0.732422
2647.249059
14.14973741
-0.002729
-4.76292E-05
0.047629237
235.5748
0.0112
15
0.366211
2646.748944
14.14706426
-0.002523
-4.40339E-05
0.044033919
235.6194
0.0104
15
0.183105
2646.826171
14.14747704
-0.002844
-4.96363E-05
0.049636332
235.6123
0.0117
15
0.091553
2646.902987
14.14788763
-0.003917
-6.83634E-05
0.068363401
235.6046
0.0161
3
23
12000
2646.393966
14.14516688
-0.735095
-0.012829613
12.82961304
174.7987
3.0232
23
6000
2641.412361
14.11853984
-0.368904
-0.006438482
6.438481512
220.2124
1.5201
23
3000
2640.578114
14.11408073
-0.192964
-0.003367801
3.367800692
231.7875
0.7954
23
1500
2640.222692
14.11218097
-0.098531
-0.001719662
1.719661543
235.0569
0.4062
23
750
2640.669596
14.11456971
-0.052189
-0.000910855
0.910854617
235.8410
0.2151
23
375
2640.654591
14.1144895
-0.030935
-0.000539909
0.539908555
236.0509
0.1275
23
187.5
2640.572238
14.11404932
-0.017424
-0.000304101
0.304101072
236.1354
0.0718
23
93.75
2640.673949
14.11459297
-0.010406
-0.000181616
0.181615918
236.1494
0.0429
-------
23
46.875
2640.782717
14.11517435
-0.006744
-0.000117703
0.117703032
236.1471
0.0278
23
23.4375
2640.66881
14.1145655
-0.005838
-0.000101891
0.101890614
236.1586
0.0241
23
11.71875
2640.619819
14.11430364
-0.004513
-7.87654E-05
0.078765389
236.1646
0.0186
23
5.859375
2640.457919
14.11343828
-0.00625
-0.000109081
0.10908125
236.1769
0.0258
23
2.929687
2639.945213
14.11069783
-0.003205
-5.59369E-05
0.055936865
236.2262
0.0132
23
1.464844
2639.37898
14.10767127
-0.002837
-4.95142E-05
0.049514161
236.2771
0.0117
23
0.732422
2638.289344
14.10184709
-0.002914
-5.0858E-05
0.050858042
236.3747
0.0120
23
0.366211
2636.721286
14.0934657
-0.002723
-4.75245E-05
0.047524519
236.5154
0.0112
23
0.183105
2634.909724
14.08378277
-0.003334
-5.81883E-05
0.058188302
236.6775
0.0138
23
0.091553
2634.234099
14.08017151
-0.003196
-5.57798E-05
0.055779788
236.7383
0.0132
Complete SIP data set for 0.5% Dowfax 8390, tap solutions.
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(|iiS/cm)
Imag Cond
(|LiS/cm)
1
8
12000
2288.109024
12.23010799
-0.61825
-0.010790317
10.79031725
222.1005
2.9409
8
6000
2283.839048
12.20728466
-0.307929
-0.005374285
5.374284837
260.2171
1.4675
8
3000
2282.923435
12.20239064
-0.159419
-0.00278234
2.782339807
269.7066
0.7601
8
1500
2282.764563
12.20154146
-0.085565
-0.001493366
1.493365945
272.1901
0.4080
8
750
2282.872488
12.20211832
-0.046714
-0.000815299
0.815299442
272.8786
0.2227
8
375
2282.844626
12.2019694
-0.028063
-0.000489784
0.489783539
273.0724
0.1338
8
187.5
2283.279462
12.20429363
-0.018576
-0.000324207
0.324206928
273.0808
0.0885
8
93.75
2282.979311
12.2026893
-0.013407
-0.000233992
0.233992371
273.1393
0.0639
8
46.875
2283.002303
12.20281219
-0.010092
-0.000176136
0.176135676
273.1472
0.0481
8
23.4375
2282.92974
12.20242434
-0.006328
-0.000110443
0.110442584
273.1643
0.0302
8
11.71875
2282.900985
12.20227064
-0.005797
-0.000101175
0.101175041
273.1686
0.0276
8
5.859375
2282.669838
12.20103514
-0.004564
-7.96555E-05
0.079655492
273.1980
0.0218
8
2.929687
2282.256214
12.19882429
-0.00426
-7.43498E-05
0.07434978
273.2479
0.0203
8
1.464844
2281.584145
12.19523204
-0.003803
-6.63738E-05
0.066373759
273.3289
0.0181
8
0.732422
2280.528035
12.18958705
-0.00324
-5.65477E-05
0.05654772
273.4560
0.0155
8
0.366211
2279.247644
12.18274327
-0.003983
-6.95153E-05
0.069515299
273.6089
0.0190
-------
On
O
8
0.183105
2278.077859
12.17649068
-0.003387
-5.91133E-05
0.059113311
273.7500
0.0162
8
0.091553
2277.409157
12.17291642
-0.004529
-7.90446E-05
0.079044637
273.8291
0.0216
20
12000
2208.603473
11.80514507
-0.797573
-0.013920042
13.92004157
197.2150
3.9304
20
6000
2200.850397
11.76370432
-0.387546
-0.00676384
6.763840338
262.3435
1.9166
20
3000
2199.100764
11.75435241
-0.19974
-0.003486062
3.48606222
277.9448
0.9886
20
1500
2198.918238
11.75337679
-0.104334
-0.001820941
1.820941302
282.0642
0.5164
20
750
2198.964501
11.75362407
-0.060261
-0.001051735
1.051735233
283.0857
0.2983
20
375
2198.888307
11.75321681
-0.032399
-0.00056546
0.565459747
283.4615
0.1604
20
187.5
2198.891691
11.7532349
-0.013453
-0.000234795
0.234795209
283.5842
0.0666
20
93.75
2198.500586
11.75114441
-0.012383
-0.00021612
0.216120499
283.6386
0.0613
9
20
46.875
2198.50436
11.75116459
-0.009992
-0.00017439
0.174390376
283.6457
0.0495
Z
20
23.4375
2198.50436
11.75116459
-0.009992
-0.00017439
0.174390376
283.6457
0.0495
20
11.71875
2198.1726
11.7493913
-0.005966
-0.000104125
0.104124598
283.6976
0.0295
20
5.859375
2197.994528
11.7484395
-0.004897
-8.54673E-05
0.085467341
283.7222
0.0242
20
2.929687
2197.644709
11.74656969
-0.00319
-5.56751E-05
0.05567507
283.7693
0.0158
20
1.464844
2197.049261
11.74338698
-0.002726
-4.75769E-05
0.047576878
283.8466
0.0135
20
0.732422
2196.081768
11.73821566
-0.002364
-4.12589E-05
0.041258892
283.9720
0.0117
20
0.366211
2194.826954
11.73150858
-0.002506
-4.37372E-05
0.043737218
284.1342
0.0124
20
0.183105
2193.683774
11.7253982
-0.002159
-3.7681E-05
0.037681027
284.2825
0.0107
20
0.091553
2193.153974
11.72256639
-0.003116
-5.43835E-05
0.054383548
284.3505
0.0155
3
30
12000
2355.153735
12.58846681
-0.896254
-0.015642321
15.64232106
165.3736
4.1418
30
6000
2344.158689
12.52969749
-0.417608
-0.007288512
7.288512424
243.1721
1.9390
30
3000
2341.849507
12.51735475
-0.223838
-0.003906645
3.906644614
259.6536
1.0403
30
1500
2341.240184
12.51409787
-0.113847
-0.001986972
1.986971691
264.6419
0.5293
30
750
2341.221238
12.51399661
-0.065411
-0.001141618
1.141618183
265.7988
0.3041
30
375
2341.022135
12.51293239
-0.037593
-0.000656111
0.656110629
266.2028
0.1748
30
187.5
2341.330199
12.51457901
-0.020492
-0.000357647
0.357646876
266.3001
0.0953
30
93.75
2341.338245
12.51462202
-0.013532
-0.000236174
0.236173996
266.3307
0.0629
30
46.875
2341.441172
12.51517217
-0.009529
-0.00016631
0.166309637
266.3313
0.0443
30
23.4375
2341.369549
12.51478934
-0.008022
-0.000140008
0.140007966
266.3430
0.0373
30
11.71875
2341.218622
12.51398262
-0.006832
-0.000119239
0.119238896
266.3625
0.0318
30
5.859375
2340.976232
12.51268703
-0.005023
-8.76664E-05
0.087666419
266.3929
0.0234
30
2.929687
2340.546657
12.51039092
-0.004605
-8.03711E-05
0.080371065
266.4424
0.0214
-------
30
1.464844
2339.750192
12.50613376
-0.004436
-7.74215E-05
0.077421508
266.5333
0.0206
30
0.732422
2338.51748
12.49954482
-0.003895
-6.79794E-05
0.067979435
266.6744
0.0181
30
0.366211
2336.995981
12.4914123
-0.003699
-6.45586E-05
0.064558647
266.8482
0.0172
30
0.183105
2335.330437
12.48250985
-0.003848
-6.71591E-05
0.067159144
267.0383
0.0179
30
0.091553
2334.268836
12.47683551
-0.004677
-8.16277E-05
0.081627681
267.1588
0.0218
Complete SIP data set for 5% Dowfax 8390, tap solutions
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
Phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
2
12000
1010.731534
5.402433047
-0.266096
-0.004644173
4.644173488
595.2903
2.8655
2
6000
1010.043436
5.398755113
-0.136782
-0.002387256
2.387256246
611.6595
1.4740
2
3000
1009.982497
5.39842939
-0.07003
-0.001222234
1.22223359
615.9501
0.7547
2
1500
1009.971475
5.398370476
-0.036974
-0.000645307
0.645307222
617.0483
0.3985
2
750
1010.006952
5.398560103
-0.021631
-0.000377526
0.377525843
617.3041
0.2331
2
375
1010.068268
5.398887842
-0.012308
-0.000214812
0.214811524
617.3643
0.1326
2
187.5
1010.123316
5.399182077
-0.007118
-0.00012423
0.124230454
617.3618
0.0767
2
93.75
1010.155394
5.399353537
-0.00497
-8.67414E-05
0.08674141
617.3502
0.0536
1
2
46.875
1010.20196
5.399602435
-0.003485
-6.08237E-05
0.060823705
617.3257
0.0375
2
23.4375
1010.221606
5.399707445
-0.00289
-5.04392E-05
0.05043917
617.3148
0.0311
2
11.71875
1010.250517
5.399861976
-0.002233
-3.89725E-05
0.038972549
617.2982
0.0241
2
5.859375
1010.251601
5.39986777
-0.002076
-3.62324E-05
0.036232428
617.2977
0.0224
2
2.929687
1010.480961
5.401093716
-0.001077
-1.87969E-05
0.018796881
617.1586
0.0116
2
1.464844
1010.588776
5.401669995
-0.000491
-8.56942E-06
0.008569423
617.0930
0.0053
2
0.732422
1010.798373
5.402790306
-0.000543
-9.47698E-06
0.009476979
616.9651
0.0058
2
0.366211
1011.19174
5.404892881
-0.00045
-7.85385E-06
0.00785385
616.7251
0.0048
2
0.183105
1011.951025
5.408951314
-0.000681
-1.18855E-05
0.011885493
616.2623
0.0073
2
0.091553
1013.324247
5.416291285
-0.001514
-2.64238E-05
0.026423842
615.4266
0.0163
2
13
12000
847.906832
4.532123255
-1.158082
-0.020212005
20.21200515
295.0032
14.8647
13
6000
848.678372
4.536247192
-1.466204
-0.025589658
25.58965841
76.7167
18.8018
13
3000
859.208573
4.592531877
-0.385242
-0.006723629
6.723628626
672.6192
4.8801
-------
On
to
13
1500
854.993265
4.570000751
-0.754454
-0.013167486
13.16748566
531.4701
9.6040
13
750
848.716173
4.536449241
-2.005046
-0.034994068
34.99406784
-309.1477
25.7080
13
375
814.434123
4.353209208
2.610395
0.045559224
-45.5592239
-660.2034
-34.8735
13
187.5
744.837845
3.981212075
0.695197
0.012133273
-12.1332732
642.9596
-10.1585
13
93.75
745.908489
3.986934745
1.989428
0.034721487
-34.7214869
-339.8690
-29.0236
13
46.875
810.125976
4.330181851
-0.252703
-0.004410425
4.410425459
745.3422
3.3951
13
23.4375
889.755071
4.755805115
0.213954
0.003734139
-3.73413916
684.9166
-2.6172
13
11.71875
905.137531
4.838025474
0.00336
5.86421E-05
-0.05864208
688.9824
-0.0404
13
5.859375
890.956806
4.762228475
-0.837165
-0.014611041
14.61104075
468.6680
10.2267
13
2.929687
890.958611
4.762238123
-0.837253
-0.014612577
14.61257661
468.6213
10.2277
13
1.464844
905.070065
4.837664864
-0.005247
-9.15759E-05
0.091575891
689.0282
0.0631
13
0.732422
905.233193
4.838536795
0.000316
5.51515E-06
-0.00551515
688.9135
-0.0038
13
0.366211
905.390979
4.839380172
-0.001024
-1.78719E-05
0.017871872
688.7931
0.0123
13
0.183105
906.363791
4.844579923
-0.001215
-2.12054E-05
0.021205395
688.0536
0.0146
13
0.091553
908.22894
4.854549279
-0.001602
-2.79597E-05
0.027959706
686.6403
0.0192
28
12000
954.822343
5.103594383
-0.419528
-0.007322022
7.322022184
596.4955
4.7822
28
6000
952.426943
5.090790797
-0.200917
-0.003506604
3.506604401
641.6056
2.2960
28
3000
951.881634
5.087876081
-0.101099
-0.001764481
1.764480847
651.8069
1.1560
28
1500
951.774205
5.087301865
-0.052562
-0.000917365
0.917364586
654.3213
0.6011
28
750
951.734478
5.087089521
-0.028894
-0.000504287
0.504286982
654.9800
0.3304
28
375
951.755359
5.087201132
-0.016062
-0.00028033
0.280330086
655.1546
0.1837
28
187.5
951.792652
5.087400466
-0.008986
-0.000156833
0.156832658
655.1870
0.1028
28
93.75
951.780014
5.087332914
-0.005175
-9.03193E-05
0.090319275
655.2134
0.0592
3
28
46.875
951.76583
5.0872571
-0.00418
-7.29535E-05
0.07295354
655.2262
0.0478
28
23.4375
951.778371
5.087324133
-0.002439
-4.25679E-05
0.042567867
655.2214
0.0279
28
11.71875
951.765044
5.087252899
-0.001968
-3.43475E-05
0.034347504
655.2312
0.0225
28
5.859375
951.725468
5.087041362
-0.001682
-2.93559E-05
0.029355946
655.2588
0.0192
28
2.929687
951.666697
5.086727227
-0.001509
-2.63366E-05
0.026336577
655.2994
0.0173
28
1.464844
951.572923
5.086225998
-0.001198
-2.09087E-05
0.020908694
655.3643
0.0137
28
0.732422
951.572923
5.086225998
-0.001198
-2.09087E-05
0.020908694
655.3643
0.0137
28
0.366211
951.302329
5.084779654
-0.001524
-2.65984E-05
0.026598372
655.5504
0.0174
28
0.183105
951.351303
5.085041423
-0.001313
-2.29158E-05
0.022915789
655.5169
0.0150
28
0.091553
951.659994
5.086691399
-0.001549
-2.70347E-05
0.027034697
655.3040
0.0177
-------
Complete SIP data set for 8% Aerosol MA 80-1, tap solutions
No.
Run
Freq./Hz
Resistance
(ohm)
Resistivity
(ohm-m)
Phase (deg)
Phase (rad)
phase (mrad)
Real Cond
(jiS/cm)
Imag Cond
(jiS/cm)
7
12000
327.641146
1.751265588
-0.154046
-0.002688565
2.688564838
1880.8462
5.1174
7
6000
326.833375
1.746947994
-0.05824
-0.001016463
1.01646272
1904.8545
1.9395
7
3000
326.610433
1.745756353
-0.02415
-0.00042149
0.42148995
1908.8353
0.8048
7
1500
326.539153
1.745375356
-0.010891
-0.000190081
0.190080623
1909.6956
0.3630
7
750
326.518751
1.745266306
-0.005897
-0.00010292
0.102920341
1909.8950
0.1966
7
375
326.500552
1.745169031
-0.002601
-4.53953E-05
0.045395253
1910.0282
0.0867
7
187.5
326.497825
1.745154455
-0.001434
-2.50276E-05
0.025027602
1910.0486
0.0478
7
93.75
326.486789
1.745095467
-0.001352
-2.35965E-05
0.023596456
1910.1134
0.0451
1
7
46.875
326.476696
1.745041519
-0.000644
-1.12397E-05
0.011239732
1910.1738
0.0215
7
23.4375
326.467827
1.744994114
-0.000449
-7.8364E-06
0.007836397
1910.2259
0.0150
7
11.71875
326.456297
1.744932485
-0.000429
-7.48734E-06
0.007487337
1910.2934
0.0143
7
5.859375
326.435853
1.74482321
-0.000143
-2.49578E-06
0.002495779
1910.4132
0.0048
7
2.929687
326.410148
1.744685815
-0.000371
-6.47506E-06
0.006475063
1910.5635
0.0124
7
1.464844
326.376522
1.744506082
-0.000528
-9.21518E-06
0.009215184
1910.7602
0.0176
7
0.732422
326.370189
1.744472231
-0.000491
-8.56942E-06
0.008569423
1910.7974
0.0164
7
0.366211
326.534219
1.745348984
-0.000716
-1.24963E-05
0.012496348
1909.8372
0.0239
7
0.183105
326.942719
1.747532446
-0.004079
-7.11908E-05
0.071190787
1907.4356
0.1358
7
0.091553
327.854288
1.752404848
-0.002557
-4.46273E-05
0.044627321
1902.1418
0.0849
2
12
12000
323.157207
1.727298611
-0.112277
-0.00195957
1.959570481
1917.6448
3.7816
12
6000
323.056564
1.726760666
-0.060753
-0.001060322
1.060322109
1926.8355
2.0468
12
3000
323.055944
1.726757352
-0.036004
-0.000628378
0.628377812
1929.1495
1.2130
12
1500
323.059096
1.7267742
-0.021473
-0.000374768
0.374768269
1929.9367
0.7234
12
750
323.080108
1.726886511
-0.010882
-0.000189924
0.189923546
1930.1419
0.3666
12
375
323.095828
1.726970535
-0.006426
-0.000112153
0.112152978
1930.1224
0.2165
12
187.5
323.091306
1.726946365
-0.002957
-5.16085E-05
0.051608521
1930.1808
0.0996
12
93.75
323.095363
1.72696805
-0.002083
-3.63546E-05
0.036354599
1930.1608
0.0702
12
46.875
323.097468
1.726979301
-0.002051
-3.57961E-05
0.035796103
1930.1484
0.0691
12
23.4375
323.100741
1.726996796
-0.001069
-1.86573E-05
0.018657257
1930.1318
0.0360
-------
On
12
11.71875
323.100501
1.726995513
-0.000626
-1.09256E-05
0.010925578
1930.1340
0.0211
12
5.859375
323.102179
1.727004482
-0.000147
-2.56559E-06
0.002565591
1930.1243
0.0050
12
2.929687
323.110338
1.727048092
-0.000273
-4.76467E-06
0.004764669
1930.0755
0.0092
12
1.464844
323.139292
1.727202854
-0.000159
-2.77503E-06
0.002775027
1929.9026
0.0054
12
0.732422
323.216797
1.727617123
0.000023
4.01419E-07
-0.00040142
1929.4399
-0.0008
12
0.366211
323.422531
1.728716786
-0.000288
-5.02646E-06
0.005026464
1928.2124
0.0097
12
0.183105
323.911373
1.731329682
0.000854
1.49049E-05
-0.01490486
1925.3018
-0.0287
12
0.091553
325.038169
1.737352488
0.00023
4.01419E-06
-0.00401419
1918.6281
-0.0077
22
12000
336.904658
1.800779729
-0.184324
-0.003217007
3.217006772
1819.6939
5.9548
22
6000
336.239917
1.79722664
-0.081832
-0.001428214
1.428213896
1848.5030
2.6489
22
3000
-269.849307
-1.44236404
10.33924
0.180450773
-180.450773
1410.2220
414.7659
22
1500
335.949257
1.795673042
-0.0208
-0.000363022
0.3630224
1855.9126
0.6739
22
750
335.938166
1.795613759
-0.011497
-0.000200657
0.200657141
1856.2528
0.3725
22
375
335.929037
1.795564964
-0.006282
-0.00010964
0.109639746
1856.3893
0.2035
22
187.5
334.872923
1.789919959
0.106859
0.00186501
-1.86501013
1851.6582
-3.4732
22
93.75
335.48912
1.793213576
0.039618
0.000691453
-0.69145295
1857.4016
-1.2853
3
22
46.875
335.641672
1.794028978
0.017438
0.000304345
-0.30434541
1857.7328
-0.5655
22
23.4375
335.676273
1.794213922
0.006201
0.000108226
-0.10822605
1857.7881
-0.2011
22
11.71875
335.683262
1.794251279
0.002453
4.28122E-05
-0.04281221
1857.7795
-0.0795
22
5.859375
335.674033
1.79420195
0.001702
2.9705E-05
-0.02970501
1857.8335
-0.0552
22
2.929687
335.652439
1.794086528
0.000397
6.92884E-06
-0.00692884
1857.9556
-0.0129
22
1.464844
335.628219
1.79395707
0.000852
1.487E-05
-0.01486996
1858.0891
-0.0276
22
0.732422
335.612495
1.793873024
0.000176
3.07173E-06
-0.00307173
1858.1768
-0.0057
22
0.366211
335.680806
1.794238152
0.000467
8.15055E-06
-0.00815055
1857.7985
-0.0151
22
0.183105
335.93873
1.795616774
-0.000306
-5.34062E-06
0.005340618
1856.3723
0.0099
22
0.091553
336.483848
1.798530469
-0.001663
-2.90243E-05
0.029024339
1853.3624
0.0538
-------
Complete pH data set
DI
Tap
1st
2nd
3rd
1st
2nd
3rd
9.33
9.23
9.27
7.67
7.61
7.61
Control
9.34
9.11
9.26
7.66
7.62
7.59
9.34
9.08
9.25
7.64
7.63
7.60
9.41
9.17
9.33
7.72
7.32
7.38
0.025% Steol
9.40
9.20
9.29
7.70
7.32
7.37
9.37
9.13
9.27
7.71
7.32
7.37
9.46
8.93
9.28
7.85
8.09
8.08
0.5% Dowfax
9.44
9.15
9.22
7.85
8.05
8.07
9.43
9.12
9.16
7.85
8.00
8.07
9.13
8.89
8.67
8.39
8.26
8.37
5% Dowfax
9.12
8.78
8.61
8.37
8.24
8.39
-
8.77
8.61
8.36
8.23
8.37
6.75
6.73
6.32
7.76
7.78
7.89
8% AMA 80-1
6.40
6.32
6.14
7.76
7.81
7.88
6.32
6.34
6.17
7.76
7.79
7.88
165
-------
Complete specific conductivity data set. Data in parenthesis was measured using the
Accumet 4-electrode conductivity probe.
DI
Tap
1st
2nd
3rd
1st
2nd
3rd
23.87
19.56
16.54
1020.93
1047.64
1075.56
Control
24.52
18.99
34.29
1019.96
1048.29
1075.52
24.43
17.76
34.91
1019.93
1049.25
1076.5
(13.1)
(17.6)
(37.6)
(1010)
(1040)
(1020)
32.37
29.15
31.54
1049.96
1068.69
1073.62
0.025% Steol
33.34
35.2
31.67
1051.87
1068.96
1079.28
32.5
36.13
33.25
1055.42
1069.23
1080.26
(25.7)
(27.8)
(26.2)
(1030)
(1040)
(1050)
355.15
366.4
362.79
1250.13
1250.73
1278.9
0.5% Dowfax
382.62
364.52
363.21
1252.17
1254
1279.3
388.18
364.44
364.57
1256.55
1255.71
1280.55
(364)
(362)
(331)
(1220)
(1240)
(1210)
2215.57
2241
2264.86
3043.41
3133.12
3117.85
5% Dowfax
2223
2248
2270.3
3047.08
3135.45
3127.09
-
2253
2274.63
3049.79
3137.72
3132.16
(2150)
(2280)
(2170)
(2780)
(2830)
(2820)
8355.21
8365.47
8364.3
8948.88
8906.66
9012.64
8% AMA 80-1
8349.7
8354.65
8366.57
8371.75
8367.92
8376.07
8957.92
8969.32
8914.17
8953.22
9029.82
9047.16
(6050)
(7830)
(6340)
(6010)
(6060)
(6220)
166
-------
Complete dissolved oxygen data set
DI
Tap
1st
2nd
3rd
1st
2nd
3rd
7231
7725
7468
8564
8139
8706
Control
7149
7679
7558
8486
8082
8675
7168
7673
7563
8527
8080
8613
7518
7681
7518
7672
4091
5205
0.025% Steol
7417
7677
7468
7658
4357
5036
7348
7417
7417
7641
4394
5323
7485
7269
7354
7798
7777
7533
0.5% Dowfax
7453
7175
7220
7793
7591
7473
7388
7108
7160
7842
7520
7377
7526
7426
7400
7939
7689
7586
5% Dowfax
7453
7307
7391
7887
7650
7543
-
7280
7426
7831
7643
7463
7537
7574
7421
7414
7641
7626
8% AMA 80-1
7430
7496
7358
7340
7588
7585
7382
7445
7338
7286
7542
7550
167
-------
Complete dielectric constant data set
DI
Tap
1st
2nd
3rd
1st
2nd
3rd
23.8
25.5
21.9
23.7
23.4
22.4
23.9
23.6
21.9
23.5
23.4
22.4
1
24.0
23.6
22.1
23.6
23.4
22.4
V-x onir 01
24.0
23.4
21.9
23.8
23.3
21.5
23.9
23.3
21.9
23.9
23.4
21.6
24.0
23.3
21.9
23.9
22.8
21.6
23.6
24.5
25.2
24.1
22.1
25.3
23.5
25.1
25.2
24.0
22.1
25.4
0.025% Steol
23.4
25.0
25.2
24.0
22.0
25.3
23.2
24.4
25.1
23.9
22.1
24.2
23.2
24.4
25.1
23.8
22.1
24.3
22.9
24.3
25.0
23.8
21.7
24.2
22.8
24.8
25.1
25.1
24.1
25.1
22.8
24.5
25.1
25.1
24.2
25.0
0.5% Dowfax
22.8
24.4
25.2
25.2
24.2
24.9
22.6
24.1
23.9
23.9
24.2
24.9
22.4
24.1
23.9
23.9
24.1
24.9
22.4
24.3
23.9
23.9
24.2
24.2
22.3
23.7
24.7
24.7
23.5
23.4
22.4
23.7
24.7
24.7
23.5
23.4
5% Dowfax
22.4
23.7
23.4
23.4
23.4
23.3
22.4
23.4
23.2
23.2
23.6
22.6
22.2
23.3
23.4
23.4
23.8
22.5
22.0
23.2
23.2
23.2
23.6
22.6
21.4
24.1
24.0
21.8
22.4
24.0
21.3
24.1
24.2
21.8
22.5
23.7
21.5
24.0
22.8
21.7
22.7
24.0
8% AMA 80-1
21.1
23.5
22.2
21.6
22.5
23.4
21.1
23.2
22.2
21.4
22.3
23.5
20.9
23.0
22.0
21.5
22.0
23.4
168
-------
REFERENCES CITED
Aal, G.Z.A., Atekwana, E.A., Slater, L.D., and Atekwana, E.A., 2004, Effects of
microbial processes on electrolytic and interfacial electrical properties of
unconsolidated sediments: Geophysical Research Letters, v. 31.
Adamson, A.W. and Gast, A.P., 1997, Physical Chemistry of Surfaces: New York,
Wiley.
Adepelumi, A.A., Solanke, A.A., Sanusi, O.B., and Shallangwa, A.M., 2006, Model tank
electrical resistivity characterization of LNAPL migration in a clayey-sand
formation: Environmental Geology, v. 50, p. 1221-1233.
Ajo-Franklin, J.B., Geller, J.T., and Harris, J.M., 2006, A survey of the geophysical
properties of chlorinated DNAPLs: Journal of Applied Geophysics, v. 59, p. 177-
189.
Anderson, M.J. and Whitcomb, P.J., 2000, DOE Simplified: Practical tools for effective
experimentation: Portland, Productivity, Inc..
Archie, G.E., 1942, The electrical resistivity log as an aid in determining some reservoir
characteristics: Transactions of the American Institute of Mining and
Metallurgical Engineers, v. 146, p. 54-61.
Atekwana, E.A., Atakwana, E.A., Rowe, R.S., Werkema, D.D., and Legall, F.D., 2004,
The relationship of total dissolved solids measurements to bulk electrical
conductivity in an aquifer contaminated with hydrocarbon: Journal of Applied
Geophysics, v. 56, p. 281-294.
Barranco, F.T. and Dawson, H.E., 1999, Influence of aqueous pH on the interfacial
properties of coal tar: Environmental Science & Technology, v. 33, p. 1598-1603.
Battelle and Duke Engineering Services, 2002, Surfactant-enhanced aquifer remediation
(SEAR) design manual, Naval Facilities Engineering Command, p. 1-110.
Black, B., Johnson, L., Cepeda, M., and Rocky Mountain Oilfield Testing Center, 2007,
Geophysical imaging technology provides insight on EOR practices, p. 133-135.
Borden, R.C., Gomez, C.A., and Becker, M.T., 1995, Geochemical indicators of intrinsic
bioremediation: Groundwater, v. 33, p. 180-189.
Box, G.E.P. and Cox, D.R., 1964, An Analysis of Transformations: Journal of the Royal
Statistical Society Series B-Statistical Methodology, v. 26, p. 211-252.
Bradford, S.A., Abriola, L.M., and Rathfelder, K.M., 1998, Flow and entrapment of
dense nonaqueous phase liquids in physically and chemically heterogeneous
aquifer formations: Advances in Water Resources, v. 22, p. 117-132.
169
-------
Bramwell, D.P. and Laha, S., 2000, Effects of surfactant addition on the
biomineralization and microbial toxicity of phenanthrene: Biodegradation, v. 11,
p. 263-277.
Brewster, M.L. and Annan, A.P., 1994, Ground-penetrating radar monitoring of a
controlled DNAPL release: 200 MHz radar: Geophysics, v. 59, p. 1211-1221.
Brewster, M.L., Annan, A.P., Greenhouse, J.P., Kueper, B.H., Olhoeft, G.R., Redman,
J.D., and Sander, K.A., 1995, Observed migration of a controlled DNAPL release
by geophysical methods: Groundwater, v. 33, p. 977-987.
Brown, C.L., Delshad, M., Dwarakanath, V., Jackson, R.E., Londergan, J.T., Meinardus,
H.W., McKinney, T., Oolman, T., Pope, G.A., and Wade, W.H., 1999,
Demonstration of surfactant flooding of an alluvial aquifer contaminated with
DNAPL, Proceedings Field Testing of Innovative Subsurface Remediation
Technologies: Washington, D.C., American Chemical Society, p. 64-85.
Chambers, J.E., Loke, M.H., Ogilvy, R.D., and Meldrum, P.I., 2004, Noninvasive
monitoring of DNAPL migration through a saturated porous medium using
electrical impedance tomography: Journal of Contaminant Hydrology, v. 68, p. 1-
22.
Conrad, S.H., Glass, R.J., and Peplinski, W.J., 2002, Bench-scale visualization of
DNAPL remediation processes in analog heterogeneous aquifers: surfactant
floods and in-situ oxidation using permanganate: Journal of Contaminant
Hydrology, v. 58, p. 15-49.
Cook, R.D., 1977, Detection of influential observation in linear regression:
Technometrics, v. 19, p. 15-18.
Cytec Industries, I., 1994, Aerosol MA 80-1 Data Sheet.
Dalton, F.N., Herkelrath, W.N., Rawlins, D.S., and Rhoades, J.D., 1984, Time-domain
reflectometry: Simultaneous measurement of soil water content and electrical
conductivity with a single probe: Science, v. 224, p. 989-990.
Dow Chemical Company, 2009, Dowfax 8390 Technical Data Sheet.
Dwarakanath, V., Kostarelos, K., Pope, G.A., Shotts, D., and Wade, W.H., 1999, Anionic
surfactant remediation of soil columns contaminated by light nonaqueous phase
liquids: Journal of Contaminant Hydrology, v. 38, p. 465-488.
Endres, A.L. and Knight, R.J., 1993, A model for incorporating surface phenomena into
the dielectric response of a heterogeneous medium: Journal of Colloid and
Interface Science, v. 157, p. 418-425.
170
-------
Fountain, J.C., Starr, R.C., Middleton, T., Beikirch, M., Taylor, C., and Hodge, D., 1996,
A controlled field test of surfactant-enhanced aquifer remediation: Groundwater,
v. 34, p. 910-916.
Grimm, R.E., Olhoeft, G.R., McKinley, K., Rossabi, J., and Riha, B., 2005a, Nonlinear
complex-resistivity survey for DNAPL at the Savannah River Site A-014 Outfall:
Journal of Environmental and Engineering Geophysics, v. 10, p. 351-364.
Harwell, J.H., 1992, Factors Affecting Surfactant Performance in Groundwater
Remediation Applications, in Sabatini, D.A. and Knox, R.C., eds., Transport and
remediation of subsurface contaminants: Colloidal, interfacial, and surfactant
phenomena: Washington, D.C., American Chemical Society, p. 124-132.
Howarth, D.J. and Sondheimer, E.H., 1953, The Theory of Electronic Conduction in
Polar Semi-Conductors: Proceedings of the Royal Society of London Series A-
Mathematical and Physical Sciences, v. 219, p. 53-74.
Hoyle, B.L., 1989, Ground-water quality variations in a silty alluvial soil aquifer,
Oklahoma: Groundwater, v. 27, p. 540-549.
Hwang, Y.K., Endres, A.L., Piggott, S.D., and Parker, B.L., 2008, Long-term ground
penetrating radar monitoring of a small volume DNAPL release in a natural
groundwater flow field: Journal of Contaminant Hydrology, v. 97, p. 1-12.
IDEX Corporation, 2007, Ismatec IP Peristaltic Pump Operating Manual.
In-Situ, I., 2008, Multi-parameter TROLL 9500, WQP-100 Operator's Manual.
Jawitz, J.W., Annable, M.D., Rao, P.S.C., and Rhue, R.D., 1998, Field Implementation of
a Winsor Type I Surfactant/Alcohol Mixture for in Situ Solubilization of a
Complex LNAPL as a Single-Phase Microemulsion: Environmental Science &
Technology, v. 32, p. 523-530.
Karapanagioti, H.K., Sabatini, D.A., and Bowman, R.S., 2005, Partitioning of
hydrophobic organic chemicals (HOC) into anionic and cationic surfactant-
modified sorbents: Water Research, v. 39, p. 699-709.
Kehew, A.E. and Passero, R.N., 1990, Ph and Redox Buffering Mechanisms in A Glacial
Drift Aquifer Contaminated by Landfill Leachate: Ground Water, v. 28, p. 728-
737.
Kelleners, T.J., Robinson, D.A., Shouse, P.J., Ayars, J.E., and Skaggs, T.H., 2005,
Frequency dependence of the complex permittivity and its impact on dielectric
sensor calibration in soils: Soil Science Society of America Journal, v. 69, p. 67-
76.
171
-------
Kirsch, R., 2006, Petrophysical properties of permeable and low-permeable rocks, in
Kirsch, R., ed., Groundwater Geophysics: A Tool for Hydrogeology: New York,
Springer, p. 1-22.
Kleinbaum, D.G., Kupper, L.L., Muller, K.E., and Nizam, A., 1998, Applied Regression
Analysis and Other Multivariable Methods: Pacific Grove, California, Duxbury
Press.
Kueper, B.H. and McWhorter, D.B., 1991, The behavior of dense, nonaqueous phase
liquids in fractured clay and rock: Groundwater, v. 29, p. 716-728.
Kueper, B.H., Redman, J.D., Starr, R.C., Reitsma, S., and Mah, M., 1993, A field
experiment to study the behavior of tetrachloroethylene below the water table:
Spatial distribution of residual and pooled DNAPL: Groundwater, v. 31, p. 756-
766.
Lesmes, D.P. and Frye, K.M., 2001, Influence of pore fluid chemistry on the complex
conductivity and induced polarization responses of Berea sandstone: Journal of
Geophysical Research, v. 106, p. 4079-4090.
Londergan, J.T., Meinardus, H.W., Mariner, P.E., Jackson, R.E., Brown, C.L.,
Dwarakanath, V., Pope, G.A., Ginn, J.S., and Taffinder, S., 2001, DNAPL
removal by a heterogeneous alluvial aquifer by surfactant-enhanced aquifer
remediation: Groundwater Monitoring and Remediation, v. 21, p. 57-67.
Longino, B.L. and Kueper, B.H., 1995, The use of upward gradients to arrest downward
dense, nonaqueous phase liquid (DNAPL) migration in the presence of
solubilizing surfactants: Canadian Geotechnical Journal, v. 32, p. 296-308.
Lowe, D.F., Oubre, C.L., and Ward, C.H., 1999, Surfactants and cosolvents for NAPL
remediation: A technology practices manual: New York, Lewis Publishers.
Lowrie, W., 2003, Fundamentals of Geophysics: New York City, Cambridge University
Press.
Mackay, D.M. and Cherry, J. A., 1989, Groundwater Contamination - Pump-And-Treat
Remediation .2: Environmental Science & Technology, v. 23, p. 630-636.
Marshall, D.J. and Madden, T.R., 1959, Induced polarization, a study of its causes:
Geophysics, v. 24, p. 790-816.
Martel, R., Golinas, P.J., and Saumure, L., 1998, Aquifer washing by micellar solutions::
3 Field test at the Thouin Sand Pit (L'Assomption, Quebec, Canada): Journal of
Contaminant Hydrology, v. 30, p. 33-48.
Martinez, A. and Byrnes, A.P., 2001, Modeling dielectric constant values of geologic
materials: An aid to ground-penetrating radar data collection and interpretation:
Current Research in Earth Sciences, v. 247.
172
-------
Mercer, J.W. and Cohen, R.M., 1990, A review of immiscible fluids in the subsurface:
Properties, models, characterization and remediation: Journal of Contaminant
Hydrology, v. 6, p. 107-163.
Montgomery, D.C., 1997, Design and Analysis of Experiments: New York, John Wiley
& Sons.
National Research Council of the National Academies, 2005, Contaminants in the
subsurface: Source zone assessment and remediation: Washington, D.C., National
Academies Press.
Nicholson, R.V., Cherry, J.A., and Reardon, E.J., 1983, Migration of Contaminants in
Groundwater at A Landfill - A Case-Study .6. Hydrogeochemistry: Journal of
Hydrology, v. 63, p. 131-176.
Olhoeft, G.R., 1985, Low-frequency electrical properties: Geophysics, v. 50, p. 2492-
2503.
Pope, G.A. and Wade, W.H., 1995, Lessons from Enhanced Oil Recovery Research for
Surfactant-Enhanced Aquifer Remediation
Surfactant-Enhanced Subsurface Remediation, Surfactant-enhanced Subsurface
Remediation: Washington, DC, American Chemical Society, p. 142-160.
Prego, M., Cabeza, O., Carballo, E., Franjo, C.F., and Jimenez, E., 2000, Measurement
and interpretation of the electrical conductivity of 1-alcohols from 273K to 333K:
Journal of Molecular Liquids, v. 89, p. 233-238.
Qin, X.S., Guang, G.H., Chakma, A., Chen, B., and Zeng, G.M., 2007, Simulatin-based
process optimization for the surfactant-enhanced aquifer remediation at
heterogeneous DNAPL-contaminated sites: Science of the Total Environment, v.
381, p. 17-37.
Radic Research, 2007, SIPLab II Operating Manual.
Ramsburg, C.A. and Pennell, K.D., 2001, Experimental and economic assessment of 2
surfactant formulations for source zone remediation at a former dry cleaning
facility: Groundwater Monitoring and Remediation, v. 21, p. 68-82.
Revil, A. and Glover, P.W.J., 1998, Nature of surface electrical conductivity in natural
sands, sandstones, and clays: Geophysical Research Letters, v. 25, p. 691-694.
Reynolds, D.A. and Kueper, B.H., 2000, DNAPL flow through fractured porous media,
in Wickramanayake, G.B., Gavaskar, A.R., and Gupta, N., eds., Treating dense
nonaqueous-phase liquids (DNAPLs): Remediation of chlorinated and recalcitrant
sompounds: Columbus, Battelle Press, p. 165-172.
173
-------
Robert, T., Martel, R., Conrad, S.H., Lefebvre, R., and Gabriel, U., 2006, Visualization of
TCE recovery mechanisms using surfactant-polymer solutions in a two-
dimensional heterogeneous sand model: Journal of Contaminant Hydrology, v.
86, p. 3-31.
Rosen, M.J., 2004, Surfactants and interfacial phenomena: Hoboken, NJ, Wiley -
Interscience, p. 1-444.
Rothmel, R.K., Peters, R.W., St.Martin, E., and Deflaun, M.F., 1998, Surfactant
foam/bioaugmentation technology for in situ treatment of TCE-DNAPLs:
Environmental Science Technology, v. 32, p. 1667-1675.
Sabatini, D.A. and Knox, R.C., 1992, Transport and remediation of subsurface
contaminants : colloidal, interfacial, and surfactant phenomena : Washington,
D.C., American Chemical Society.
Sauck, W.A., 2000, A model for the resistivity structure of LNAPL plumes and their
environs in sandy sediments: Journal of Applied Geophysics, v. 44, p. 151-165.
Schoen, J.H., 1996, Physical properties of rocks: Fundamentals and principles of
petrophysics: Tarrytown, NY, Pergamon, p. 1-583.
Schwarz, G., 1962, A theory of the low-frequency dielectric dispersion of colloidal
particles in electrolyte solution: Journal of Physical Chemistry, v. 66, p. 2636-
2642.
Slater, L.D. and Glaser, D.R., 2003, Controls on induced polarization in sandy
unconsolidated sediments and application to aquifer characterization: Geophysics,
v. 68, p. 1547-1558.
Slater, L.D. and Lesmes, D., 2002, IP interpretation in environmental investigations:
Geophysics, v. 67, p. 77-88.
Sogade, J.A., Scira-Scappuzzo, F., Vichabian, Y., Weiqun Shi, Rodi, W., Lesmes, D.P.,
and Morgan, F.D., 2006a, Induced-polarization detection and mapping of
contaminant plumes: Geophysics, v. 71, p. B75-B84.
Soilmoisture Equipment Corporation, 2005, MiniTrase 6050X3K1 Operating
Instructions.
Stat-Ease, I., 2007, Design Expert 7.1.1, Design Expert.
Stepan Company, 2005, Steol CS-330 Material Safety Data Sheet.
Telford, W.M., Geldard, L.P., and Sheriff, R.E., 1990, Applied Geophysics: New York,
Cambridge University Press.
174
-------
Topp, G.C., Davis, J.L., and Annan, A.P., 1980, Electromagnetic determination of soil
water content: Measurements in coaxial transmission lines: Water Resources
Research, v. 16, p. 574-582.
U.S.EPA, 1991, Dense nonaqueous phase liquids: A workshop summary: U.S. EPA
Report EPA/600/R-92/030, 81p.
U.S.EPA, 1996, Surfactant-enhanced DNAPL remediation: Surfactant selection,
hydraulic efficiency, and economic factors: U.S. EPA Report EPA 600/S-96/002,
15p.
U.S.Silica Company, 1997, ASTM 20/30 unground silica product data sheet.
Vanhala, H., 1997, Mapping oil-contaminated sand and till with the spectral induced
polarization (SIP) method: Geophysical Prospecting, v. 45, p. 303-326.
Vanhala, H. and Soininen, H., 1995, Laboratory technique for measurement of spectral
induced polarization response of soil samples: Geophysical Prospecting, v. 43, p.
655-676.
Vinegar, H.J. and Waxman, M.H., 1984, Induced polarization of shaly sands:
Geophysics, v. 49, p. 1267-1287.
Waxman, M.H. and Smits, L.J.M., 1968, Electrical conductivities in oil-bearing shaley
sands: Society of Petroleum Engineering, v. 8, p. 107-122.
Werkema, D.D., 2008, Report on the geoelectrical detection of surfactant enhanced
aquifer remediation of PCE: Property changes in aqueous solutions due to
surfactant treatment of perchloroethylene: Implications to geophysical
measurements: U.S. EPA Report EPA/600/R-08/031, 77p.
West, C.C. and Harwell, J.H., 1992, Surfactants and subsurface remediation:
Environmental Science Technology, v. 26, p. 2324-2330.
Williams, J.H., Lapham, W.W., and Barringer, T.H., 1993, Application of
Electromagnetic Logging to Contamination Investigations in Glacial Sand-And-
Gravel Aquifers: Ground Water Monitoring and Remediation, v. 13, p. 129-138.
Willumsen, P.A., Karlson, U., and Pritchard, P.H., 1998, Response of fluoranthene-
degrading bacteria to surfactants: Applied Microbiology and Biotechnology, v.
50, p. 475-483.
Zhong, L., Mayer, A., and Glass, R.J., 2001, Visualization of surfactant-enhanced
nonaqueous phase liquid mobilization and solubilization in a two-dimensional
micromodel: Water Resources Research, v. 37, p. 523-537.
175
-------
Zonge, K., Wynn, J., and Urquhart, S., 2005, Resistivity, induced polarization, and
complex resistivity, in Butler, D.K., ed., Near-surface geophysics: Tulsa, Society
of Exploration Geophysicists, p. 219-263.
176
-------
-------
SERIV
United States
Environmental Protection
Agency
Please make all necessary changes on the below label, detach
or copy and return to the address in the upper left hand corner.
If you do not wish to receive these reports CHECK HERE
detach, or copy this cover, and return to the address in the
upper left hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA PERMIT No. G-35
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-10/041
April 2010
www.epa.gov
Recycled/Recyclable
Printed wfth vegetable-based ink ©?
paper that contains a minimum of
SjP „ » U-;
processed chlorine free
------- |