SPE 59532

Electrical Dissipation Factor: A Novel Way To Detect Changes in Shale Water Content
P.S. Adisoemarta, L.R. Heinze, Texas Tech U., and S.L. Morriss, U of Texas at Austin
SPE Members

Copyright 2000, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 2000 SPE Permian Basin Oil and Gas
Recovery Conference held in Midland, Texas, 2123 March 2000.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
The alteration of shales, caused by adsorption of water while
drilling, is one of many contributors to the wellbore stability
problem that costs the industry in the order of $ 400-500
million annually (Bol et.al., 1992). This alteration of shale
problem has acquired a logging perspective due to the
increasing use of measurements while drilling. The capability
of taking real time and time-lapse measurements while still
drilling introduces the possibility of detecting a swelling
problem while something can still be done about it.
The objective of this study (Adisoemarta, 1999) is to
observe the changes of complex electrical characteristic of
shale as a function of water content. The word complex in
the electrical characteristic term means this study will not only
observe the in-phase electrical response but also the out-ofphase response.
By taking the ratio of the in-phase to the out-of-phase
electrical response, this study found that this ratio, the
dissipation factor, changes linearly as a function of shale
water content. This method can be easily applied to both the
drilling / wellbore stability or formation evaluation areas.
Electromagnetic Theory
Maxwells equations and certain constitutive relationships
describe the macroscopic electrical behavior of conducting
dielectrics subjected to a harmonic sinusoidal field.
All electromagnetic fields are created from distributions of
charges and currents in which the electric field (resulting from
the charge distributions), and the current densities (from the
current distributions) are related through the complex transfer
functions that result from Maxwells equations:
(1)
E B / t = 0

H D / t J = 0

(2)

and with the conservation of charge defined as:

J q / t = 0

(3)

are the components of the electromagnetic field, where:

E = electric field intensity (volt/meter)
B = magnetic flux density (webers/meter2 )
H = magnetic field intensity (ampere-turn/meter)
D = electric displacement (coulomb/meter2 )
J = electric current density (ampere/meter2 )
q = charge density (coulomb/meter3 ).
For the case of homogeneous, isotropic, and a non-zero
electrical conductivity medium, the electric field becomes
E = 0

(4)

For the case of materials that exhibit electromagnetically

linear behavior, the following relationships are valid:
D = E

(5)

J = E

(6)

B = H

(7)

where
= dielectric permittivity (farad/m)
= electric conductivity (mho/m or siemens/m)
= magnetic permeability (henry/m).
The equation for total current density, JT, is the result of
solving equations (2), (5), and (6):
J T = H = E + dE / dt

(8)

where E is the conduction current component (in phase with

applied voltage), and dE/dt is the displacement current
component (the out of phase response).

ADISOEMARTA, HEINZE AND MORRISS

For the time-harmonic case, in which temporal behavior,

ET, described by
ET = Eoe

jt

(9)

where
j = (-1)1/2
= frequency (radians / sec),

(10)

where * and * are, in general, complex scalars dependent

upon frequency (Fuller and Ward, 1970).
By substituting these complex scalars, with their respective
real (shown with a single prime, ' ) and imaginary (shown with
a double prime, " ) components,

=
j

(11)

=
j

(12)

equation (10) can be rewritten as the following:

J T = (eff j eff ) ET

(13)

where eff = ' + " and eff = ' - "/ are real functions of
frequency that constitute the experimentally measured rock
parameters (Fuller and Ward, 1970).
In the simplest form, equation (13) becomes the familiar
Ohms law that relates electric current to the voltage across
the medium and the conductance of the medium itself.
Ohms law is as follows:
I = GV

(14)

where
I = electric current (amperes)
G = electric conductance (mho or siemens)
V = electric potential (volt).
Conductivity and Susceptivity
The output from the measurement equipments are not directly
comparable to each other, and they have to be converted to
conductivity and susceptivity before they can be compared to
each other. For the impedance analyzer, the outputs are
resistance and reactance. The conversions to resistivity,
reactivity, conductivity and susceptivity are calculated using
the equations below:
r=

Rd 2
t

g=

b=

the total current density in a medium becomes the following:

J T = ( j ) ET

x=

(15)

SPE 59532

Xd 2
t

r
r + x2
2

x
r + x2
2

(16)

(17)

(18)

where
R = resistance (ohm)
X = reactance (ohm)
r = resistivity (ohmmeter)
x = reactivity (ohmmeter)
t = sample thickness (meter)
d = electrode diameter (meter)
g = conductivity (Siemens/meter)
b = susceptivity (Siemens/meter).
For the network analyzer, the outputs are in real and
imaginary (in-phase and out-of-phase) permittivity, and the
conversions to conductivity and susceptivity are as follows:
= eff
g = 0

(19)

b = 0
= eff

(20)

where
g = conductivity (Siemens/meter)
b = susceptivity (Siemens/meter)
' = relative permittivity, real part (dimensionless)
" = relative permittivity, imaginary part (dimensionless)
0 8.854x10 -12 (Farads/meter).

Dissipation Factor
A dielectric parameter that will be used in this research is the
dissipation factor. The dissipation factor is defined as the ratio
of in-phase to quadrature current components. This term
describes the ratio of electromagnetic energy that is given to a
medium to the amount that is lost as heat

D=
(21)

Electromagnetic Measurement
To cover a wide range of frequency, this study uses the fourelectrode parallel-plate setup for the low frequency
electromagnetic measurements, and the open-ended coaxial
probe method for the high frequency electromagnetic
measurements.

SPE 59532

ELECTRICAL DISSIPATION FACTOR: A NOVEL WAY TO DETECT CHANGES IN SHALE WATER CONTENT

Parallel-Plate Method
The material under test is formed as a thin flat sheet and
sandwiched between two electrodes, thus making a capacitor,
as shown in Figure 1. This method is very dependable and
very easy to implement.
The disadvantages of this method are a frequency limit of
around 15 MHz; fringe effect or the stray current path at the
sample edge; necessity for a smooth, flat with parallel faces
sample; and sample maximum thickness of around 10 mm.
Methods for parallel-plate measurement can be separated
into two general categories, which are two-electrode and fourelectrode measurements. Figure 2 shows these two methods.
The two-electrode measurement is where the setup is exactly
as shown in the Figure 1. The sample is placed between two
current electrodes, and these electrodes are also the voltage
electrodes. This method is prone to errors at the lower end of
the frequency, less than 10 kHz, due to the contact impedance
from the ionic build-up at the sample-electrode interface
(Lewis et al., 1986; Garrouch, 1992). The ionic build-up is the
zone where the frequency changes are slow enough that ions
have sufficient time to pile up at the sample-electrode
interface. As the frequency increases, the possibility of buildup diminishes, and at sufficiently high frequency, the ionic
build-up zone vanishes completely. To reduce the effect of
ionic build-up, Scott et al. (1967) used a combination of
platinized platinum electrodes and blotter pads to increase the
surface area at the sample-electrode interface, hence
minimizing the ionic build-up.
In a four-electrode set-up, separate voltage electrodes are
used at each sample face located away from the current
electrodes, outside the ionic build-up zone. As the voltage
electrodes have a high impedance, they will draw almost no
current, so that essentially no ionic build-up will occur at the
voltage electrodes.
Open-Ended Coaxial Probe
The open-ended coaxial probe is basically a cut-off coaxial
line where the sample touches the face of the probe, as shown
in Figure 3. The face of the sample must be flat and smooth.
The electromagnetic field is reflected off the sample, and the
reflection (S11) is related to the complex permittivity of the
sample.
The advantages of this method are that the frequency range
is relatively broad (100 MHz up to 2.6 GHz), the probe is very
small and portable, the probe can work with both solid and
liquid samples, and there is no need to machine the sample for
a perfect fit, such as for the coaxial-method. The disadvantage
of this method is that because the depth-of-investigation is
very small, the reading is mostly affected by a small portion of
the sample and is thus very prone to heterogeneity of the
sample.
Frequency Coverage
Each shale sample in this study will be measured for its
electromagnetic properties in the range from 5 Hz up to 1.3
GHz (1.3 x 109 Hz) using two different measurement setup,
shown in Figure 4. The low frequency range, 5 Hz up to 13

MHz (1.3 x 107 Hz), is with the parallel-plate method while

the high frequency range, 20 MHz - 1.3 GHz (1.3 x 109 Hz), is
with the open-ended coaxial probe. Overlaid in the figure are
the commercially available logging tools.
The width of the frequency coverage is about 7.5 orders of
magnitude (5x101 to 1.3x109 Hz) that encompasses almost half
of the span of the available measurement equipment capability
(10-6 - 1011 Hz) (Olhoeft, 1981); hence this study is very
significant to enhance our knowledge about electrical
properties of shales.
Shale Sample Characteristics
The shale sample that was used in this study is Wellington
shale. This shale is surface shale that was cored in central
Utah. Table 1 shows the results of the X-ray diffraction test.
Using nominal values of density for each component, the
grain density for Wellington is 2.71 g/cc. Water content of the
shale samples were determined from their native and dried
weight, and were found as 2.86% by weight. Total porosity,
defined here in terms of the moveable water by the drying
process, is 7.0%. The CEC this shale is 9.4 meq /100 gram
and composition is mostly Illite.
Sample Preparation
This study uses samples of 2-inch diameter by 1-cm thick
sample disks. This dimension was determined to satisfy the
requirement of both the parallel-plate and open-ended probe.
To ensure that each sample will have a good contact to the
measurement probe, its parallellness and flatness have to be of
less than 3-thousandth of an inch. This requires several hours
of hand lapping the individual sample surface.
Each individual shale sample for this study was
preconditioned to ensure that no air was entrapped in the shale
pore space. This procedure is required to avoid any
measurement artifacts due to entrapped air (Forsans and
Schmidt, 1994). First, the moveable water was removed from
the sample by oven drying at 99OC for several days until the
sample weight remained constant. Next, the sample is
resaturated in a controlled way by increasing, in steps, the
level humidity environment inside a dessicator until it reached
its original sample weight. This process usually takes several
weeks to finish.
Reference and Standard Materials
To ensure that the equipment is working properly, methanol
and brines have been measured for their electrical properties
and compared against their published values.
The permittivity (both the real and imaginary part) of
methanol that was measured for this study and published data
by Jordan et al. (1978) and Marsland (1987) is shown in
Figure 5. The measured values for this study agree with the
published data, suggesting that the high-frequency open-ended
coaxial probe is working properly. Accuracy of the probe
measurement is demonstrated over the frequency range of 20
MHz to 1.3 GHz.
Figure 6 shows the effective conductivity of NaCl and KCl
brines as 2000-ppm concentration, at ambient (60oF)

ADISOEMARTA, HEINZE AND MORRISS

temperature, measured with the open-ended coaxial probe.

Also shown on the plot is the conductivity of NaCl and KCl
taken from Charts Gen-8 and Gen-8 of the Log Interpretation
Charts (Schlumberger, 1995).
To ensure a consistent reading on the day-to-day
measurement and to detect any drift that might occur from the
measurement equipment, a set of standard materials has been
selected and they are measured before and after the actual
sample measurement. The standard materials are as follows:
- AlrcynTM ALC-1130BK, a thermoset synthetic rubber,
for the low-frequency equipment
- Deionized water for the high frequency equipment.
Figure 7 shows the complex electrical properties of Alcryn
as a function of frequency. Alcryn was selected as a reference
material because at 1 MHz, it has a high dielectric constant,
near that of shale, and the material is not hazardous.
The transition between the parallel-plate and the openended coaxial probe measurements is smooth, although the
probe data are increasingly noisy below 200 MHz as shown by
bigger error bars. Alcryn rubber is a relatively non-conductive
material.
Figure 8 shows the dimensionless effective permittivity
(i.e. effective dielectric constant, eff) vs. frequency of
deionized water measured with the parallel-plate fixture and
the open-ended coaxial probe. Overlaid on the plot is the
dielectric constant formula presented by Hasted (1973).
Measurement Timeline
Each sample will be measured for its electromagnetic
properties while the sample dries out, starting from fully
saturated to a much drier condition.
One method to reduce the water content in a shale sample
is to expose the sample to ambient air for a while, letting a
small amount of pore water at the sample edge evaporate, and
then allow the water to redistribute evenly before a set
electromagnetic measurement is done. A carefully arranged
air exposure procedure has to be executed so that the amount
of water remaining in the sample can be monitored in a
controlled way.
The procedure exposes the sample to air for 5 minutes
while the sample sits in an opened digital analytical balance,
and then resaturate the sample for 2 hours while the sample is
immersed in inert fluid. Figure 9 shows the timeline of one
shale sample.
As shown in Figure 9, a fully water saturated shale sample
is taken out of a dessiccator that has water activity (relative
humidity) of 0.96, the water activity of native Wellington
shale, and then its electromagnetic properties are measured.
This measurement is for the fully saturated condition. After
this measurement is done, the sample is then exposed to air for
5 minutes in an opened digital balance. To re-equilibrate the
sample to the new water saturation condition after the sample
edges are exposed to air, the sample is immersed in inert fluid,
Isopar, for two hours. After the sample has assumed the new
equilibrium condition, another set of electromagnetic

SPE 59532

measurement is executed. The sequence repeats again until

the sample shows a visible crack on the sample surface that
indicates air has entered the sample.
Figure 10 is the plot of the sample weight at the end of air
exposure plotted against the cumulative air exposure time. By
dividing the actual water loss (the decrease in sample weight)
to the total moveable water (determined from the oven drying
of the sample), the actual water loss in percent is calculated,
which is plotted on the secondary Y-axis.
Dissipation Characteristics
Figure 11 shows the dissipation factor of Wellington as a
function of both frequency and water content. The plot shows
a distinct peak (local maxima), and the magnitude of the peak
decreases as the sample gets drier. Also a local minima can be
seen on the figure, but the details are obstructed behind the
mountain.
Figure 12 shows the characteristics of the local the maxima
and minima of Wellington shale as a function of water loss.
As can be seen on the plot the frequency of the maxima and
minima point get lower as sample gets drier. The magnitude
of the maxima and minima are also dropping. This outcome is
very interesting because either by using these characteristics
individually or in combination, samples water content can be
determined. One potential application is in the time lapsed
MWD situation where the dissipation factor characteristics can
be displayed at the drill floor real time and can be used to
detect changes in boreholes water content and predict the
onset of wellbore instability. There are several technical
obstacles for applying these methods: the tool must operate at
a much lower frequency than that of the current MWD tools,
and also be of variable frequency instead of the single
frequency mode of the current tools.
Another application of this method is in well logging. By
comparing the dissipation factor characteristic from different
logs on the same formation type, the formation water
saturation can be qualitatively approximated. Time lapsed
analysis of dissipation factor can be used to determine changes
in water content in the shale zone as a function of time.
Dissipation Factor at Downhole Condition
Due to higher pressure and temperature that insitu, downhole,
shale has to bear, a fully saturated downhole shale zone has a
lesser amount of water content than a fully saturated surface
shale.
The fully saturated insitu shale can be represented as a
point at around the 5% water loss on Figure 11. By looking at
the figure, one can see that for insitu shale, the dissipation
factor changes not only for a drying case (moving to the right
on the x-axis) but also for a swelling case (moving to the left
on the x-axis).

SPE 59532

ELECTRICAL DISSIPATION FACTOR: A NOVEL WAY TO DETECT CHANGES IN SHALE WATER CONTENT

Conclusions
Dissipation factor plot of shale shows a distinct maxima
and minima.
The location and magnitude of the dissipation factors
maxima and minima decreases as water loss increases
For the case of insitu shale, this study shows the changes
in dissipation factor for both the drying and swelling of
shale.
Dissipation factor characteristics from MWD tools can be
displayed at the drill floor in real time and can be used to
predict the onset of wellbore instability.
MWD tools need to be modified to work at low
frequencies, and also be of variable frequency
By comparing the dissipation factor characteristic from
different well-logs on the same formation type, the
formation water saturation can be qualitatively
approximated.
Nomenclature
B
= magnetic flux density (webers/meter 2 )
b
= susceptivity (Siemens/meter).
D = electric displacement (coulomb/meter 2 )
D = dissipation factor (dimensionless)
d
= electrode diameter (meter)
E
= electric field intensity (volt/meter)
G = electric conductance (mho or siemens)
g
= conductivity(Siemens/meter)
H = magnetic field intensity (ampere-turn/meter)
I
= electric current (amperes)
J
= electric current density (ampere/meter 2 )
j
= (-1)1/2
q
= charge density (coulomb/meter 3 ).
R
= resistance (ohm)
r
= resistivity (ohmmeter)
t
= sample thickness (meter)
V = electric potential (volt).
X = reactance (ohm)
x
= reactivity (ohmmeter)
= electric conductivity (mho/m or siemens/m)

= magnetic permeability (henry/m).

= frequency (radians / sec),

= dielectric permittivity (farad/m)

'
= relative permittivity, real part (dimensionless)
" = relative permittivity, imaginary part
(dimensionless)

at the 1994 SPE/ISRM Rock Mechanics in Petroleum

Engineering Conference held in Delft, The Netherlands, 29-31
August 1994.
Fuller, Brent D., and Ward,Stanley H.: Linear System Description of
the Electrical Parameters of Rocks, IEEE Transactions on
Geoscience Electronics, Vol GE-8, No.1 (January 1970) 7-15.
Garrouch, Ali: Dielectric Properties of Fluid Saturated Rocks,
Ph.D. Dissertation, The University of Texas at Austin, 1992.
Hasted, J B.: Aqueous Dielectrics, Chapman and Hall, London 1973.
Huang, F.S.C., and Shen, Liang C.: Analysis of Error Due to
Presence of Gaps in the Measurement of Rock Samples in a
Coaxial Line, Geophysics, v.48, 1983, 206-212.
Javalagi, Mohan I.: Electrical Resistivity of Hydrated Shales,
Masters Thesis, The university of Texas at Austin, 1990.
Jordan, B.P., Sheppard, R.J. and Szwarnowski, S.: The Dielectric
Properties of Forma-mide, Ethanediol and Methanol, J. Phys.
D (1978) 695-701.
Lewis, M.G., Sharma, M.M., and Dorfman, M.H.: Techniques for
Measuring the Electrical Properties of Sandstone Cores, paper
18178 presented at the 63rd Annual Technical Conference and
Exhibition of the Society of Petroleum Engineers, Houston, TX,
Oct 2-5, 1986.
Marsland, T.P. and Evans, S.: Dielectric measurement with an openended coaxial probe, IEEE trans. 1987.
Olhoeft, G.R.: Electrical Properties of Rocks, Physical Properties
of Rocks and Minerals, 11-2, edited by W.R. Judd, Y.S.
Toulokian, R.F. Roy, McGraw-Hill, New-York, 1981, 257-310.
Schlumberger.: Log Interpretation Charts, 1995.
Scott, J.H., Carroll, R.D., and Cunningham, D.R.: Dielectric
Constant and Electrical Conductivity Measurements of Moist
Rock: a New Laboratory Method, Journal of Geophysical
Research (1967), 5101-5115.

Table 1. X-Ray Diffraction Bulk Analysis Result

(Javalagi, 1990)

Crystalline Components
Smectite
Illite
Kaolinite
Chlorite Fe
Calcite
Dolomite
Pyrite
Feldspar Na
Feldspar K
Quartz

Weight % in Crystalline
Portion Wellington Shale
0
12
2
1
14
5
4
5
2
56

0 8.854x10 -12 (Farads/meter).

References:
Adisoemarta, Paulus S.: Complex Electrical Permittivity of Shale as
a Function of Frequency and Water Content, Ph.D.
Dissertation, Texas Tech University, Lubbock TX, 1999.
Bol, G.M., Sau-Wai, W., Davidson, C. J., and Woodland, D.C.:
Borehole Stability in Shales, paper SPE 24975, Proc.
EUROPEC92, 1992.
Forsans, T. M. and Schmitt, L.: Capillary forces: The neglected
factor in shale instability studies?, paper SPE 28029 presented

Material
under test

Figure 1: Parallel-Plate Method

Electrodes

ADISOEMARTA, HEINZE AND MORRISS

Voltage Electrodes

Voltage
Electrodes
Ion Buildup

Ion Buildup

NaCl and KCl (2000 ppm)

Ion Buildup

Ion Buildup

Current Electrodes

SPE 59532

1.0
DC Conductivity from
Schlumberger Chart
Book
chart book
NaCl = 0.37 S/m
KCl = 0.35 S/m

Current Electrodes

Figure 2: The Two and Four Electrode Setup

Figure 3: The Open-Ended Coaxial Probe

Open-Ended
Coaxial Probe

Effective conductivity,

eff (S/m)

0.8

0.6
NaCl from
Schlumberger
chart book
NaCl

KCl

0.4

KCl

KCl from
Schlumberger
chart book

NaCl

0.2

High-Frequency Setup

Parallel-Plate
Method

0.0
10 4

10 5

10 2

10 3

10 4

10 5

10 8

10 9

10 10

10 6

10 7

10 8

Figure 6: Effective conductivity of NaCl and KCl brines at 2000

0
ppm concentration, at ambient (60 F) condition.

EPT

MWD

Dual
Induction

Shallow
Laterolog

Deep
Laterolog

10 1

10 7

Frequency (Hz)

Low-Frequency Setup

10 0

10 6

10 9 10 10

Frequency (Hz)

Figure 4: Frequency Span of this Study, from 5 Hz up to 1.3 GHz

overlaid with current well logging tool frequencies and the two
measurement methods.

Alcryn ALC-1170BK
15
Sample thickness: 2 mm for parallel plate
25 mm for coaxial probe
Parallel plate diameter: 5 mm

10

35

30

25
Jordan
Marsland
Adisoemarta

-5

20
10 7

10 8

10 9

10 10

Frequency (Hz)
Figure 5: Electrical properties of Methanol from this study
overlaid with two published data.

Dimensionless Effective Conductivity

and Effective Permittivity

40

eff = eff / 0

15

Effective dielectric constant,

Dimensionless effective conductivity,

eff / 0

Methanol

10
eff = eff /0

eff

Parallel-plate

Open-ended

-5
10 4

10 5

10 6

10 7

10 8

10 9

10 10

Frequency (Hz)
Figure 7: Alcryn ALC-1130BK as standard material used for
calibrating the equipment before and after every measurement.

SPE 59532

ELECTRICAL DISSIPATION FACTOR: A NOVEL WAY TO DETECT CHANGES IN SHALE WATER CONTENT

Deionized Water
Wellington WN4 Absolute Reading

Symbols are plotted for all data from

parallel-plate, but not for probe since
data density is too high (201 points).

10 4

3
Dissipation Factor

Effective dielectric constant,

Parallelplate

10 3

10 2

Probe

0
10

Hasted formula

10

10

25
10

Fre
que
ncy
,

Hasted data

10

20
10
Hz

15

10

10

5
10

10 4

10 5

10 6

10 7

10 8

10 9

2 hrs. in Isopar
equilibration

pct
oss,
ater L
ble W
a
e
v
Mo

Figure 11: Dissipation Factor of Wellington WN4 as a function of

both frequency and Water Loss. Dissipation Factor shows a
distinct peak (maxima) and valley (minima) that the height drops
as water loss increases.

5min air

5min air

5min air

Figure 8: The dimensionless effective permittivity vs frequency of

deionized water.

2 hrs. in Isopar
equilibration

10 10

Frequency (Hz)

Fully saturated sample

Wellington WN4

dessicator
@0.96

Dissipation Factor
4

10000000

3.5

1000000

EM Measurements

Wellington Shale
Assuming 2.4g of water OIP

Minima
3

100000
Frequency (Hz )

Figure 9: Timeline for one shale sample starting from a fully

saturated sample until the sample shows a visible crack.

2.5
10000
2

Maxima
1000

1.5
100

67

Maxima

25

10

66.8
15
66.7
10
66.6
5

66.5
66.4

10

20

30
40
50
Air Exposure Time (min)

60

70

0.5
Minima

20
Water in place loss (% )

Sample Weight

66.9

Magnitude

eff = eff / 0

10

0
80

Figure 10: Sample Weight and Water Loss as a Function of

Cumulative Exposure Time

10
15
Water Loss (pct)

20

0
25

Figure 12: Dissipation Factor Characteristics as a Function of

Water Loss. Both the maxima and minima points move with
respect to frequency and magnitude as water loss increases.