Kiberu PDF

Induced polarization and
Resistivity measurements on a
suite of near surface soil
samples and their empirical
relationship to selected
measured engineering
parameters
Johnmary Kiberu
March, 2002
Induced polarization and Resistivity
measurements on a suite of near surface
soil samples and their empirical
relationship to selected measured
engineering parameters
by
Johnmary Kiberu
Thesis submitted to the International Institute for Geo-information Science

and Earth Observation in partial fulfilment of the requirements for the degree
in Master of Science in Applied Geophysics.
Degree Assessment Board

Thesis advisor Dr. Jean. Roy
Thesis examiners Dr. Ir. Slob (External Examiner, TU Delft)
Prof. C. Reeves (Chairman)
Prof. F. Van der Meer (Member)
Ir. Sporry (Member)
INTERNATIONAL INSTITUTE FOR GEO - INFORMATION SCIENCE AND EARTH OBSERVATION
ENSCHEDE , THE NETHERLANDS

Disclaimer
This document describes work undertaken as part of a programme of study at
the International Institute for Geo-information Science and Earth Observa-
tion (ITC). All views and opinions expressed therein remain the sole respon-
sibility of the author, and do not necessarily represent those of the institute.
Abstract
Time–domain induced polarization (IP) and resistivity laboratory

measurements were performed on 43 natural soil samples obtained from
Antequera (Spain), Nairobi (Kenya) and Enschede (The Netherlands).
Soils are polarizable when they contain clay minerals, which are the main
causes of swelling and shrinking. These two effects are a major concern
for foundation engineering because they cause extensive damage to struc-
tures. Therefore, there is a need to develop a technique that would be able
to identify, at the outset of an investigation, those soils that posses unde-
sirable expansion characteristics.
Clayey soils exhibit membrane polarization which is due to ion accu-
mulations and variation in ion mobility within the pore passages. The use
of IP was an attempt to identify clay minerals. The following parameters
were measured: Chargeability (m) and resistivity (ρ) using a transmitter
and receiver. Soils were saturated with distilled water and emphasis was
given to the analysis of m as a function of water content. The relationship
of m to particle size distribution was illustrated by use of triangular plots.
The laboratory results showed that the IP response was dependant
on the amount of water and cation exchange capacity (CEC), which is
directly related to clay minerals. Decay curve analysis was performed in
terms of 3 exponential decays and the respective amplitudes (Aj) were
found to be statistically dependant on water content. However, the ratios
of Aj did not show any dependency on water content, and were therefore,
used to differentiate soil samples according to their origin.
Keywords
Induced polarization, time domain, chargeability, decay curve analysis,
resistivity, cation exchange capacity, grain size, porosity, water content
i
Abstract
ii
Acknowledgements
I would wish to express my thanks and gratitude to my supervisor: Dr. Jean Roy
who proposed this project and guided me up to the final handing in of this thesis. His
critical comments and helpful suggestions contributed to enrich my thesis.
All the staff of Applied Geophysics division of ITC Enschede are greatly acknowl-
edged. Prof. Dr. C. Reeves, Dr. S. Barritt, Ir. R. Sporry and Mr. W. Hugens are
thanked for their support and encouragement.
I would wish to extend my thanks to the Head of Geology Department, Makerere
University; Dr. Andrew Muwanga, for giving me permission to attend this course.
All the staff of ITC and DISH hotels are appreciated for ensuring the optimum aca-
demic environment that has helped me in the completion of the study period. The
Netherlands government is thanked for awarding me a scholarship that enabled me
to attend this course at ITC.
The research would not have taken foot without the help of Mr. Patrick Kariuki
who provided me with the necessary data and the soil samples. I thank Drs. B. de
Smeth for the advice he offered during the experiment and for the permission to use
the Chemistry Laboratory.
To my classmates, without you, life would have been hard for the 18 months.
I would lie to express my sincere appreciation and gratitude to Mr. John Kironde
for the relentless help, he has offered me by staying and keeping a keen eye on the
property at home. Thank you very much for your prayers and encouraging words.
My gratitude to the computer cluster manager, secretary and ITC reproduction
department for their assistance at all moments.
iii
Acknowledgements
iv
Contents
Abstract i
Acknowledgements iii
List of Tables ix
List of Figures xi
1 Introduction 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Objectives of the research . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Research Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Sources of Data and Information . . . . . . . . . . . . . . . . . . . . . . 5
1.6.1 Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Research methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Background to the research 9

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 The use of geophysical techniques: electric properties and the IP tech-
nique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.1 Position of the IP technique among other
geophysical methods . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 General characteristics of clays . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.1 Formation and types of clay minerals . . . . . . . . . . . . . . . 15
2.3.2 Cation exchange capacity (CEC) . . . . . . . . . . . . . . . . . . 17
3 Physicochemical basis of the Induced Polarization method 19

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Electrochemical nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
v
Contents
3.3 Electrode impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.3.1 Fixed Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.2 Diffusion layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Origin Of Induced Polarization . . . . . . . . . . . . . . . . . . . . . . . 22
3.4.1 Electrode polarization (Overvoltage) . . . . . . . . . . . . . . . . 22
3.4.2 Membrane (electrolytic) polarization . . . . . . . . . . . . . . . . 23
3.5 Methods of measurement of IP effect . . . . . . . . . . . . . . . . . . . . 26
3.5.1 Time domain method . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.2 Frequency domain method . . . . . . . . . . . . . . . . . . . . . . 28
3.5.3 Complex Resistivity (Spectral IP) . . . . . . . . . . . . . . . . . . 29
4 Study areas and soil samples 33

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.1 Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.2 The Netherlands . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.3 Kenya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2.1 Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2.2 The Netherlands . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2.3 Kenya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Geologic setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3.1 Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3.2 The Netherlands . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3.3 Kenya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4 Description of soil samples . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4.1 Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4.2 The Netherlands . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.4.3 Kenya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5 Data acquisition methodology 41

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.2 Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.2.1 Saturation of soil samples . . . . . . . . . . . . . . . . . . . . . . 43
5.3 Potential electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.4 Geometric factor K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.5 Determination of water content . . . . . . . . . . . . . . . . . . . . . . . 47
5.6 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.7 Laboratory techniques and methods of
measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.7.1 Time domain method . . . . . . . . . . . . . . . . . . . . . . . . . 52
vi
Contents
5.8 Test measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.9 Sources of errors and their magnitude . . . . . . . . . . . . . . . . . . . 57
6 Experimental data and processing 61

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.2 Water content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.3 Analysis of IP response characteristics . . . . . . . . . . . . . . . . . . . 62
6.3.1 Chargeability and resistivity . . . . . . . . . . . . . . . . . . . . 62
6.3.2 Chargeability and water content . . . . . . . . . . . . . . . . . . 63
6.3.3 Chargeability and clay mineralogy/CEC . . . . . . . . . . . . . . 67
6.3.4 Chargeability and granulometrical composition of soils . . . . . 68
6.4 Decay curve analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.5 Interpretation and discussion of results . . . . . . . . . . . . . . . . . . 79
7 Models 83
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.2 Theoretical aspects of model development . . . . . . . . . . . . . . . . . 83
8 Conclusion and Recommendation 87

8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
8.2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Appendix A 97
Appendix B 113
vii
Contents
viii
List of Tables
5.1 Window widths (ms) for the decay curve semi logarithmic mode . . . . 54
6.1 Water content (%) values obtained with sample 1 (Spain) . . . . . . . . 61

6.2 Typical original receiver readings . . . . . . . . . . . . . . . . . . . . . . 73
6.3 Results of statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . 75
1 Soil sample description 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

2 Soil sample description 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3 Geometric constant (K) values . . . . . . . . . . . . . . . . . . . . . . . . 116
4 IP/Resistivity measurements reliability . . . . . . . . . . . . . . . . . . 117
5 Causes for the unreliable determination of Aj and τj constants . . . . . 118
6 Calibration of the transmitter . . . . . . . . . . . . . . . . . . . . . . . . 119
ix
List of Tables
x
List of Figures
1.1 Research methodology flow chart . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Position of the IP amongst other geophysical techniques . . . . . . . . 13

2.2 The silicon tetrahedron (after Bioag, 2001) . . . . . . . . . . . . . . . . 14
2.3 The Aluminium octahedron (after Bioag, 2001) . . . . . . . . . . . . . . 15
2.4 1:1 type clay-kaolinite (Sherman, 1997) . . . . . . . . . . . . . . . . . . 16
2.5 2:1 type clay-montmorillonite (Sherman, 1997) . . . . . . . . . . . . . . 17
3.1 Potential across the fixed and diffuse layers with (a) current into and
(b) out of the interface ( (Sumner, 1976) . . . . . . . . . . . . . . . . . . 21
3.2 Grain electrode polarization (Reynolds, 1997) . . . . . . . . . . . . . . . 23
3.3 Membrane polarization associated with constriction between mineral
grains (Reynolds, 1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4 Membrane polarization associated with negatively charged clay parti-
cles (Reynolds, 1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.5 The integrated decay voltage used as a measure of chargeability m
(Reynolds, 1997) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.1 Location of study area (Spain) (Dizon-Bacatio, 1992) . . . . . . . . . . . 34

4.2 General map of study area (The Netherlands) (Magellan, 1992) . . . . 34
4.3 General map of study area (Kenya) (Fairburn, 1963) . . . . . . . . . . . 35
5.1 pF rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.2 PVC made sample tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.3 Saturation process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
5.4 Sample holder and potential electrodes . . . . . . . . . . . . . . . . . . 45
5.5 Cylindrical sample tube . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.6 Weighing process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.7 Schematic representation of the IP instrumentation set up . . . . . . . 49
5.8 IP instrumentation set up . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.9 Transmitted time domain current wave . . . . . . . . . . . . . . . . . . 51
xi
List of Figures
5.10 IRIS Elrec–6 receiver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

5.11 Calibration of the transmitter . . . . . . . . . . . . . . . . . . . . . . . . 53
5.12 Decay curve sampling mode . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.13 Measurement process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.14 Data transfer to a PC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.15 illustration of the diagonal cut of the sample tubes . . . . . . . . . . . . 59
5.16 illustration of the set up . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6.1 A plot of chargeability against resistivity . . . . . . . . . . . . . . . . . 63

6.2 A plot of chargeability (mV/V) against water content (%) . . . . . . . . 65
6.3 A plot of chargeability (mV/V) against water content (%) . . . . . . . . 66
6.4 A plot of chargeability (mV/V) against CEC (meq/100g) . . . . . . . . . 67
6.5 Effect of grain size on the chargeability of soil samples . . . . . . . . . 69
6.6 Effect of grain size on the chargeability soil samples . . . . . . . . . . . 70
6.7 Effect of grain size on the chargeability of Spain and Kenya samples . 71
6.8 Decay curves obtained from sample 10 from Spain at different water
contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.9 Illustration of fits of one exponential and three exponentials to the data
points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.10 A plot of A1 against water content . . . . . . . . . . . . . . . . . . . . . 76
6.11 Ratio A2/A3 against water content . . . . . . . . . . . . . . . . . . . . . 77
6.12 Chargeability, Resistivity and Ratio A2/A3 from the analysis of decay
curves against water content . . . . . . . . . . . . . . . . . . . . . . . . 78
7.1 Anion trap number against clay concentration (Keller, 1966) . . . . . . 85
1 Chargeability against water content for sample 1 . . . . . . . . . . . . 97

xii
List of Figures

32 A1 against water content . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
35 Ratio A1/A2 against water content . . . . . . . . . . . . . . . . . . . . . 109
xiii
List of Figures
xiv
Chapter 1
Introduction
1.1 Introduction
The use of Induced electrical Polarization (IP) technique over the past 30 years has
proven to be one of the most successful geophysical methods in providing in situ in-
formation about rock mineralogy especially in search for disseminated minerals with
electronic conductivity. For this case, the technique is being applied to quite a new
field, which concerns materials that do not contain conductive minerals but rather
clay minerals. The method has been used, though rarely in the fields of hydrogeology
(Vacquier et al., 1957), (Marshall and Madden, 1959), oil and gas field exploration
(Sternberg and Oehler, 1990) and in environmental studies such as mapping of pol-
luted land areas (Towel et al., 1985). However, it is worth noting that previous IP
studies of non metallic earth materials have focused on clay mineral soils, sandy and
shaly sediments containing clay minerals (Klein and Sill, 1982). Laboratory studies
on the electrical characteristics of such soils is able to show diagnostic signatures
of what they consist and thus, hopefully, can lead to a proper classification of soil
material in terms of clay presence and other materials.
Polarization is a geophysical phenomenon which measures the slow decay of volt-
age in the ground after the cessation of an excitation current pulse (time domain
method) or low frequency variation of the resistivity of the earth (frequency domain
method) (Sumner, 1976). In simple terms, the IP effect reflects the degree to which
the subsurface is able to store electric charge, analogous to a leaky capacitor. It occurs
when an electric current passes through a rock/soil. If the current is interrupted, a
difference in potential, which decays with time is observed. The rate of decay of this
potential (induced polarization potential) depends on the lithology of the rock, its
pore geometry and degree of water saturation. Schlumberger (1920) noted that this
phenomenon was taking place in the bulk volume of the rock and not on the field elec-
trodes used to measure it. The dependency of polarizability of rocks/soils upon their
1
1.1. Introduction
lithological composition and hydrogeological properties favors the application of the

IP method for hydrogeological (groundwater) and engineering geologic investigations.
This study is based on laboratory induced polarization and resistivity measure-

ments on a suite of near surface soil samples from Spain, The Netherlands and
Kenya. The empirical relationship between electrical IP measurements and the mea-
sured engineering parameters will be a subject of interest in the analysis of the re-
sults. For this research, the engineering parameters are taken to be CEC, clay con-
tent, porosity and grain size, however the results will be compared with other engi-
neering parameters from the main soil swelling project1 . These near surface natural
soils rarely contain conductive particles (e.g graphite), and their electrical properties
are attributed to the presence of clay minerals. Soils that contain clay minerals and
change volume with water content are termed as swelling soils. An attempt to use
this technique on soil samples in the laboratory, will hopefully provide a low cost
non–invasive means of studying swelling soils for engineering purposes and to asses
the suitability of the sub surface. This may ultimately decrease the cost of evaluation
and help to make the right decision at such an important stage in decision making
for an engineering programme.
The IP effect in earth materials having non metallic minerals is usually referred
to as membrane polarization. This requires the presence of particles such as clays,
with a negative surface charge, which in the presence of an electrolyte like water,
attracts cations and forms an electrical double layer on the particle surface. The
volume property of soil related to this phenomenon is the Cation Exchange Capacity
(hereafter referred to as CEC) (Olhoeft, 1985) which is proportional to the surface
charge density and the specific surface area of the solid phase (Bolt and Bruggenwert,
1978) and mostly related to clay minerals. This property supports the alternative
procedure of IP in trying to detect clay minerals apart from the other conventional
laboratory techniques like Xray diffraction.
The study will attempt to relate polarizability and resistivity parameters obtained
from Induced electrical polarization measurements with other properties measured
through other techniques and these include CEC, clay content, clay mineralogy, par-
ticle size, and water content. All soils that show a membrane type of polarization
have their own characteristic sets of electrochemical reactions. These reactions are
the basis of the IP technique in volume characterization and evaluation.
1
PhD program involved in mapping swelling clays
2
Chapter 1. Introduction
1.2 Problem definition
A systematic and thorough laboratory study, which is the purpose of this study, is
needed to understand the effects of clay within the soil samples on induced electri-
cal polarization laboratory measurements. The swelling behavior of soils, which is
caused by presence of clay minerals, is a destructive phenomenon from an engineer-
ing point of view and given its widespread distribution within the soils, a solution to
this problem has to be sought. However the problem of identifying, at the outset of
an
investigation, those soils likely to possess undesirable swelling characteristics re-
mains one of the most troublesome phases of the subject.
In a separate project, a study of this behavior by the Ground & Space-based short-
wave infrared (SWIR) reflectance spectrometers to identify clay minerals is under
way. These methods are designed for a low–cost large scale mapping. However, for
foundation engineering, it requires volume characterizations which involve taking
measurements on soils in situ. In many parts of the world, the possibility of dam-
age to structures due to the swelling of clays constitutes a severe problem in design
and construction. Owing to the serious economic implications as a result of this phe-
nomenon there is a need for an in-situ information provider and in this context, the
technique of induced electrical polarization will be used in an attempt to identify clay
minerals and provide information for volume investigations.
The IP technique in this research assumes the measurement of both resistivity
ρ and the polarizability η. This procedure will provide a convenient means of ex-
amining the dependence of IP response on variables such as grain size, clay type
(mineralogy), cation exchange capacity, and water content. The swelling potential of
clay minerals is directly linked to their ability to adsorb water and also depends on
the clay type. Therefore, this experimental investigation is focused on natural soil
samples that consist of phyllosilicates (clays). The fact that clay minerals have the
ability to conduct electricity through ion exchange reactions with pore fluids e.g. wa-
ter in soil samples, will provide the basis of evaluating the potential role of induced
polarization technique in swelling potential evaluation.
This project is run as a component in support of a wider project2 to detect the
swelling potential of clays by low cost non–invasive means. Swelling soils and expan-
sive clays are major hazards because they cause extensive damage worldwide every
year. Houses, offices, schools and factories are always subject to structural damage
once constructed on such soils.
2
PhD program involved in mapping swelling clays
3
1.3. Objectives of the research
1.3 Objectives of the research

The main objectives of this research project are the following:
• Study the technique of IP and concentrate on its possible application to the clay
swelling potential mapping.
• Observe the inter-relationship between variables such as IP response, resistiv-

ity, water content, CEC and grain size.
• Attempt to identify all sources of error encountered in the experiment in an

effort to improve the IP technique in this field of application.
• Attempt to quantify the Cation Exchange Capacity (CEC) and acquire infor-
mation about the nature of the clay mineralogy in the soil samples by the IP
technique.
1.4 Research Questions

1. Is there any relationship between the amount of water in the soil samples with
the chargeability and resistivity values so obtained?
2. Can the IP technique be used to differentiate between different soil types ac-
cording to their electrical response?
3. Does the geological and geographical background of the soil samples have an
effect on their electrical and water adsorption behaviors?
4. Is there any relationship between the granulometrical composition of soils and

their polarizability?
5. Is there any effect on the polarizability of soil samples based on the preparation
methods (remoulding, re–sampling and saturation)?
6. Can the analysis of the exponentials of the individual decay curves obtained
lead to parameters related to granulometry and water content?
7. Is it possible to use the IP technique to quantify CEC given other data such as
clay content and water content? Can it also identify the clay minerals?
4
1.5 Hypothesis
• By measuring the IP response when the water content and granulometric dis-
tribution (including clay content) are provided by other means, the IP technique
can be used to quantify CEC.
1.6 Sources of Data and Information

1.6.1 Literature
The time domain electrical induced polarization technique has been used in the study
of chargeability of fine sediments and soils. Some literature on the time domain
technique was obtained from the library to support this research and notably papers
from journals.
The software programs used in this research are categorized into two:
• Data Processing
– SPSS
• Data documentation and presentation
– MSword, Grapher, MsPowerpoint,and LaTex
1.6.2 Data
Ancillary data was obtained from the main swelling project (Mr. Patrick Kariuki3 ).
This included a set of data about particle size analysis and cation exchange capacity
values (CEC). These data sets covered soil samples from Spain and Kenya, no data
set was available for samples from The Netherlands. Therefore, the laboratory
study with samples from The Netherlands involved water content determination and
chargeability measurements.
A particle size distribution (PSD) analysis is a necessary index test for soils be-
cause it represents the relative proportions of different sizes of particles. From this
it is possible to tell whether the soil consists of predominantly gravel, sand, silt or
clay sizes and to a limited extent which of these sizes ranges is likely to control the
engineering properties.
There are mainly two procedures of carrying out a particle size analysis. The first
one is sieving which is used for gravel and sand size (coarse) particles that can be
separated into different size ranges with a series of sieves of standard aperture. The
3
PhD student, Earth System Analysis Department
5
1.7. Research methodology
second procedure is the sedimentation which is suitable for small size particles like
clay and silt (Head, 1992). Measurements in this case are made either by sampling a
suspension of soil particles in water by means of special pipette or by determining the
density of the suspension using a special hydrometer (the hydrometer method). The
data on particle size analysis for clay and silt fractions in this research was obtained
by the sedimentation procedure and in particular by means of a special pipette. The
sedimentation procedure is based on the fact that largest particles fall more quickly
through water than smaller ones. Thus, if a suspension is made up of clay and silt
in water, the silt will settle out more quickly than the clay and by measuring the
speed at which the suspension as a whole settles out, it is possible to determine the
distribution of particle sizes.
The CEC expresses the number of positive charges from the ions that neutralize
the negative charge of soil particles. It was determined through chemical means by
the following procedure. The soil samples were percolated with ammonium acetate
and the bases were measured in the percolate. The samples were then subsequently
percolated with sodium acetate, made by mixing dilute solution of Na (1000mg/l) with
25 ml of IM NH4 OAc solution, where the excess salt was then removed and the ad-
sorbed sodium exchanged by percolation with ammonium acetate. However, there
are two procedures for doing this:
1. Percolation tube procedure
2. Mechanical extractor procedure. For this study, the method used was the per-
colation tube procedure (Reeuwijk, 1995). For Na measurements, 12.5 ml of 2%Cs
suppressant was added to the sodium acetate solution to overcome ionization inter-
ference, before mixing it with the soil samples. Exchangeable Na was then measured
by flame emission spectrophotometry (FES) at a wavelength of 589.0 nm. The sodium
in this percolate is a measure for CEC.
In addition, information about geology, location, depth at which the soils where
collected and other relevant short descriptions were also made available.
1.7 Research methodology

The general research methodology was designed to make an effort to fulfill the ob-
jectives. The main induced polarization data was treated in collaboration with other
data obtained from the swelling clay project. The methodology for the research work
is graphically summarized in the flow–chart (Figure 1.1). The dotted line separates
the data obtained from the swelling clay project (Particle size analysis and CEC) and
that obtained through this research.
6
• Step 1
Literature review: This involved collecting published material through liter-
ature search tools from the Library about the subject of Induced Polarization.
This was done before the IP instrumentation was obtained.
• Step 2
Set up of measuring apparatus: This involved the modification of the ITC–
IP transmitter by installing a new current regulator and a crystal timer. The
building up of the non polarizable potential electrodes was done and the IP
sample holder put back in service after cleaning and lubrication. The Elrec–6
receiver was acquired on rental basis from IRIS France. Designing and cutting
the sample tubes from the Polyvinyl (PVC) tubes was done together with the
purchase of the red caps to be used for preventing soil spill over.
• Step 3
Sample preparation: Though the samples were already collected from the
field, there was a need to prepare them. This involved putting them in specially
prepared sample polyvinyl chloride (PVC) tubes and saturating them with wa-
ter. The sample tubes were labelled using numbers (1,2...43) and then weighed.
• Step 4
Data acquisition: This involved taking weights of the soil samples with their
containers every day. Secondly, calibrating the IP Transmitter before IP mea-
surements. The same procedure was followed with the Elrec–6 IRIS receiver.
The IP measurements were made in the time domain mode.
• Step 5
Data Processing: This involved the use of Grapher , Excel and SPSS soft-
wares for data analysis. Decay curve decomposition was done using non linear
least squares fitting algorithms. Graphical displays were used to relate the
IP response (Chargeability) with water content, CEC, resistivity and particle
size. MSWord, MSPowerpoint and LaTex were used for data documentation
and presentation.
• Step 6
Data interpretation: This was based on the results obtained (i.e IP response
dependence on clay content, clay mineralogy, and moisture content). After the
experimentation research, conclusions and recommendations were given.
7
1.7. Research methodology
LITERATURE REVIEW
SAMPLE PREPARATION
From the Clay

Swelling Project
LABORATORY EXPERIMENTS
Particle Size INDUCED POLARIZATION WATER CONTENT

Cation Exchange
Capacity Values (CEC) Distribution DETERMINATION
Values (PSD)
Time domain IP
IP parameter
Resistivity parameter
ANALYSIS OF THESE DATA SETS
INTERPRETATION
CONCLUSIONS
Figure 1.1: Research methodology flow chart
8
Chapter 2
Background to the research
2.1 Introduction
Work on the use of remote sensing techniques to identify clay minerals, notably, Space
and Ground based shortwave infrared reflectance spectrometry techniques, which
have the discriminatory capacity to identify clay minerals has been reported by Van
der Meer (1999). Basically these techniques allow the measurement of distinctive
spectra of smectite, illite and kaolinite (here listed from high to low swelling poten-
tial). Such work aims at low cost, large scale mapping of clay mineralogy however
a certain proportion of coverage of the soil response would be masked especially for
areas in or near urban areas, or under conditions of thick and dense vegetation. Each
of these techniques essentially provides surface measurements , however, for founda-
tion engineering, information about in situ properties of soils is required for volume
characterizations and also for ground truthing and calibration of satellite–borne sen-
sor data.
Standard geotechnical engineering practice for the delineation of areas of high
swelling potential builds on extensive laboratory analysis including x-ray diffraction
analysis for identifying clay mineralogy and Atterberg limits for deriving the swelling
index, which are both labour intensive and expensive, and do not provide in situ
measurements and results.
The swelling of clay minerals in soils is caused by the chemical attraction of wa-
ter, where water molecules are incorporated in the clay structure in between the clay
plates. When more water is available, the clay plates are further separated, destabi-
lizing the mineral structure and hence causing expansion. There are factors that are
involved in clay swelling and these include: type and amount of clay particles in soils,
clay particle size, soil moisture content and structure. It is these factors that lead to
the possible use of geophysical techniques including the IP technique that may de-
crease the amount of labour and money when undertaking such characterizations
9
2.2. The use of geophysical techniques: electric properties and the IP technique
that can provide a tool that may contribute to the mapping of a worldwide economic
problem.
2.2 The use of geophysical techniques: electric

properties and the IP technique
The technique of Induced Polarization was used in studies of Schlumberger (1920)
who observed a voltage transient after a cut–off of direct current pulse that had been
fed into the ground. In an attempt to describe the observed phenomenon, he at-
tributed the transient voltage to an electrical polarization of the ground and thus
used the same phenomenon in exploiting and locating unexposed ores.
The polarization phenomenon was previously studied in detail by Wait (1959a)
and its modern development stems largely from the work done by Bleil (1953). In
addition, Vacquier et al. (1957) and Marshall and Madden (1959) described, re-
spectively, the time domain IP measurements on artificial clay sand mixtures and
a theoretical model for membrane (clay) polarization. Ogilvy and Kuzmina (1972)
described additional time domain measurements on artificial mixtures, while Roy
and Elliot (1980) used horizontal layers of varying clay sand composition to model
negative apparent chargeabilities due to geometric effects. Vanhala and Soininen
(1995) carried out laboratory measurements using spectral induced polarization on
soil samples in order to assess the effect of mineralogy, grain size distribution, mois-
ture content and electrolyte composition on the resistivity of soil material.
Rocks, which contain clay mineralogy often display electrical properties which
can not be predicted by the bulk electrical properties of the constituents (Cohen,
1981). Interactions between clay minerals and ground water can produce polarization
phenomena and decreases in resistivity.
The electrical properties of minerals as they pertain to IP create three main
classes: insulators, electronic conductors (sulphides, oxides and graphite) and lastly
ion exchangers (clays). The two basic electrical properties are taken as resistivity
ρ measured in Ohm.m and the dielectric behaviour. Resistivity is a measure of the
opposition to flow of charge in a material and the dielectric behaviour refers to a sub-
stance that exhibits a polarization due to charge separation in an electric current.
Electrical resistivity is a physical property of soils and rocks that depends directly
on the pore fluid’s electrical resistivity and saturation, as well as the clay content.
When clay is present, the resistivity becomes frequency dependant because of a fun-
damental charge storage mechanism that introduces a capacitive-like element in the
equivalent circuit representation of the ground.
Vacquier et al. (1957) studied the membrane polarization effect in groundwa-
10
Chapter 2. Background to the research
ter prospecting, and found that the ratio of two chargeability values measured at
different times after cut–off of the current pulse, i.e., a quantity roughly dependant
on the relaxation time, could be related to the grain size of the sediment. Marshall
and Madden (1959) presented a membrane–polarization model which, although an
ultimate simplification for sediment texture, gave qualitatively correct predictions of
how grain size affects the phase spectrum of the sediment.
Polarization is caused by concentration gradients that develop at zone boundaries
in response to current flow in clay rock materials. Keller and Frischknecht (1966)
gave a qualitative development of this concept based on anion blockage to demon-
strate that theoretically there is a peak polarization response for different clays and
their concentrations. Variation in the intensity of polarization implies that clay con-
tent is also changing from place to place in the rock.
The dependency of resistivity on clay type and content has been examined to be
able to predict porosity and hydrocarbon saturation from well logs. Waxman and
Smits, 1968; Waxman and Thomas, 1974 assumed that the resistivity of rocks is
frequency dependant, which is true at low frequencies, and therefore presented a
semi empirical model for describing the dependence of resistivity on clay content,
expressed as CEC per unit volume.
Vacquier et al. (1957) found that the chargeability of sand–clay mixtures was
proportional to the constituent clay and its ionic exchange characteristics. The im-
portance of rock texture for the IP effect is manifested by a weak IP in compact clays
(low CEC), and a strong IP in sediments with disseminated clay particles (high CEC)
on the surface of larger grains. Increased electrolyte salinity (ion concentration) and
electrical conductivity decrease the IP effect (Klein and Sill, 1982).
Parkhomenko (1971) noted that for a fixed clay content, the chargeability of a
rock is greater for clays having higher ion exchange capacities, however, this applies
only in the active part of the IP response i.e to the left of the IP response drop, fol-
lowing the IP peak response in the bell–shaped IP response curve. However, this
does not apply in the case of massive or pure clay. She observed that the largest in-
duced polarization effects are obtained for clay contents in the range 3–10%. Higher
or lower clay contents will give rise to a lower induced polarization. She also noted
that induced polarization increases with increasing water content until an optimum
saturation is reached beyond which IP decreases. The salinity and composition of
the electrolyte have a profound influence upon the mode of occurrence of this max-
imum. With increasing concentration the maximum IP response is depressed and
shifts slightly towards lower water contents. In a study with shaly sands, Ogilvy
and Kuzmina (1972) demonstrated the occurrence of a maximum IP response for an
optimum water content.
Illiceto et al. (1982) presented data to the effect that the maximum IP response
11
2.2. The use of geophysical techniques: electric properties and the IP technique
for an optimum water content occurred within the range 0.2 < Sw < 0.6 (where Sw
is water saturation) for clean sands. When the finer (< 74 µm) fraction was washed
out, this maximum shifted to lower values of Sw. When working with fine sediments,
they found that the ratio A2/A3 using a decay model:
V (t) = A1e−t/T 1 + A2e−t/T 2 + A3e−t/T 3 (2.1)
with T1<T2<T3, was useful for lithotype identification, because for each lithotype
(clays, loams, silts and sands), the ratio was falling in distinct ranges and it was
statistically independent of water content.
Fraser and Ward (1965) worked with sandstone samples and observed that the
IP effect was increasing with the degree of saturation almost linearly. He further ob-
served that it was approaching zero at very small saturation. Keller and Frischkneit,
1966 and Ogilvy and Kuzmina, 1972 studied the effect of grain size parameters on
the IP response. They have reported that induced polarization is low in the extreme
cases of clean gravels or pure shales but that the IP effect attains a maximum value at
some intermediate grain size. Dakhnov (1962) noted with the fully saturated reser-
voir rock that the induced polarization response was approximately proportional to
the specific surface area of the constituent grains. It must be noted that specific
surface area increases with a decrease in grain size.
Laboratory measurements on soil samples showed that their frequency depen-
dant electrical response (FDER) can be used to estimate hydraulic conductivity and
porosity using inversion and regression models, and in that regard
Boadu and Seabrook (2000) used a double Cole–Cole model. The FDER resistivity
and phase spectra of a soil contains valuable information about its porosity, hydraulic
conductivity, texture and fluid properties. The transport properties of soils and rocks
such as fluid flow permeability and electrical conductivity, are important in near–
surface environmental and engineering applications (Mazac et al., 1985). According
to basic laboratory measurements on a variety of shaly sands, silts and clays, it is
showed that the main feature of their conductivity spectra in the frequency range
between 10−3 and 103 Hertz is nearly a constant phase angle (Borner et al., 1996).
Wet porous soils are generally very heterogeneous multi phase systems with a com-
plicated internal structure whose electrical and hydraulic properties depend on the
pore space geometry and also related to the microstructure of the pore space.
Polarization and the complex nature of electrical rock conductivity are attributed
to zones of unequal ionic transport properties along the pore channels caused by
charged interfaces and constrictions. Interactions at and near the contact area of
the solid and liquid phases are the main causes of the formation of an electrical
double layer. Various electrical phase boundary phenomena are of special interest,
because they result in a dispersion of the electrical conductivity in the frequency
12
range (1mHZ to 10kHz) (Olhoeft, 1985). Using Induced polarization measurements,

the complex electrical rock properties such as frequency effect or phase angle can
be obtained. They depend on pore space structure and microstructure of the internal
rock boundary layer. Therefore, they contain information about internal surface area,
rock porosity, rock permeability and fluid properties.
Draskovits and Smith (1990) worked with data from drill hole geophysical and
lithology well logs. They observed that coarse sand and gravel were characterized
by high resistivity and low polarizability. Maximum polarizability was associated
with silty beds having medium apparent resistivities whereas clay lithologies were
indicated by low resistivities and low polarizabilities.
2.2.1 Position of the IP technique among other

geophysical methods
The IP technique is a galvanic linear electrical method in which polarization of the
ground is made by a physical electrical contact of an electrode also known as an
ohmic contact. The essence of the IP technique is derived from the measurement
of the slow decay of voltage in the ground following the cessation of an excitation
current pulse (time domain method) or variations of the resistivity of the earth as
a function of frequency (frequency domain method). This leads to classifying the
technique, amongst other geophysical methods under galvanic electrical techniques
(Figure 2.1).
GEOPHYSICAL TECHNIQUES
POTENTIAL FIELD SEISMIC TECHNIQUES ELECTRICAL TECHINIQUES NUCLEAR TECHNIQUES OPTICAL

TECHNIQUES TECHNIQUES
INDUCTIVE GALVANIC Gamma ray Neutron

GRAVITY MAGNETICS MAGNETO-TELLURIC
spectrometry MRS
TECHNIQUE
TIME DOMAIN EM FREQUENCY DOMAIN EM
TELLURIC RESISTIVITY MISE-A-LA- IP MIP SP COMPLEX RESISTIVITY

MASSE
KEY TO SYMBOLS:
IP=Induced Polarization
MIP=Masgnetic Induced Polarization
SP=Spontaneous Polarization
MRS=Magnetic Resonance Sounding
Figure 2.1: Position of the IP amongst other geophysical techniques
13
2.3. General characteristics of clays
2.3 General characteristics of clays

Clays are one of the oldest ceramic raw materials and are recognized by certain prop-
erties. When they are mixed with water, they form a coherent, sticky mass, that is
readily mouldable and if dried, becomes hard, brittle and retains its shape. Clays
may take on various forms where they can easily be recognized as the sticky, tena-
cious constituent of soils and can also occur as rocks, which owing to compression, are
so hard and compacted that penetration and action of water are very slow processes
(Worrall, 1968). The commonest impurities in natural clays are quartz and mica-
ceous material but minor impurities such as hydrated iron oxide, ferrous carbonate
and pyrites also occur.
The basic building blocks of clay minerals are the tetrahedral layer and the oc-
tahedral layer. The tetrahedral layer is composed of either Si or Al in tetrahedral
coordination (4 oxygens) with oxygen. The tetrahedral layer is obtained by joining
Si tetrahedra (Figure 2.2) at their basal oxygens. The octahedral layer is composed
of cations in octahedral co–ordination (6 oxygens) with oxygen and can be obtained
by linking the Al octahedra (Figure 2.3) by the side or edge oxygens. Whether a
cation is in tetrahedral or octahedral co–ordination with oxygens depends on the size
of the cation. The cation is stable in a particular environment, as long long as it is
able to keep the oxygen anions from touching, thereby preventing repulsive forces
from destabilizing the structure. For example, Si is so small that only 4 oxygens are
able to fit around it and the most stable arrangement of these oxygens is in tetrahe-
dral co–ordination. Cations such as Mg2+ , Fe2+ and Fe3+ are larger and thus able
to accommodate 6 oxygens in their co-ordination environment. Aluminium size is
in between Si and Fe/Mg, therefore it has the ability to fit in either octahedral or
tetrahedral co-ordination.
Figure 2.2: The silicon tetrahedron (after Bioag, 2001)
Charge development within the clay structure develops under the following con-
14
Figure 2.3: The Aluminium octahedron (after Bioag, 2001)
ditions. If an Al tetrahedra substitutes for a Si tetrahedra in the tetrahedral layer,

excess negative charge will develop because of the difference in charge of the two
cations i.e Al+3 and Si+4 . Due to this charge difference, the negative charge (-2) on
the oxygens that is being shared between the Al and Si tetrahedral is not satisfied.
2.3.1 Formation and types of clay minerals

Clay minerals form at the expense of primary rock forming minerals. Primary min-
erals are unweathered minerals with relatively large crystals which formed under
constant conditions. Examples include Mica, quartz, muscovite and feldspar. Sec-
ondary minerals are highly weathered clay minerals with tiny crystalline structure
which formed under conditions of intense weathering.
Clays are formed from three distinct processes:
1. Alteration
Chemical and physical changes of primary mineral, however, with light changes
in structure
Primary minerals → intermediate minerals → secondary minerals
Muscovite + H2 O → vermiculite + K+
Vermiculite + H2 O → Montmorillonite + Mg2+
2. Recrystallization
Solubilizsed aluminium and silicon oxides from weathering clays recrystallize
to form kaolinite (Grim, 1968).
3. Weathering of clays to form other types of clay

) Montimorrilonite + K+
Illite *
15
Montmorillonite → Kaolinite + SiOx

The longer or the more intense the weathering process, the more the silica that
is lost and the lower Si:Al ratio.
For example, Vermiculite: 2Si:1Al layer (youngest)
Kaolinite: 1Si:1Al layer (more weathered)
Fe and Al oxides: no silica at all (highly weathered)
Clay minerals include kaolinite with a 1:1 type lattice, low CEC (3–15 meq/100g),
low surface area (15 m2 /g) and a low shrink–swell capacity. Its structure is shown in
Figure 2.4. Illite has got a 2:1 type lattice, medium CEC (10–40 meq/100g), medium
surface area (80 m2 /g) and a medium shrink–swell ability. Lastly, montmorillonite
with a 2:1 type lattice, high CEC (80–150 meq/100g), high surface area , high shrink–
swell capacity. It structure is shown in Figure 2.5. The CEC values are quoted from
Keller and Frischknecht, 1966.
Tetrahedral layer
Octahedral layer
Figure 2.4: 1:1 type clay-kaolinite (Sherman, 1997)
16
Tetrahedral layer
Octahedral layer
Tetrahedral layer
Interlayer cations (Ca/Na)
Figure 2.5: 2:1 type clay-montmorillonite (Sherman, 1997)
These minerals give rise to the CEC
2.3.2 Cation exchange capacity (CEC)

This is the total number of positive charges from exchangeable cations that neutral-
ize the negative charges on the soil particles. The exchangeable cations are ions that
neutralize the negative charge of the soil particles and exchange or equilibrate read-
ily with others in the soil solution. Soils that contain clays have cations such as Ca,
Mg, H, K and Na, which are loosely held to the surface and can subsequently be ex-
changed for other cations or essentially go into solution should the clay be mixed with
water hence the reason why they are called exchangeable ions. The CEC is quanti-
fied in meq/100 g i.e the weight of ions in milliequivalents adsorbed per 100 grams of
clay. The cation exchange capacity is a typical property of clayish material because
it relates to the clay mineral structure and is the basis for the electrical behavior of
such soils.
There are three main sources of the exchange capacity:
1. Isomorphous substitution
This involves mainly the substitution of Al3+ for Si4+ in the tetrahedral sheet
and Mg2+ or Fe+2 for Al3+ in the octahedral sheet. This is the major source of
clay exchange capacity, except for the kaolin minerals (Mitchel, 1993). However,
isomorphic substitution of Fe+3 for Al+3 in the octahedral layer does not give
rise to charge because both cations have a charge of +3.
2. Broken bonds
17
These occur along particle edges and on non–cleavage surfaces and may provide
exchange sites. They are the major source of the exchange capacity of kaolinite.
3. Replacement
Ionized hydrogen from OH groups is replaced by another type of cation (Bioag,
2001).
Earlier research work done in the field of electrical induced polarization on rocks
and soils shows that more work is needed to understand this phenomenon. Materi-
als that contain clay minerals such as montmorillonite, illite and kaolinite give IP
response due to their ability to have exchangeable cations in their lattice. A brief
description of the formation of clay minerals and the causes of the cation exchange
capacity have been indicated. Therefore, a detailed theoretical basis of the IP tech-
nique and its link to the structural characteristics of the clays needs to be given
attention, which is the subject of the next chapter.
18
Chapter 3
Physicochemical basis of the

Induced Polarization method
3.1 Introduction
The information discussed in this chapter provides the physical and chemical back-
ground of the Induced polarization technique.
3.2 Electrochemical nature

Empirical studies of the IP properties of mineralized rocks or synthetic ore samples
have been done to investigate effects of certain changes in the solution chemistry
(Henkel and Collins, 1961). Diffusion processes involve the motion of ions in solu-
tion resulting from the presence of a concentration gradient whereas the reactions
between ions in solution are referred to as chemical reactions. In the earlier studies
conducted by Marshal and Madden (1959) , it was indicated that diffusion processes
seemed dominant in the frequency range of interest for field measurements however
these studies were unable to identify any of the chemical reactions involved. In view
of the fact that chemical environments can undergo tremendous variations in nature,
a possibility exists that the IP phenomenon may also be very variable because of the
chemical environmental factors.
A study conducted by Angoran and Madden (1977) was directed at identifying the
processes that control the IP phenomenon and understanding how these processes
are effected by the chemical environment. Therefore, an investigation of the effect
of the in situ chemistry on the induced polarization phenomena was carried out by
means of laboratory studies of the electrode impedances of metallic and sulfide min-
erals. However, the reaction rate theory shows that this effect is largely due to the
19
3.3. Electrode impedance
impedance associated with the diffusion of ions involved in the charge transfer re-
action to and from the reaction sites. The conclusion from the laboratory study was
that the impedance is inversely proportional to the concentration of the reacting ions
and inversely proportional to the square root of the frequency. The basic parameter
of concern in this study is the electrode impedance.
3.3 Electrode impedance

This is the total resistance to the flow of alternating current across the interface be-
tween an electrode and an electrolyte; includes solution resistance, capacitances in
the fixed and diffuse layers, and warburg impedance. However the interpretation
of this impedance requires the assumption of a particular model electrical equiva-
lent circuit. When metallic or any other conducting minerals are part of the electric
current paths in an otherwise ionic conducting medium, an excess impedance arises
at the solution–mineral interface associated with the charge transfer reactions that
are needed to maintain the continuity of the electric current flow. The presence of
this impedance and the fact that it is highly frequency–dependant is the basis of the
electrode IP method. The concept of using an equivalent circuit to describe electrode
impedances was firmly established by the work of Grahame (1952). The environment
in the immediate vicinity of the electrode–solution interface must be taken into ac-
count. According to an excellent review of the properties of this zone as given by
Grahame (1952), this zone is made of two parts as described in the following section.
3.3.1 Fixed Layer

This contains a compact layer of ions and molecules rather rigidly held in place on the
electrode by chemical and adsorption forces (Figure 3.1). It also contains a net electric
charge and it is within this layer that the charge transfer reaction between the solu-
tion and the electrode is assumed to take place. This layer is thin however there is
an appreciable electrical capacitance coupling the diffuse zone to the electrodes. This
capacitance is known as the fixed layer capacitance (CF) whose typical values are
in the range of 5–50 µF/cm2 and this capacitance usually dominates the electrode
impedance at frequencies above 30–100Hz.
3.3.2 Diffusion layer

This is adjacent to the fixed layer on the solution side and is considered to be simi-
lar to the rest of the solution, except that any net charge in the fixed layer creates
an electric field which unbalances the positive and negative ion concentrations in the
20
Chapter 3. Physicochemical basis of the Induced Polarization method
zone (Figure 3.1). It has the thickness given by the Debye length, which for room tem-
peratures and .001N salinity is about .01µm. Using a simple model for the conduc-
tion phenomenon, the diffusion layer ions can be assumed to move as point charges
through a viscous medium under the influence of imposed electric fields or ion con-
centration gradients or both. The potential drop across the diffuse layer is known as
the zeta potential, the effects of which can be directly observed. The value of the Zeta
potential is, however, indirectly measured. Note, however, that at high frequencies,
when the separation between sites is large compared to the diffusion distance, this
acts just like a Warburg impedance. At low frequencies, the impedance acts as a ca-
pacitive impedance and is due to the accumulation of the reaction products between
the reaction sites.
Figure 3.1: Potential across the fixed and diffuse layers with (a) current into and (b) out of
the interface ( (Sumner, 1976)
21
3.4. Origin Of Induced Polarization
3.4 Origin Of Induced Polarization

The phenomenon of induced polarization and its electrochemical mechanisms are ex-
tremely complex. Nevertheless, attempts have been made through recent research
to establish the important phenomena that cause induced polarization in rocks/soils.
These include diffusion of ions next to metallic minerals and ion mobility within pore–
filling electrolytes (Sumner, 1976). Most authors agree that these interactions take
place at the contact of solid particles and electrolytic solutions which exist in the
ground, when a current flow is applied. The surface charges of these particles induce
in the solution an ion concentration of opposite sign close to the interface. A double
electric layer is formed in which the ions are impeded to be mobilized. When a cur-
rent is turned off, the initial equilibrium is re–established and the energy consumed
in the modification of the occurred interactions is restored. This storage of chemical
energy takes place because the mobilities of various ions through the rocks vary from
point to point. When a current is applied across such a rock, excesses or deficiencies
of certain ions will occur at the boundaries between the zones with different mobili-
ties. The concentration gradients thus developed oppose the current flow and cause a
polarizing effect. The exact causes of induced polarization are still unclear but most
probably, induced polarization results from the combined effects of several physical
chemical processes. However, there are two main mechanisms that are reasonably
understood:
• Grain (electrode) polarization (overvoltage)
• Membrane (electrolytic) polarization.

They both occur through electrochemical processes.
3.4.1 Electrode polarization (Overvoltage)

In the geological situation, current is conducted through the rock mass by the move-
ment of ions, within groundwater, passing through interconnected pores or through
the fracture and micro crack structure within the rock. If the passage of these ions
is obstructed by certain mineral particles which, like common metals, transport the
current by electrons, ionic charges pile up at particle–electrolyte interface, positive
ones where the current enters the particle and negative ones where it leaves as in
(Figure 3.2) . The piled up charges create a voltage that tends to oppose the flow of
electric current across the interface and the particle is said to be polarized. When the
current is interrupted, a residual voltage continues to exist across the particle, due
to the bound ionic charges, but it decreases continuously as the ions slowly diffuse
back into the pore electrolytes, which process gives the induced polarization effect
(Parasnis, 1966).
22
The induced polarization effect opposes the build up and collapse of the primary
potential difference, and is referred to as the overvoltage because an extra voltage
above that required to overcome the ohmic resistance is required to drive current
through the ore zone. The secondary voltage Vs that must be overcome when the
current is switched on is the same as the residual value that the potential falls to
at switch off (Griffiths and King, 1965). The level of IP is controlled by a number of
factors, as each blocked pore becomes polarized the number of pores, hence, the state
of dissemination and the volume of the ore will have a large effect. IP is greatly influ-
enced by the total surface of the polarizable material. A disseminated mineralization
has a lot of surface (possibly much more than a massive mineralization) and there-
fore, depending on polarization properties may have a larger IP response than its
massive counterpart. However, many massive bodies have a disseminated halo thus
increasing the overall surface. High porosity and high groundwater conductivity re-
duce IP as both lead to a short circuiting of energizing current through unblocked
paths (Griffiths and King, 1965).
Foremost among the ore mineral having an electronic mode of conduction and
therefore exhibiting strong IP are pyrite, pyrrhotite, chalcopyrite, graphite, galena,
bornite, magnetite and pyrolusite.
Figure 3.2: Grain electrode polarization (Reynolds, 1997)
3.4.2 Membrane (electrolytic) polarization

This type of polarization explains the IP effects that are observed when no metallic-
type minerals are present in the ground. This type of polarization applies to the
soil material under this study. There are two causes of membrane or electrolytic
polarization:
• By constriction within a pore channel

There is a net negative charge at the interface between most minerals and pore
23
3.4. Origin Of Induced Polarization
fluids. Positive charges within the pore fluid are attracted to the rock surface
and build up a positively charged layer up to about 100µ m thick, while negative
charges are repelled (Figure 3.3) . Should the pore channel diameter reduce
to less than this distance, the constriction will block the flow of ions when a
voltage is applied. Negative ions will leave the constricted zone and positive
ions will increase their concentration, so producing a potential difference across
the blockage.
Figure 3.3: Membrane polarization associated with constriction between mineral grains
(Reynolds, 1997)
When the applied voltage is switched off, the imbalance in ionic concentration
is returned to normal by diffusion, which produces the measured IP response.
• Presence of clay particles or filaments of fibrous minerals

Both of these tend to have a negative charge. In the absence of conductive
minerals, IP owes its origin to the existence of clay particles contained within
the pore structure of the rock (Parasnis, 1973). The surface of clay particle is
negatively charged and thus attracts positive ions from the electrolytes present
in the capillaries of a clay aggregate. An electrical double layer is, therefore,
formed at the surface of the particle as shown in Figure 3.4, the concentration
of the positive ions being the greatest at the surface of the clay particle. If the
positively charged zone persists far enough into the capillaries, it effectively
repels other positively charged ions and so acts like an impervious membrane
impeding their movement through the capillaries.
Clay minerals have ion–exchange capacities which means that they have cations
in ion–exchange positions in the lattice. When they electrolyze (ionize) the ex-
changeable ions go into solution, leaving behind the clay particle, which carries
a net charge. This clay particle then forms a highly charged immobile anion,
blocking free ion flow through the pore in which it may be situated. This causes
an imbalance in ion concentration along the double layer formed. When an elec-
24
Figure 3.4: Membrane polarization associated with negatively charged clay particles
(Reynolds, 1997)
tric current is forced through the clay, the positive ions are displaced (in fact,
their displacement constitutes part of the current) and on the interruption of
the current the positive charges redistribute themselves in their former equi-
librium pattern. The process of redistribution manifests itself as a decaying
voltage between two electrodes in contact with the clay as an IP effect.
It might be expected that the amount of polarization in clay bearing rocks and
soils increases in direct proportion to the ion exchange capacity. However, this
is not true because from experiments done on clay-rich rocks such as shale
show relatively little ability to polarize in comparison with siltstone which has
a lower content of clay minerals (Keller and Frischknecht, 1966). The expla-
nation to this is that in a rock containing a large proportion of clay, almost all
the negative charges may fixed in exchange positions within the lattice with
virtually no anions existing in solution. This causes the amount of charge ac-
cumulating at a potential barrier in a rock to be small because there are no
anions.
The nature of surface conduction in silicates including some clay minerals is

a phenomenon that is less known but very important in water bearing rocks.
Rock–forming minerals usually fracture in such a way that one species of ion
in the crystal is commonly closer to the surface than others. In silicates, the
oxygen ions are usually close to the surface. When an electrolyte is in contact
with such a surface, it will seem to the ions in solution that the surface is
negatively charged . Cations will be attracted to the surface by coulomb forces
and adsorbed, while anions will be repelled. The same effect takes place with
water molecules, which are polar. The water molecule is not symmetrically
25
3.5. Methods of measurement of IP effect
constructed from an oxygen atom and two hydrogen atoms; rather, there is an
angle of 1050 between the bonds from the oxygen to the two hydrogen atoms.
As a result, the center of mass of the positive charges (the hydrogens) does
not coincide with the center of mass of the negative charges (the oxygen atom).
Consequently, when viewed from the oxygen side, the water molecule appears
to carry a negative charge; and when viewed from the hydrogen side, it appears
to carry a positive charge (Keller and Frischknecht, 1966). When an electrolyte
is in contact with such a silicate surface, several layers of water molecules will
become adsorbed to the surface. This layer may be several molecules thick since
each layer of oriented water molecules absorbs other water molecules. A single
adsorbed cation will neutralize a surface charge which would hold several water
molecules in a chain. So the conductivity of water in this oriented adsorbed
phase is higher than the conductivity of free water and so contributes to the
overall conductivity of a rock.
It should be noted, however, that the increased pressure in the adsorbed layers
increases the viscosity of water and decreases the mobility of the ions. This
means that if many ions are adsorbed, the conductivity of the electrolyte may
be significantly reduced.
3.5 Methods of measurement of IP effect

In theory, induced polarization is a dimensionless quantity whereas in practice it is
measured as a change in voltage with time or frequency. The time and frequency IP
methods are fundamentally similar, however, they differ in a way of considering and
measuring electrical waveforms. In the former, a direct current is applied into the
ground, and what is recorded is the decay of voltage between two potential electrodes
after the cut off of the current (time–domain method). In the latter, the variation
of apparent resistivity of the ground with the frequency of the applied current is
determined (frequency–domain method).
In another type of frequency method, which is called Complex Resistivity (CR)
method, a current at frequency range (0.001 Hz to 10 kHz) is injected in the ground
and the amplitude of voltage as well as its phase with respect to the current is mea-
sured. That is a phase–angle IP measurement. It is important to note that many CR
implementations are measuring the real and quadrature component of the response.
Of course it is very easy to transform these two parameters into amplitude and phase
as shown in Section 3.5.3.
In Laboratory experiments, either soil/rock samples are used or scale models of
typical field situations are simulated as analog models for the measurement of the
26
IP response. A basic response parameter measured in the IP method is the amount

of change of voltage (∆V ) or resistivity (∆ρ) seen as a function of either time or fre-
∆ρ
quency. The fundamental basic IP function can be written as ∆V V or ρ ; this being
the basic IP response ∆V or ∆ρ normalized with respect to Voltage V or resistivity ρ
in order to form an independent IP parameter.
3.5.1 Time domain method

One measure of the IP effect is the the chargeability m (Siegel, 1959) and is usually
expressed as :
Vs
m= (mV /V ) (3.1)
Vp
where Vp is the ON–time measured MN voltage, and Vs is the OFF–time measured
MN voltage where AB is the current injection dipole and MN is the voltage mea-
suring dipole. Instrumentally, it is extremely difficult to measure Vs at the moment
the current is switched off due to electromagnetic effects which produce a transient
disturbance on switching , so it is measured at specific times (e.g 0.5s) after cut off.
Measurements are then made of the decay of Vs over a very short time period (0.1s)
after discrete intervals of time (also around 0.5s). The measured parameter in the
time domain is the area under the decay curve of Voltage V(t) corresponding to the
time interval (t1 , t2 ). The integration of these values with respect to time gives the
area under the curve (Figure 3.5) , which is an alternative way of defining the curve.
When the integral is divided by Vp , the resultant value is called the chargeability (m)
and has units of time (milliseconds). It is expressed as:
1
Z t2 A
m= Vs(t) dt = (3.2)
Vp t1 Vp
where Vs is the OFF–time measured MN voltage at time t, and Vp the observed

voltage with an applied current. Note, however, the major advantage of integration
and normalizing by dividing by Vp is that the noise from cross–coupling of cables
and from background potentials is reduced. Care has to be exercised in selecting
appropriate time intervals to maximize signal to noise ratios without reducing the
method’s diagnostic sensitivity.
27
Figure 3.5: The integrated decay voltage used as a measure of chargeability m

(Reynolds, 1997)
3.5.2 Frequency domain method

IP effects can be observed using observations in the frequency domain. The IP decay
characteristics observed in time domain can be transformed to the frequency domain
using Fourier techniques. Frequency domain measurements are made at two differ-
ent frequencies which are usually less than 10Hz (e.g 0.1 and 5Hz, or 0.3 and 2.5 Hz).
The measured parameter is the steady state voltage response after filtering and the
derived parameter is the Frequency Effect (FE) which is expressed as:
Vlo − Vhi
FE = (3.3)
Vhi
OR
ρlo − ρhi
FE = (3.4)
ρhi
where Vhi and Vlo are the stead state voltage responses at the filtered high and low
frequencies respectively. Since the current is held at a constant peak amplitude while
varying the frequency, the FE can as well be expressed as in Equation 3.4, where
ρhi and ρlo are the respective magnitudes of apparent resistivity at frequencies hi
and lo. The apparent resistivity at low frequency (ρlo ) is greater than the apparent
resistivity at a higher frequency (ρhi ) because the resistivities of rocks decreases as
the frequency of the alternating currents is increased. The two apparent resistivities
are therefore used to determine the Frequency Effect (FE) (unitless) which can be
expressed, alternatively, as the Percentage Frequency Effect (PFE) (Wait, 959a).
P F E = F E ∗ 100% (3.5)
28
It should be noted that although FE and PFE are defined in terms of two frequen-
cies, modern frequency domain IP surveys are often done either at single frequency
(resistivity and phase measurements) or at larger number of frequencies (e.g Zonge
offers 5 frequency schemes)
Metal factor (M.F) is another parameter in frequency domain method, which was
suggested by Madden to correct (partially) for the resistivity of the country rock (Mar-
shall and Madden, 1959). It is calculated from FE magnitude to compensate for
variations in effective resistivity of the host rock including changes in electrolyte,
temperature, pore size that are unrelated to the capacitive effect of the metal content
(Telford et al., 1976). It is defined as the FE divided by the low frequency apparent
resistivity (ρlo ). However, as the number, thus obtained is inconveniently small, it is
multiplied by 2π ∗ 105 , so that the practical definition of the metal factor becomes:
(ρa0 − ρa1 )
MF = A = A(σa1 − σa0 ) (3.6)
ρa0 ρa1
where ρlo and ρhi are the apparent resistivities whereas σlo and σhi are the appar-
ent conductivities at low and higher frequencies respectively; ρlo > ρhi and σlo < σhi ;
and A = 2π ∗ 105 .
If the resistivities are expressed in Ω m, the dimensions of the metal factor are
Ω−1 m−1 , which are those of electric conductivity.
In the field, the IP electrode system is the same as the resistivity system, so
electrode arrays such as Wenner, Schlumberger, Pole–dipole and double dipole are
used. IP field measurements always incorporate resistivity measurements.
Noise sources, besides spontaneous polarization (SP) which is easily compensated
for, includes telluric currents, capacitive couple (due to current leakage between cur-
rent electrodes and potential wires, or between current and potential wires), electro-
magnetic coupling (due to mutual inductance between current and potential wires)
and the IP effect from barren rock (Telford et al., 1976). Other sources of noise in-
clude fences, power lines, pipelines and other extensive man made structures with
earthed grounds which produce spurious induced polarization anomalies by acting
as secondary sources and sinks of current. IP is now a well established technique in
exploration geophysics, especially for base metal exploration, as indicated by the ex-
tensive use of the metal factor parameter. Both time domain and frequency domain
IP are commonly used .
3.5.3 Complex Resistivity (Spectral IP)

This involves taking amplitude and phase measurements, which in IP are defined as
the difference in phase angle between the received polarization voltage signal and the
29
stimulating current signal by Ohm’s law, in the case when both are sinusoidal wave-
forms. If the input current is a square wave, the phase measurement is defined as
the phase angle between the fundamental harmonic of the transmitted and received
signals.
The phase angle is defined as the angle, whose tangent is the ratio between the
imaginary and real components of the received voltage (V) or resistivity and is ex-
pressed as:
V imag. ρimag.
φ = tan−1 = tan−1 (3.7)
V real ρreal
The in–phase (real) and the out–of–phase (imaginary/quadrature) components

are more easily measured with modern electronic systems than the phase angle alone
and the two components provide additional useful information about IP phenomena.
The magnitude of the complex resistivity (|Z(ω)|) and phase (φ) of the polarization
voltage are measured over a wide range of frequencies (0.3 to 4 kHz) of applied cur-
rent. The frequency dependency is usually plotted as a binary function in the form of
logarithms to base 2 rather than to base 10. The magnitude is determined from the
real and imaginary components of the above mentioned parameters and is expressed
as:
p
Z(ω) = Im2 + Re2 (3.8)
It has been shown by Zonge et al. (1972) that IP phase angle and frequency domain
measurements correlate very well however the difference lies in the fact that by con-
ventional frequency domain IP method, the results are presented in various ways but
it is basically only the change in complex resistivity with frequency that is employed.
The complete theory about induced polarization is a subject of further study, how-
ever, the chapter has presented what is known so far in regard to the physicochemical
basis of the method. An attempt to explain the electrochemical nature of this method
is made by use of electrode impedance. This is made possible by use of an equivalent
circuit. There are two layers that make up the zone in the immediate vicinity of the
electrode–solution interface. These are the fixed and diffusion layers. The origin of
induced polarization is complex but two known mechanisms have been presented.
Electrode polarization (overvoltage) and membrane (electrolytic) polarization both of
which occur through electrochemical processes.
There are three methods of measurement of IP effect that have been discussed.
Chargeability being the measured parameter in the Time domain measurements as
the Percentage Frequency Effect is for the Frequency domain measurements. Com-
plex Resistivity (CR) is the third measurement method in which the magnitude of the
complex resistivity and the phase of polarization voltage are measured over a wide
range of frequencies.
30
The laboratory experiment for which this study is for, applied one of the measure-
ment methods (time domain method) on the soil samples. But where were these soil
samples obtained and which factors influenced the choice of the sampling sites, is a
question that can be best answered by the next chapter.
31
32
Chapter 4
Study areas and soil samples
4.1 Introduction
Study areas are places where soil samples were collected. A description of such places
and their soils is the basis of this chapter. In the selection of the soil sampling sites
for this research, important consideration was given to the geology, soils and climate
of the study area. The sampling locations had to be rich in all the three-targeted
mineral types, namely kaolinite, smectites and illites especially for Spain and Kenya.
4.1.1 Spain
The Spain study area shown in Figure 4.1 is located between longitudes 40 310 1000 E
and 40 510 1000 E and latitudes 370 000 0400 N and 370 100 0400 N. It is about 45 km north of
Malaga in the southern part of Spain and covers the municipalities of Antequera,
Campillos and Molina. It spans an area of approximately 30*18 km2 with varying
topography. It is easily accessible by road.
4.1.2 The Netherlands

The Netherlands study area (Enschede) lies in the eastern part of the country
(Figure 4.2). It is located between longitudes 60 500 E and 60 550 E and latitudes
520 100 N and 520 150 N.
4.1.3 Kenya
The Kenya study area (Nairobi) shown in Figure 4.3 is bounded by 10 000 to 10 300 S
latitudes and 360 300 to 370 300 E longitudes. It lies in the central part of the country.
33
4.1. Introduction
Figure 4.1: Location of study area (Spain) (Dizon-Bacatio, 1992)
Figure 4.2: General map of study area (The Netherlands) (Magellan, 1992)
34
Chapter 4. Study areas and soil samples
Figure 4.3: General map of study area (Kenya) (Fairburn, 1963)
4.2 Climate
Climate is a composite concept, which includes the temperature, rainfall, humidity
and evapotranspiration among other atmospheric variables. Soils and other near
surface features are usually determined not only by the current climate but also by
the paleo–climate. These factors strongly determine the type of clay mineral formed
in an area. The temperature sets the speed of any chemical reaction; rainfall provides
a medium in which chemical reactions can take place as well as a medium for the
movement of soil constituents. Leaching of calcium carbonate in the soil is rainfall
dependant, the higher the rainfall the deeper the horizon where calcium carbonate
appears in the soil.
4.2.1 Spain
The study area falls within the Continental Mediterranean type of climate and has a
dry summer with an overall mean annual temperature of 17.60 C. The highest mean
temperature is in the month of August at 25.50 C while the lowest is in the month
of January with a mean of 7.50 C. The rainfall is concentrated between the months of
October to April. The month of December receives the highest rainfall averaging 76.8
mm with the lowest being in the month of August averaging 6.1 mm. These climatic
conditions have enabled the development of soils rich in clay minerals with a high
Si/Al ratio (2:1). Lack of enough rainfall has however led to the calcareous layer to be
shallow in the area which leads to less concentrations of bivalent cations (Ca), which
35
4.3. Geologic setting
are important for the formation of clay minerals with Si/Al ratio of 2:1.

The study area has a temperate maritime climate, with temperatures averaging 20 C
in January and 16.60 C in July. Average annual rainfall is almost 800 mm distributed
fairly evenly throughout the year.
4.2.3 Kenya
The study area has got a pleasant climate for most of the year with temperatures
ranging between 14 and 280 C, rainfall is moderate and fairly uniform throughout
the year and increases as one moves from the plains at Athi River to the Kikuyu
escarpment. Due to the high rainfall and the underlying volcanic rocks, areas to the
west of Nairobi City are fertile and well served by streams. To the east, the area is
mainly savannah grassland. Deep weathering of the volcanic rocks gives rise to an
extremely thick soil cover on the watersheds.
4.3 Geologic setting

4.3.1 Spain
The area under study (Antequera Spain) is situated in the Subbetic zone of the Betic
Cordillera, one of the mountain chains of the Iberian Peninsula that is character-
ized by extensive parallel mountain ranges. According to geological time scale, a
geosyncline that encircles the Spanish Meseta, a stable core of the Iberian Peninsula,
formed a marine environment in which the sediments were deposited over a very
long period of time. A re–orientation and rotation of the Iberian peninsula from the
Lias to Cretaceous was followed by an approach of the Iberian Shield to the African
shield. During this period, the rather strong forces applied to the geosynclinal sed-
imentary rocks surrounding the Meseta resulted in their uplifting and subsequent
erosion. The by–products of this process, the flysch, form part of the main sequence
of the study area.
Another period of compression followed in the Oligocene and Miocene resulting
in big overthrusts in which the Triassic evaporites played an important role acting
as lubricants. Upper Miocene till late Pleistocene was dominated by deposition of
weathered and eroded materials from the uplifted regions into the depressions. These
materials have been referred to as Molasse and dominate the study area. They are
generally coarse textured occurring as impure sandstone intercalated with marls.
36
During the periglacial period of the Quaternary, the deposition continued mainly
around the hills. Large footslopes from different materials, developed at the foot
of the undulating and rolling topography and are referred to as solifluction glacis.
The change of climatic conditions since the last geological event has resulted in more
stable slopes and the development of deep soil profiles on these glacis (Dizon-Bacatio,
1992).
The geological periods and events had an important bearing on the soil forming
processes and properties. This has resulted in 2:1 (silica: aluminium) type of clay
domination in the soils due to high concentration of bivalent cations such as Ca and
Mg during their formation. The geology of the area consists mainly of marls, lime-
stone, dolomites, gypsum and bioclastic sandstones.

The rocks that were formed during the Tertiary period are not generally visible on
the surface in The Netherlands, but their direction of dip is from the southeast to the
northwest. Rocks of this period include limestone, which outcrops in a few places in
the southeast and the east of the country (Meijer, 1994).
The surface of the Netherlands consists almost entirely of sediments deposited
in the Quaternary period. In the penultimate pleistocene ice age–the Riss or Saale
period ice sheets were pushed from Northern Europe on to The Netherlands. In
the river valleys of the present provinces of Utrecht, Overijssel and Gelderland (the
central Netherlands), which run from south to north, the valley walls were pushed
sideways and upwards. This resulted into a series of elongated ice ridges or push
moraines, which reach a height of more than 100 meters and are now a popular
recreational area.
During the ice ages, the North Sea coast line laid much further to the northwest,
due to a considerable drop in the sea level. After the ice had melted, an extensive
area in the north of the country was covered with ground moraine, consisting of sand,
gravel, glacial till, boulder clay and erratic boulders. Prior to the ice age, the Rhine
and Mass rivers had deposited thick layers of sand and gravel from the erosion of
Alps and the lower European mountain ranges into the central and southern parts of
the country.
After the ice sheets had disappeared, westerly and northwesterly winds, partly
blowing across the dry floor of the North Sea, dispersed sand over the Netherlands.
The sand largely covered the ground moraines, the ice pushed ridges and the river
deposits.
In the period that followed, the Holocene, the climate became warmer resulting
into the rising of the sea level. The North Sea flooded the west of The Netherlands up
37
4.4. Description of soil samples
to the line formed by Groningen, Utrecht and Breda. The cross section of the western
Netherlands shows quite clearly that the pleistocene strata dip towards the west. In
the low Netherlands, the subsurface consists of Holocene clays, peats and pleistocene
sands.
During the period of sea level rise, sand bars were formed in the West running
parallel to the present coastline. In the course of time, the sand bars were trans-
formed into low dunes, which are now referred to as old dunes. In approximately
AD 1000, dunes were once again formed on top of the old dunes. The younger dunes
are higher reaching about 20 to 40 meters above sea level. Behind the old dunes,
a lagoon formed, to which the sea could penetrate through gaps in the dune chain.
Initially marine clay was deposited in the lagoon but peat formation ensued as the
lagoon became increasingly shallow. However the peat was sometimes washed a way
by the sea which led to new deposits of marine clay.
Peat formation also occurred in a number of marshy areas of the high Nether-
lands. During the Holocene period, rivers deposited sediments especially sand and
clay, however this was largely restricted to the region of the major rivers.
4.3.3 Kenya
The geology of the study area can be described as succession of lavas and pyroclastics
of cainozoic age overlying a foundation of folded Precambrian schists and gneisses
of the Mozambique Belt. The crystalline rocks are rarely exposed but occasionally,
fragments are found in agglomerates derived from the former Ngong volcano. The
geology consists of mainly metamorphic and volcanic rocks namely crystalline rocks
(fine grained schists and coarse gneisses), granitoid gneisses, phonolites and phono-
litic trachytes, tuffs and lake beds.
4.4 Description of soil samples

The sampling was done by Mr. P. Kariuki 1 and the samples were kindly made avail-
able to this MSc research project.
4.4.1 Spain
The selection of these soil samples was aided by an existing soil database that has
been built over the years in the soil division of ITC 2 . These soil formations have
been described as consisting of mainly the 2:1 clay minerals and greatly affected
1
PhD student, Department of Earth Systems Analysis
2
International Institute for Geoinformation Science and Earth Observation
38
by the geological/geomorphological processes and the climate as well. The degree

of development of these soils has a high correlation with the physiographic position
and pedogenesis prevailing in the area. Rainfall is enough to wash lime into the soil
but not out of the soil. This has resulted in the soil being mainly calcareous (Dizon-
Bacatio, 1992). The soils have textures ranging from loamy sand to clay loam on the
surface, gravel clay loam to sandy clay loam, silty clay loam to clay in the subsurface
horizons, sandy clay to clay. Their colour ranges from dark/dark brown to reddish
brown and dark grey to dull yellowish orange.
There was no distinctive spatial sampling interval. The samples were obtained
from near surface at a sampling depth as shown in appendix B, which is above the
water table. Most of the samples were obtained from areas that are not densely
covered with vegetation and others using hand augers because the surface was hard.
These soil samples were labelled as 1–24 for the IP/Resistivity measurements.

As discussed in section 4.3.2, most of these soils contain sands,clays, and Peat. Cover
sandy soils were deposited in the period of the last ice ages (Pleistocene). The damp
sandy soils are used as meadowland, the high sandy soils as meadows and woodland,
whereas the medium high sandy soils are used as arable land. The parental material
of these soils is the material mixed up by sub soiling (humose sand, peat and pod-
zolized sand) overlying undisturbed sand (aeolian sand of Weichsel age)
(De Bakker, 1978). Their colour ranges from brownish for sands to dark colour for
peat. Soil samples under this category were labelled as 27–30.
4.4.3 Kenya
These soils were selected due to their known richness in both high and low swell
potential soils. Soils of this area are products of weathering of mainly volcanic rocks
under relatively high temperature and rainfall. The principle soils overlying the
trachytic rocks of the northwest part of the area include strong brown to yellow–red
friable clay and red friable clays with high humus layer overlying clay. These are
developed from lava, volcanic tuff and ash in humid region with rainfall of more than
1000 mm per year and give way southeastwards to red friable clays developed on
similar rock types in area where annual rainfall is 762–1000 mm. Where drainage
is impeded over the Athi plains, the soils are black to dark grey clays comprising
black cotton soils with calcareous and non–calcareous variants. They overlie the
Nairobi and Kapiti phonolites, which form impermeable strata. The soils that have
been sampled were formed under different conditions but contain clay minerals that
39
4.4. Description of soil samples
are responsible for swelling especially the 2:1 (Si/Al) ratio. For the experimental
procedure, the soil samples were labelled as 25-26 and 31-43 .
The general description of the study areas including location, climate and geo-
logical setting has been discussed. The climate has been described as an important
factor in the formation of the three types of clays (montmorillonite, illite and kaoli-
nite). The geological setting of the study areas has been characterized by geological
periods and events that had an important bearing in the soil forming processes and
soil properties. It has been indicated that presence of high concentrations of biva-
lent cations such as Ca and Mg has resulted into a 2:1 (Si/Al) type of clay dominance
in the soils during their formation. The samples that have been obtained from the
above mentioned locations were prepared in order to make them suitable for both
gravimetric and IP/Resistivity measurements. The next chapter shows the processes
and methods used.
40
Chapter 5
Data acquisition methodology
5.1 Introduction
A great number of laboratory measurements have been made by a lot of investigators
to study the basic phenomena of induced polarization using different methods. Under
controlled conditions it is possible to examine the influence of different factors such
as clay content, clay mineralogy, and moisture content on the IP effect. A set of
Laboratory measurements has been done on a suite of near surface soil samples to
study their IP response under varying conditions of clay mineralogy, clay content and
moisture content. As natural rather than synthetic samples were used, the range of
these parameters were limited by the composition of the available samples.
5.2 Sample preparation

Owing to the fact that the IP response of soil depends not only on the mineralogy and
moisture content but also on the texture (Klein and Sill, 1982), the sample prepa-
ration procedure was made in such way that the original structure of soils was not
distorted. To meet this, soil samples were taken with sample rings of known volume
and diameter. However, recent comparisons by Vanhala and Soininen (1995) of labo-
ratory measurements of spectral induced polarization with field results suggest that
sampling and packaging procedures have minor influence on the electrical response
of a soil, particularly the phase spectrum.
The undisturbed soil material from Spain was collected by using a metallic pF
ring of dimension (diameter 50 mm and length 50 mm) shown in Figure 5.1. pF
rings are sample stainless steel tubes that are smooth inside and out with a bottom,
which has got cutting edge for easier soil entry. The pF ring was hammered into
the ground using an impact absorbing hammer. The soil samples so obtained were
41
5.2. Sample preparation
properly labelled, tightly covered using red caps and insulating tape and then put in
a storage box.
The Netherlands samples were collected using a long Polyvinyl Chloride (PVC)
tube by hammering it into the ground using the same type of hammer. However, the
cutting edge at the bottom of this tube, was done by use of a knife before any soil
sample could be collected.
The Kenya samples were collected as loose sediments, and put into plastic bags.
Neither the pF rings nor the PVC tubes were used in the collection for this type of
soils.
Figure 5.1: pF rings
Sample tubes for the IP measurements were designed from PVC plastic tubes.
Several pieces of electrically inactive tubes (50 mm long and 40 mm wide) were cut
from a long PVC plastic tube. Figure 5.2 shows the already cut sample tubes with
red caps, which were used to prevent soil particles from spilling over.
Holes for potential electrodes were drilled in between the 15 mm and 35 mm marks
and temporarily closed with black insulating tape as shown in Figure 5.2 in order to
stop loss of soil material.
42
Chapter 5. Data acquisition methodology
Figure 5.2: PVC made sample tubes
5.2.1 Saturation of soil samples
In order to use the soil samples for a laboratory experiment, a simple procedure
was devised to saturate them using distilled water. This procedure was important
because the soil had lost its field moisture. Because the soil samples under study
were collected from different locations and by different methods, the procedure of
saturating them differed.
The samples that were collected from Spain were being kept in metallic pF rings,
and it was by use of these pF rings that the soil was put in a basin containing distilled
water. The pF rings were carefully opened, and at one end of the opening of every
pF ring, a piece of cloth was wrapped and fastened by use of rubber bands whereas
the second opening was left free as shown in in both Figure 5.2 and Figure 5.3. All
of these pF rings were then put in a basin containing distilled water to a level of
about 3 cm and left to saturate as shown on the left hand side of Figure 5.3. Since
the pF rings are metallic and therefore conduct electricity, the soils were resampled
and put into standard sample plastic tubes designed for the IP measurement. Re-
sampling involved removing the soil material from the pF rings to the plastic PVC
tubes for electrical measurements. For samples from The Netherlands, since they
were collected in longer tubes than the standard sample tube size, it was necessary to
resample the soil into the standard sample tubes. Both tubes were of the same diam-
eter but with different lengths, therefore, it was a matter of aligning them and with
43
5.3. Potential electrodes
Figure 5.3: Saturation process
gentle knock on the longer tube, the soil could just collect into the standard sample
tube. These were then wrapped with pieces of cloth and put in a basin containing
distilled water. The soils were then left to absorb water.
For samples from Kenya, there was a slight change in the procedure in that the
soil was first remolded into the specially prepared sample plastic tubes. This involved
taking the loose sediments from the big plastic bags by use of a metallic spoon, and
physically putting them in the sample tubes. Thereafter, the tubes were wrapped
with pieces of cloth and then put in distilled water to saturate as shown on the right-
hand side of Figure 5.3.
The process of leaving the above mentioned type of soils to absorb water took
seven days.
5.3 Potential electrodes

Non polarizable potential electrodes were locally assembled from transparent ball
pen plastic containers, pieces of sisal material, copper wires and saturated copper
sulphate solution (Figure 5.4). A transparent model (ball pen plastic container) was
used to allow control of the level of the copper sulphate solution and to ensure that
44
the copper wire is well immersed in it. Epoxy glue was used to fix the movable parts
of the plastic ball pen containers.
A piece of sisal material cut from a long one, which was being used to tie a paper
box, was used to make a porous contact between the potential electrode and the soil
sample. Saturated copper sulphate was then put into the ball pen plastic container.
To make sure that the copper sulphate solution is saturated, hot water was used
to dissolve the copper sulphate crystals until the color of the solution became dark
blue, however the solution was left to cool before use (Scintrex, 1981). In addition,
copper sulphate crystals were put inside the plastic ball pen containers before the
cold copper sulphate solution was put. A copper wire was then inserted into the
container full of copper sulphate solution and then connected to the measuring cables.
Small pieces of sisal material were kept as short as possible in order to keep the
contact resistance below 20 kΩ. The potential electrodes were kept damp and an at-
tempt was made to stop them from dripping, as the sample must not be contaminated
with the electrolyte. These Cu–CuSO4 non polarisable type of electrodes is known to
be stable and establish an almost non–polarizable, somehow low–impedance contact
between the sample and the measurement apparatus (Vanhala and Soininen, 1995).
Figure 5.4: Sample holder and potential electrodes
45
5.4. Geometric factor K
In order to properly use the electrodes, a CTU core sample holder shown in
Figure 5.4 was used. It has output terminals to which the potential electrodes were
connected and linked to the receiver by measuring cables. It has copper disks at its
ends, and for this experiment, acted as the current electrodes.
Wires transmitting current were then fixed to the current electrode terminals of
the CTU, which are just on the round disks of the core holder and then connected to
the ITC Laboratory Transmitter.
For the IP measurements, the samples were put in the core holder of the same de-
vice shown in Figure 5.4. This core holder though designed to accommodate mineral
exploration drill hole core sizes, was some how able to accommodate this soil sample
size because of the elastic springs around its ends. This would hold the sample dur-
ing an IP measurement and on change of the sample, one of the handles could just
be moved in the opposite direction to the other. The copper contacts were regularly
cleaned to ensure good electrical contact.
5.4 Geometric factor K

This is a constant numerical factor, which when multiplied by the measured resis-
tance will convert the resistance to the resistivity for a uniform medium. It is de-
pendant on the type of the electrode array and spacing used. It was essential to
determine the K because the experiment involved measurement of resistivity. The
sample tubes are considered to be cylindrical, and owing to the electrode array used,
the K is computed in the following way:
Consider a cylindrical sample tube as shown in Figure 5.5, (`) being the length
between the potential electrodes.
Figure 5.5: Cylindrical sample tube
46
The area of the circular end (A) of the tube is calculated by the following expres-
sion:
A = π(d/2)2 (5.1)
The following expression is then used to find the geometrical factor K:
π(d/2)2
K= (5.2)
`
Where d is the diameter of the circular end and ` is the distance between the potential
electrodes P1 and P2 .
For all the sample tubes, the ` was measured with a ruler and by use of Equa-
tion 5.2, the K values were obtained and are presented in Appendix B.
5.5 Determination of water content

This is a test that is intended to measure the mass of water contained in a soil sample
as a percentage of the dry sample mass. In this study, water content of the soil is
defined as the ratio expressed as a percentage, of the weight of water to the weight of
the dry soil in the given mass. It is required as a guide to classification of natural soils
(Franklin and Vogler, 1979) and is a useful property to know as it enables electrical
induced polarization to occur once a current is fed into the soil sample.
A standard operating procedure was followed to determine the amount of water
in the soil. This involved weighing clean and dry plastic (PVC) containers using
a Mettler top pan balance (dynamic range=400g) to the nearest 0.01g as shown in
figure 5.6.
The wet soil samples and their containers were weighed again to the nearest
0.01g. The sample tubes were opened, the soil left to dry in air for an overnight and
then weighed again the next day. IP measurements were made in between each of
these steps. Care was taken when drying the samples not to lose any material as a
spill over. The same procedure was repeated 15 times after an interval of 1 day after
each balance measurement.
In order to completely dry the sample, a gravimetric method was used in which
soil samples were heated and dried in an oven at 1050 C to a constant weight for 24
hours. The sample container was made up of PVC material and resisted the high
temperature for 24 hours. The samples were then put in a desiccator for cooling
before being weighed. This ensured that no moisture could be absorbed by the soil
during the cooling process. A balance of adequate capacity (Figure 5.6), capable of
weighing to a resolution of 0.01g of the sample weight was used. The dry weight of
the soil was used as the divisor in the calculation because it expresses the absolute
47
5.5. Determination of water content
Figure 5.6: Weighing process
quantity of soil present. The following mathematical expression was then used to
find the percentage of water that is contained in every sample.
Mw
W = ∗ 100 (5.3)
Ms
Where Mw is the mass of the water
Ms is the mass of the dry soil.
In the same experiment, the degree of water saturation was determined by taking
the mass of water (Mw ) as obtained from above and dividing it by the total mass of
water contained in the pores at saturation. Since 1 g of water is equivalent to 1 cm−3
of water, then the expression for the degree of water saturation (Sw ) becomes
Vw
Sw = (5.4)
Vtotal
where Vw is the volume of water and Vtotal is the total volume of water in the pores
(Franklin and Vogler, 1979).
Volume of cylinder is given by:
V = π ∗ r2 ∗ h (5.5)
Other measurements involved the determination of porosity for all the soil sam-
ples. Porosity is defined as the ratio of the total pore volume that is not occupied
48
by the solid constituents to the total volume of a representative soil sample. It is a

dimensionless quantity and expressed as either a decimal fraction or a percentage.
Porosity measurements were made on the soil samples by taking volume of water in
the pores of samples at saturation and then dividing it by the volume of the sample
tube as determined by Equation 5.5. In particular the following expression was used
to compute for the porosity (Φ)
Vwater
Φ= ∗ 100 (5.6)
Vtotal
Where Vwater is the volume of water contained in the soil sample when saturated,
and Vtotal being the bulk soil sample volume (Franklin and Vogler, 1979).
5.6 Instrumentation
The laboratory measurement system used in this study was the ITC Laboratory
Transmitter and IRIS Elrec–6 Receiver (Figure 5.7) units employing the
four–electrode technique similar to that used by (Vanhala and Soininen, 1995). These
are respectively a laboratory transmitter and a digital multichannel IP receiver
adapted to small–scale laboratory measurements of induced polarization.
PC
30uA to 30mA
CH1
Elrec-6 Receiver
ITC IP-TX
_
+
Lab. Transmitter Current output Cu-CuSO4
electrodes
Tx- &Tx+=current electrodes

Tx- Tx+
Sample holder
Figure 5.7: Schematic representation of the IP instrumentation set up
Figure 5.7 shows an illustration of the instrumentation employed for the time
domain IP measurements in the laboratory. The ITC Laboratory Transmitter was
modified to suit laboratory work in addition to the field measurements of small
49
5.6. Instrumentation
Figure 5.8: IP instrumentation set up
A–MN–B separations. The A–MN–B refers to the arrangement of electrodes where

AB are current electrodes between which, are the potential electrodes indicated here
as MN.
It consists of two parts:
(a) ITC IP–TX Transmitter which generates a square pulse electric current that is fed
to the current electrodes AB. This transmitter has the following specifications: a crys-
tal controlled timer and a relay driver where the On+=Off=On-=Off=2.000 sec with
the ’On’ time and ’Polarity’ relays switching signals. The polarity is automatically
changed between the pulses. The waveform shown in Figure 5.9 indicates how the
current is usually transmitted. The pulse times as indicated from above are T=2.000
sec. It has a 7 steps current regulator with a nominal range from 30 µA to 30 mA
with 9 defined selector switch positions: Battery, Off, 30 µA, 100 µA, 300 µA, 1 mA, 3
mA, 10 mA, 30 mA. The transmitter is powered by batteries with an output voltage
of nominally 18 volts because it is made with batteries of 6 Volts.
(b) IRIS Elrec–6 Receiver is a six dipole multiwindow Time domain Receiver de-
signed for induced polarization DC electrical measurements (Figure 5.10). It is pow-
ered by either six 1.5 D size alkaline dry cells or one 12 V external battery or two
internal rechargeable 6 V batteries connected in series (=12 V). In the time domain
mode, and besides the classical logarithmic and arithmetic sampling modes of charge-
ability, the programmable mode offers 10 fully independent IP windows. It measures
the voltage between the receiving electrodes and displays the apparent resistivity
and the chargeability values. The measurement is made fully automatically through
50
Current
ON+ ON+
+I
OFF OFF
Time
T = 2s
T T T T
-I (T = Time)
ON-
Figure 5.9: Transmitted time domain current wave
the control of a micro processor that does the automatic self potential correction, the
digital stacking for signal enhancement and the error display in case of procedure
troubles. A real time clock introduced in the instrument permits the recording of
each reading with the date and the time of the measurement.
Figure 5.10: IRIS Elrec–6 receiver
51
5.7. Laboratory techniques and methods of
measurement
5.7 Laboratory techniques and methods of

measurement
5.7.1 Time domain method

The IP time domain measurement mode was used in this experiment. After the set up
of the experiment as shown in Figure 5.7, the Transmitter had to be calibrated by the
following procedure: First the battery voltage reading was recorded using the digital
voltage meter (DVM) set at voltage 20 VDC when the switch is put in Battery position
as shown in Figure 5.11 . Secondly the current output was measured using the same
DVM but using the Ammeter section. The switch was progressively put on each suc-
cessive current settings and the value was read and recorded with strong attention
to the DVM settings at a bigger scale when the range is getting exceeded. When all
that was done the system was then connected for IP measurements as shown in Fig-
ure 5.8. The calibration of the transmitter was done before every measurement and
the values of voltage and current written down (Appendix B, Table 6. The regulator
was put in Off position and then two jumpers were connected from the regulator to
the back of ITC IP–TX. The terminals in the front of the ITC IP–TX were connected
to current electrodes of the sample holder as shown in figure 5.11. An attempt was
made to keep the cables as straight as possible to minimize electromagnetic coupling.
Before the actual IP measurements could be made, the Elrec–6 Receiver had to
be properly set. The following operations had to be made after the setting up of the
experiment and these included: a BATTERY CHECK in order to check the voltage
of the internal batteries (IRIS, 1998). This was done when the unit (Elrec–6) had
been switched on. For proper operation of the instrument , the voltage of its dry cells
have to be more than 5.6 V.
In order to select the sampling I.P decay curve, the mode that consists of four
types of sampling of Time domain I.P decay curves had to be set. The four modes
include the Arithmetic, semi logarithmic, logarithmic and programmable. For this
particular study, the semi logarithmic mode was selected and according to the
charging time which is 2000 ms, this mode shows 10 I.P windows as shown in fig-
ure 5.12. Table 5.1 shows the window widths (ms) for the semi logarithmic mode
with charging time of 2000 ms.
Due to the fact that for any I.P measurement a certain configuration of electrodes
has to be used, it was important to set the E–array which is the electrode configura-
tion. For this experiment, the electrode array was described amongst the ”OTHER”
category which includes all those electrode arrangements that are not incorporated
into the instrument by the manufacturers.
52
Figure 5.11: Calibration of the transmitter
Other important settings were under the configuration menu which concern the
type of IP values, type of readings, sign of voltage, type of grid unit and set date
options. Raw values of IP option was selected because it gives the true chargeabil-
ity values during the measurement as opposed to normalized value option that gives
chargeability values according to a standard IP decay curve. The type of readings
chosen was the running readings whereby the values displayed during the mea-
surements are the average values of the three last consecutive pulses. On the other
hand, the cumulative readings show values that are the average of all the pulses re-
ceived from the beginning of the measurement. For the voltages, it was suitable to
53
measurement
Figure 5.12: Decay curve sampling mode
Table 5.1: Window widths (ms) for the decay curve semi logarithmic mode
TIME = 2000 ms
MODE V_DLY M_DLY 1 2 3 4 5 6 7 8 9 10
SEMI LOGARITHMIC 1260 80 80 80 80 80 160 160 160 320 320 320
choose the Unsigned voltage option because it gives the primary voltages at the
end of the measurement in absolute value regardless of the polarity i.e they will al-
ways be positive. For the signed voltage option, the voltages are signed with respect
to the polarity of the dipole and consequently, the resistivity values are also signed.
The metric grid was selected as the type of grid unit. This means that some param-
eters were to be introduced in meters such as the geometrical factor (K) and in this
case, the apparent resistivity value was computed by the system in Ohm–meters.
The above described operations do not change so frequently if the experiment is
carried out under similar conditions. However, there other operations that have to be
changed for every measurement because they are sample specific. This leads to the
description of the procedure of the measurement of induced polarization effect.
After putting the soil sample in the sample holder, both current and potential elec-
trodes were connected to the system. From the alphanumeric key board of the Elrec–
6 receiver shown in Figure 5.10, the geometrical factor (K) was introduced through
the SPACING option. The K values were obtained as described in Section 5.4. The
introduction of the K value at this point was very important because of the electrode
array which was selected with the E–ARRAY key and the need for the computation
of the resistivity values.
In order to ensure proper connection of the electrodes to the instrument channel
(Channel 1) of the receiver before current was transmitted, the ground resistance
54
Figure 5.13: Measurement process
of the electrodes were checked by use of the Rs CHECK key. It is recommended

to operate with the lowest possible values of ground resistances preferably around
4 k Ω though a value of 20 k Ω should not be exceeded. For this study, the contact
resistance was maintained below 20 k Ω.
The current being transmitted has to be entered into the system and this was
done through the SET UP KEY. In addition, the measuring domain (TIME) and
pulse duration were selected. The current was set in mA and the time was 2000
ms and constant throughout. The measuring domain was selected as Time T. wave
which permits to compute the global apparent chargeability and partial chargeabil-
ities by sampling the I.P. decay curve. This option uses a symmetrical Time domain
55
measurement
waveform curve: ON+, OFF, ON-, OFF as shown in Figure 5.9. In this domain, the
sampling rate is 10 ms with an accuracy in synchronization of 10 ms. The chargeabil-
ity values measured have a resolution of 0.1 mV/V with a typical accuracy of 0.6 %,
and a maximum of 2 % of reading 1 mV for Vp>10 mV. It was very important before
any measurement to control the level of noise by use of the MONITOR KEY. This
function features a voltmeter that can be used to display the value of voltage coming
from the soil sample in mV.
After making all those settings, measurements were taken by pressing the
START key and after seven pulses, the process was stopped and the readings
recorded in the memory of the receiver (Figure 5.13). After the experiment, the data
that had been stored in the memory of the receiver, would be transfered into the
computer using a serial link and a data transfer software utility provided with the IP
receiver as shown in Figure 5.14.
Figure 5.14: Data transfer to a PC
What was given as a result were chargeability values which are described in the
following way. In each I.P window (TM i : width), the value of the partial apparent
chargeability (Mi) is computed and these Mi values permit to obtain the weighted
value (M) of these partial chargeabilities (Mi).
At the end of the measurement, the average values of partial apparent charge-
abilities (Mi) are computed. The weighted average value (M) of these average partial
apparent chargeabilities is also computed.
Pnip
i=1 TM i .M i
M= Pnip (5.7)
i=1 TM i
56
Where
nip is the number of I.P. windows: up to 10.
TM i are the width of the I.P. windows, expressed in ms.
Mi are the average values of the partial apparent chargeabilities, expressed in mV/V
.
M is the weighted average value of the average partial apparent chargeabilities, ex-
pressed in mV/V )(IRIS, 1998).
It was observed that as the samples were becoming dry, the chargeability was
decreasing. The contact resistance potential electrodes were observed to increase
when the samples lost water.
5.8 Test measurements

• Ten pairs of potential electrodes were tested in terms of their contact resistance.
It was observed that potential electrodes with match sticks as their porous con-
tacts had contact resistances ranging between 100–350kΩ whereas those with
sisal material contacts were below 20kΩ.
• A test run was made to check the effect of a charger connection to the receiver
on the chargeability values as compared to its in–built DC voltage output. It
was observed that there was no change in the chargeability values given by the
same sample under the two conditions e.g sample 1 (Spain) at current of 0.29
mA gave a chargeability value of 0.29 mV/V with and without a charger.
• Measurements were made on the same sample to test the effect of different
values of current. The currents used were 0.03mA, 0.10 mA, 0.29 mA and 0.97
mA and the chargeability values given were 2.90 mV/V, 3.77 mV/V, 2.30 mV/V
and 2.03 mV/V respectively.
• Soil samples were weighed before they were prepared for IP/Resistivity mea-
surements, to check whether the dynamic range of the Mettler top balance is
not exceeded.
5.9 Sources of errors and their magnitude

When performing this experiment, sources of errors were identified and are outlined
as follows
1. The samples were dried in air (room temperature) when the caps have been
removed. The Oven was not used for the daily drying because the same samples
57
5.9. Sources of errors and their magnitude
were to be taken for another test (NMR) and it was thought that heating them
before that test would distort the structure of the clay mineralogy. This air
drying did not ensure even distribution of water in the samples giving rise to a
possible distortion on IP readings.
2. For the IP measurements, all samples were put in a CTU sample holder stand
(see Figure 5.4). After the measurements, they were removed and replaced.
The problem was that when these samples were still wet, when removing them,
they would leave some soil particles on the current copper contacts (current
electrodes)(Figure 5.16). A laboratory tissue was used to clean the contacts but
it was impossible to recover all the soil from the tissues back to the samples.
This gave errors in the procedure for the determination of the water content to
the magnitude of 0.80g. When the samples were dry, particles would spill over
onto the surface of the sample holder stand, a small brush was used to collect
them but it was impossible to get all of them.
3. The potential electrodes used were locally made, which caused problems. The
sisal material used as a porous contact between the electrode and the soil sam-
ple was dripping and as result contaminating the samples. This means that
when measurements are done on the following day, the resistivity and IP val-
ues may be different. Owing to this dripping nature, the contact resistance of
the potential electrodes was very hard to keep constant. The effect due to this
is that chargeabilities may change if contact resistances are not kept constant
(Scintrex, 1981). In order to quantify this error, samples were weighed before
being put in the experimental set up and then after. This gave an average 0.5
g of extra weight. The contact resistance of the potential electrodes was always
above the optimal ceiling (≤ 4 kΩ) but within the acceptable range of 0–20 kΩ
(IRIS, 1998).
4. The sample tubes were locally designed from a long Polyvinyl tube to the dimen-
sions mentioned in section 5.2. Cutting these sample tubes was not perfectly
done because they did not have flat ends. The ends of some of the tubes were
a little bit diagonal (Figure 5.15). The effect due to this is that when the ends
are not flat, there is no good current electrode contact. Errors were also given
by irregular end surfaces of soils themselves. For example sample number 9,
the end surface was irregular and the measurement taken gave with it gave a
chargeability value of 0.33 mV/V and after flattening the end surfaces, it then
gave a chargeability value of 1.56 mV/V. This is caused by heterogeneous flow
of electric current in the sample due to a localized contact.
5. When some samples were saturated with distilled water, they overgrew and
58
Diagonal cut
Sample tube inside the

holder
Current
electrode
Figure 5.15: illustration of the diagonal cut of the sample tubes
in an attempt to make the surface flat, the overgrown part was trimmed off.
On drying, these samples shrunk and became smaller, which provided a poor
current electrode contact.
6. The springs on the CTU–2 sample holder were not strong enough to hold the
size of sample that was used in this experiment. The problem was easily identi-
fied when improvising for the current electrode contact by using the aluminium
foil or the old two and half Dutch guilder coins. When the rods bearing the
springs are physically pushed towards each other (Figure 5.4 ), the voltage
output from the transmitter and observed with receiver is increased thus indi-
cating a better current electrode contact than before.
Excitation source
Current regulator
A
Ammeter
IP-Rx
receiver
soil sample
in its holder
Cu contacting electrode
Cu contacting electrode
Figure 5.16: illustration of the set up
7. Some IP measurements were carried out when the excitation was not current
regulated i.e the regulator not being in its linear range. This was being checked
by taking a spot check at the current electrodes (Figure 5.16) and making sure
that the voltage across them is less than 5 volts. Because of this error, the
59
5.9. Sources of errors and their magnitude
results obtained from samples (13, 28 and 30) were removed from the data set
of this study together with half of the results obtained from samples (9, 12 and
27). For the rest of the samples, the last two measurements that were taken
when the samples were getting drier were also removed. The requirement of
linearity for electrode polarization was only fulfilled with current (≤ 2.84 mA),
so IP measurements taken with current (9.07–25 mA) were not considered in
this study.
A table of the IP/Resistivity measurements reliability is presented in Appendix B

(Table 4). It shows what happened to each of the soil samples during the experiment.
Data acquisition is an important aspect of this research. The chapter has outlined
the main steps taken to prepare, saturate and weigh the soil samples. The IP instru-
mentation (ITC–IP Transmitter and Elrec–6 Receiver) has been described in regard
to use, calibration and maintenance. However, there were errors observed during the
experiment and these have been clearly stated. The experimental data that has been
obtained need to be analyzed in order to acquire more information about soil samples
, which is the subject of the next chapter.
60
Chapter 6
Experimental data and

processing
6.1 Introduction
With the main variable parameters of the response obtained i.e chargeability and
resistivity, an attempt was made to relate them with the other variables including
water content (%), clay content (%), grain size (mm) and CEC (meq/100g). The data
/results were displayed using graphs and ternary plots.
6.2 Water content

A standard operating procedure described in Section 5.5 was used to obtain the
amount of water contained in the soil samples. A set of data for sample 1 (Spain)
is shown in Table 6.1.
Table 6.1: Water content (%) values obtained with sample 1 (Spain)
Field number Lab. sample no.
1 WATER CONTENT DETERMINATION
st
1
reading at
saturation 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 14th 15th
Mass of container (g) 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12 17.12
mass of wet soil + Container (g) 113.61 113.03 111.91 110.67 108.50 107.67 107.07 105.14 104.13 104.12 103.54 101.7 100.68 99.42 97.35
Mass of wet soil (g) 96.49 95.91 94.79 93.55 91.38 90.55 89.95 88.02 87.01 87.00 86.42 101.7 83.56 82.30 80.23
Mass of Dry Soil + Container (g) 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35 97.35
Mass of dry soil (g) 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23 80.23
Mass of water (g) 16.26 15.68 14.56 13.32 11.15 10.32 9.72 7.79 6.78 6.77 6.19 21.47 3.33 2.07 0.00
Water content (%) 20.27 19.54 18.15 16.60 13.90 12.86 12.12 9.71 8.45 8.44 7.72 26.76 4.15 2.58 0.00
61
6.3. Analysis of IP response characteristics
It was observed that with some soil samples as shown in Appendix B (Table 4)
increased in volume on taking up water, and with a gradual loss of water, they de-
creased in size. This is analogous to clay soils which expand with increase in moisture
content whereas moisture loss is accompanied by shrinkage. However, the amount of
shrinkage is thought to be effected by the packing arrangement of the particles. It
was further observed that the water content of the samples varied depending on the
soil type and sampling site.
6.3 Analysis of IP response characteristics

The IP study of samples was developed in successive steps. Initially, a comparative
analysis was made of chargeability and resistivity of the different samples. This was
followed by a comparative examination of chargeability and water content, charge-
ability and CEC for fixed clay percentages, and finally chargeability and grain size.
This type of analysis provided information regarding the phenomenon of IP in soils
under study.
6.3.1 Chargeability and resistivity

From the time domain measurements were obtained readings for the chargeability
in mV per volt (mV/V) and resistivity in Ohm–meters (Ω.m). Figure 6.1 shows a
graphical display of chargeability (mV/V) against resistivity (Ω.m) at saturation. The
legend shows 6 classes of CEC values and the origin of samples. The CEC classes
were obtained through grouping the clustered points together after sorting them in
an ascending order. A group of sample points of the same origin (Spain) with CEC
classes (11–15 meq/100g and 16-22 meq/100g) occupy the lower part of the graph be-
tween 0–5 mV/V and 0–200 Ω.m The soil samples that are included in the above CEC
classes are number 21 with 27.4% clay content, 31 (71.5%), 6 (44.5%), 15 (33.2%), 5
(42.3%), 7 (37.0%), 14 (38.0%) and number 17 with 48.0% clay content, 34 (27.4%),
8 (44.6%), 3 (46.0%), 33 (70.7%), 23 (41.0%), 20 (50.4%), 25 (41.1%) respectively (Ap-
pendix B, Table 1).
In the upper part of the graph, there are samples with higher CEC classes domi-
nated by a single origin (Kenya). For examples, samples in the (23–26meq/100g) CEC
class include number 13 with 51.8% clay content, 42 (35.5%), 16 (48.0%), 4 (65.4%),
19 (63.7%), 18 (57.2%), 24 (46.1%),37 (56.0%) whereas in the class (27–30 meq/100g)
there are number 1 (57.7%), 2 (61.4%), 9 (67.9%) and 43 (53.7%). This group of sam-
ples is characterized by higher chargeability values (13–21 mV/V) but with the same
range of resistivity as the first group.
62
Chapter 6. Experimental data and processing
CEC
2-10 meq/100g (Kenya)
2-10 meq/100g (Spain)
25 27-30 meq/100g (Spain)
20
Chargeability (mV/V)
15
10
0 100 200 300 400

Resistivity (Ohm.m)
Figure 6.1: A plot of chargeability against resistivity
6.3.2 Chargeability and water content
Figure 6.2 shows the results of measurements done on samples 2, 17, 19 from Spain
and 27 from Holland. In general, their chargeability values increase with increasing
water content up to a maximum beyond which they decrease. Sample 2 attains a
maximum of 1.56 mV/V at 36% water content. It contains 36% of clay with a CEC
value of 19.0 meq/100g. It is a type of soil formed from lacustrine sediments (silt
(21.8%) and sand (16.8%) and was collected from a lacustrine plain. Sample 17 was
formed from colluvium material and collected from a hill land. It’s maximum charge-
ability (7.39 mV/V) is achieved at 23% water content. Its clay content is 48.8% with
16.8 meq/100g CEC. Other materials include silt (19.0%) and sand (32.2%). Sample
19 is formed from bio–sands and was collected from a piedmont. The value of charge-
ability obtained at its peak is 3.49 mV/V at 27% water content. The CEC value is 24.4
meq/100g and other materials are clay (63.7%), silt (10.3%) and sand (25.9%). And
63
lastly, sample 27 which is sandy has a maximum of 3mV/V at 30% water content.
Figure 6.3 shows another type of relation between chargeability and water con-
tent. For these samples, the chargeability increases with increasing amount of water
almost linearly. Sample 25 has the highest value of chargeability (22.28 mV/V). This
soil sample was formed from residual soils and collected from low lying land. It has
41.1% clay, 21.1 meq/100g CEC with silt (40.6%), sand (18.3%). Sample 36 attains
a maximum of 14.95 mV/V at 48% water. It was formed from residual soils and col-
lected from a hill slope. It is made of 52.3% clay, 4.3% silt and 43.5 % sand with CEC
value of 4.3 meq/100g. For sample 38, the clay content is 85.6%, silt (3.6%) and sand
(10.8%). However it has got a CEC value of 8.1 meq/100g. It was formed from resid-
ual soils and collected from a hill slope. Sample 39 was collected from an undulating
plateau, having been formed from residual soils. Its highest value for chargeability is
18 mV/V at 43% water content. Sample 10 has the lowest chargeability value of 5.42
mV/V at 38% water content in this group. It was formed from colluvium material and
collected from a piedmont. It has 71.6% of clay, silt (24.7%) and sand (3.7%). The
CEC value is 45 meq/100g.
For the rest of the samples, more graphs are presented in Appendix A and their
descriptions in terms of particle size constituents and parent materials in Appendix
B (Tables 1 and 2).
64
Soil samples
Sample 2 (Spain)
Sample 17 (Spain)
Sample 19 (Spain)
Sample 27 (The Netherlands)
6
10 15 20 25 30 35 40
Water content (%)
Figure 6.2: A plot of chargeability (mV/V) against water content (%)
65
Soil samples
Sample 25 (Kenya)
Sample 36 (Kenya)
Sample 38 (Kenya)
Sample 39 (Kenya)
Sample 10 (Spain)
25
20
15
10
0 20 40 60 80 100
Water content (%)
Figure 6.3: A plot of chargeability (mV/V) against water content (%)
66
6.3.3 Chargeability and clay mineralogy/CEC

In order to study the behaviour of chargeability with clay mineralogy which is char-
acterized here by CEC, a graphical display shown in Figure 6.4 was used. The legend
shows the clay percentage classes, which were fixed following a similar procedure as
described in Section 6.3.1
Figure 6.4 shows that for a fixed class of clay percentage, the chargeability in-
creases with increasing CEC. With respect to the origin of samples, Kenya samples
are represented by higher chargeability values than the Spain samples.
Clay classes
15-39% Clay (Kenya)
15-39% Clay (Spain)
40-49% Clay (Kenya)
40-49% Clay (Spain)
50-60% Clay (Kenya)
50-60% Clay (Spain)
61-85% Clay (Kenya)
61-85% Clay (Spain)
25
20
15
10
0 10 20 30 40 50
CEC (meq/100g)
Figure 6.4: A plot of chargeability (mV/V) against CEC (meq/100g)
67
6.3.4 Chargeability and granulometrical composition of

soils
Soils consist of an assemblage of discrete particles of various sizes and so the granulo-
metric composition is made up of separate ranges of sizes that have been determined
in relative proportions, by dry mass, of each size range. This grain size classification
has a direct application in engineering studies and hydrology.
In an attempt to analyze the relationship between chargeability and the granulo-
metrical composition, a triangular representation was used as shown in Figures 6.5.
The corners of the triangle represent the three different grain sizes for Sand (> 0.05
mm), Silt (0.002–0.05mm) and Clay (<0.002mm). Each corner of the triangle corre-
sponds to hundred percentage of the respective grain sizes. Samples plotted at the
sides of the triangle represent bimixtures of the two grain sizes while any sample
lying within the triangle represents a trimixture.
The chargeability of each sample was determined at saturation and plotted in the
equilateral triangle according to its grain size. The different symbols represent indi-
vidual soil samples with the three constituent particle sizes as shown in Figure 6.5.
The number written against each of the symbols is the chargeability value in mV/V.
In order to get a close examination, a triangular plot representing samples with
CEC values greater than 20 meq/100g was used (Figure 6.6). This was an attempt to
fix the mineralogy/CEC and investigate the effect of the individual particle sizes on
the chargeability.
Figure 6.6 was divided into two ternary plots according to the origin of the sam-
ples. It was observed that samples showed different behavior according to their origin
as illustrated below.
Figure 6.7 shows the relationship of chargeability and grain size distribution for
both Spain and Kenya samples. The plot shows a loose trend of increasing charge-
ability from very fine Clay (< 0.002 mm) to less fine fractions (Silt (0.002–0.050 mm)
and Sand (> 0.05 mm)) in the case of Spain. However, for Kenya samples, there is no
clearly defined trend between their chargeability and particle sizes.
68
CEC
Sand (>0.05mm) 27-30 meq/100g (Kenya)
0 100
20 80
40 60
3.3
7.2
3.9
3.6
13.4 1.8
60 7.3 40
15.0
4.5 2.2
18.3 1.8
12.5 1.0
1.1
3.8 2.3 1.7
80 2.9 16.5 20
21.1 2.2
12.5 2.8
17.3
6.8
22.3 2.8 0.4
6.9 17.7 19.7
3.9 3.5
13.9 11.5
100 0
14.1 5.4
0 20 40 60 80 100
Silt (0.002-0.05mm) Clay (<0.002mm)
Figure 6.5: Effect of grain size on the chargeability of soil samples
69
CEC
Sand (>0.05mm)
0 100
20 80
40 60
60 13.4 40
1.8
1.1
3.8 2.3 1.7
80 21.1 2.2 20
17.3
0.4
6.9 17.7
3.9 3.5
13.9
100 0
14.1 5.4
0 20 40 60 80 100
Silt (0.002-0.05mm) Clay (<0.002mm)
Figure 6.6: Effect of grain size on the chargeability soil samples
70
CEC CEC
23-26 meq/100g (Spain) 23-26 meq/100g (Kenya)
Sand (>0.05mm) Sand (>0.05mm)

0 100 0 100
20 80 20 80
40 60 40 60
60 40 60 13.4 40
1.8
1.7
3.8
80 20 80 21.1 20
2.2
17.3
0.4
17.7
3.9 3.5
13.9
100 0 100 0
3.0 5.4
0 20 40 60 80 100 0 20 40 60 80 100
Silt (0.002-0.05mm) Clay (<0.002mm)Silt (0.002-0.05mm) Clay (<0.002mm)
Figure 6.7: Effect of grain size on the chargeability of Spain and Kenya samples
71
6.4. Decay curve analysis
6.4 Decay curve analysis

In order to get more information from the IP response of these soil samples, the
exponentials of the individual decay curves were analyzed as suggested in the work
of Illiceto et.al (1982), Ogilvy and Kuzmina (1972) and Roussel, 1962. This was made
in an attempt to obtain parameters related to granulometry and water content.
The shape of a decay curve or the time dependence of chargeability M(t) may be
described as the sum of an exponential function of the form
n
X −t
M (t) = Aj ∗ exp( ) (6.1)
j=1
τj
where τ is the time constant of the relaxation mechanism.

The decay curve (Figure 6.8) was constructed from the chargeability values ob-
tained from the IP channels and Time windows. Decay curves that are shown were
obtained from soil samples at saturation and at intermediate water content.
1.6
Water content
38%
32%
28%
1.2
Chargeability (%)
0.8
0.4
0 400 800 1200 1600 2000

Time (msec)
Figure 6.8: Decay curves obtained from sample 10 from Spain at different water contents
The chargeability values were first converted from mV/V to % by dividing them
by 10 before plotting them. As mentioned in Section 5.7.1, the semilogarithmic mode
72
Table 6.2: Typical original receiver readings
V= 525.532 Sp= -12 I= 0.98 Rs= 10.43

Ro= 160.9 Ohm.m M= 2.24 E= 0.0
M1= 7.69 M2= 5.77 M3= 4.62 M4= 3.83
M5= 3.04 M6= 2.34 M7= 1.86 M8= 1.43
M9= 1.02 M10= 0.78
cycl= 8 Time= 2000 V_D= 1260 M_D= 80

T_M1= 80 T_M2= 80 T_M3= 80 T_M4= 80
T_M5= 160 T_M6= 160 T_M7= 160 T_M8= 320
T_M9= 320 T_M10= 320
of measurement with a time base of 2000 ms was used. This mode measures charge-
ability in 10 IP windows and these values, referred to as M1...M10 in Table 6.2, make
up the Y–axis of the decay graphs.
Consider Table 6.2, V is the average value of voltage, in mV, Sp is the Self Po-
tential at the beginning of the measurement, in mV, I is the introduced value of the
current, in mA, Rs is the contact resistance at the beginning of the measurement, in
kΩ, Ro is the resistivity value, in Ohm.m, M is the weighted average value of the par-
tial chargeabilities, in mV/V, E is the standard deviation of the chargeability value,
M1....M10 are the average values of the partial chargeabilities, in mV/V, V DLY is the
on time delay in ms, M D is the off time delay in ms, T M1....T M10 are the widths of
the IP windows, expressed in ms. The time represented on the X– axis corresponds
to the delay time plus half time interval of the first window (T M1) and then for the
next time window, it becomes delay time plus the time interval of the first window
plus half the time interval for the second IP time window. The same procedure is
repeated until the ten time windows are processed. The points effectively represent
a digitized decay curve.
In order to extend the IP study beyond the direct chargeability evaluation, decay
curves were analyzed using the following interpretation model .
3
X −t
m= Aj ∗ exp( ) (6.2)
j=1
τj
The constants Aj and τj were determined by least–squares non–linear regression

and the Levenberg–Marquardt version of this algorithm was used (Marquardt, 1963).
The three exponential model gives three values of co-efficients (A1 , A2 , A3 ) and
73
three time constants (τ1 , τ2 , τ3 ). The advantage of fitting a curve with three expo-
nentials is that it fits the data points well compared to a fit with one exponential as
illustrated in Figure 6.9
1.6 Curve fitting

Data points
1 exponential
3 exponential
1.2
Chargeability (%)
0.8
0.4
0 400 800 1200 1600 2000

Time (ms)
Figure 6.9: Illustration of fits of one exponential and three exponentials to the data points
(0) (0)
The initial estimates of Aj τj of the parameters in the model 6.1 were deduced
from a mean of values graphically determined from some decay curves. The set of
data that was considered for decay curve analysis involved 251 readings.
It was observed that some values of Aj and τj for 127 readings obtained from soil
samples were not reliably determined by the algorithm due to un regular distribu-
tion of current inside the sample due to a localized contact with current electrodes,
sample cohesion due to loss of water and lack of proper electrical contact during the
IP measurements . By applying a percentage of good fit, it was observed that a half
of the 251 readings had a fit of 99.998%, one sixth of the 251 had a moderate fit
(95.5%), whereas a a third of the readings were unreliable. An outline of the causes
for the unreliable determination for the three exponentials (Amplitude and decay
time constants) in regard to the individual samples whose part of the readings were
problematic, has been given (Appendix B, Table 5 )
However, for the remaining 124 readings, the values of the constants were deter-
mined and the following observations made:
1. The shortest time constants for τ1 , τ2 and τ3 were 25 ms, 199.1 ms and 500 ms
74
respectively.
2. The co-efficients A1 , A2 , and A3 were observed to vary as a function of water as

shown in Figure 6.10 and others in Appendix A.
Graphical analysis of the A1 constant and the amount of water was performed and
analyzed statistically for increasing or decreasing trends. The trends were tested for
significance using the F–statistic (Draper and Smith, 1981).
The F statistic is defined as the ratio
M SQr
F = (6.3)
σr
where M SQr = mean square due to regression; and σr = residual variance. The mean
square due to regression is the component of total variance that can be explained by
the linear trend. As such, trends with greater significance have a higher F. The sig-
nificance level value p was used in assessing the significance of the trends. And this
is the probability that the trend is deemed erroneously significant. The significance
of a trend increases as p becomes smaller.
Fig 6.10 shows a plot of A1 against water content for all the samples from the
three locations. The plot visibly indicates a trend of increasing A1 with an increase
in water content. To ascertain the significance of this trend, the A1 constant values
were regressed linearly on the water content values (SPSS., 1998). The F statistics,
F=168.635 and p value=0.000, were obtained, which means that the trend is highly
significant because the p value for the F is less than 0.0005 (Table 6.3).
Table 6.3: Results of statistical analysis
ANOVA
Model Sum of df Mean F Sig.

Squares Square
Regression 90.052 1 90.052 168.635 0.00
Residual 55.537 104 0.534
Total 145.589 105
a Predictors: (Constant), Water content
b Dependent Variable: A1
75
soil sample origin

Spain
Holland
Kenya
3
A1
0 20 40 60 80 100
Water content (%)
Figure 6.10: A plot of A1 against water content
76
CEC classes
4 2-10 meq/100g (Spain)
3 16-22 meq/100g (Spain)

2
A2/A3

1
-1
0 20 40 60 80 100
Water content (%)
1000
Figure 6.11: Ratio A2/A3 against water content
The Aj were found to vary as a function of water however the ratios between any
100
pair of the co-efficients were observed as not depending on the water content.
Resistivity (Ohm.m)
Therefore, an attempt was made to utilize these ratios for the purpose of differentia-
tion between soil types .
Figure 6.11 shows the ratios obtained from the analysis of decay curves from the
10
samples. The values of this parameter range between 0.5 and 3 for all samples. After
plotting various ratios of Aj as a function of water content, ratio A2/A3 was found
to 25have the least spread of values and represented a distinction between samples
according
1 to their origin (Figure 6.11).
Figure 6.12 shows a comparative analysis of the three parameters (chargeability,
20
electrical resistivity, ratio A2/A3) as a function of water content.

The following can be noted:
15
1. Samples cluster in groups according to their origin

10
2. The range of water content of samples from the same origin can be determined
5 from this graph. This information is also useful for the evaluation of porosity
of samples.
0
77
3. The clusters according to CEC classes occur in the same places with regard to
origin and it is not possible to isolate the different classes. However, there are
imperfect clusters that can be observed.
More graphs of the Aj ratios are presented in Appendix A.
4
CEC Classes
2-10 m e/100g (Spain)

3
2 16-22 m eq/100g (Spain)

A2/A3
16-22 m eq/100g (Kenya)
1 23-26 meq/100g (Spain)

0
-1
1000
Resistivity (Ohm.m)
100
10
25
20
15
10
0 20 40 60 80 100
Water content (%)
Figure 6.12: Chargeability, Resistivity and Ratio A2/A3 from the analysis of decay curves
against water content
78
6.5 Interpretation and discussion of results

The observations made during the saturation process of soil samples can be explained
in the following way. The properties of clay soils are effected by the total water con-
tent and by the energy by which that water is held. Due to changes in temperature,
extent of saturation and chemical composition or concentration of pores solutions, ef-
fects such as shrinkage on drying and swelling on water uptake are observed. How-
ever, the magnitudes of these changes depend upon interrelated variables of geolog-
ical history, fabric, extent of consolidation /cementation, mineralogical composition
and nature of exchange ions (Gillot, 1968). The high porosities observed with sam-
ples from Kenya are indication of the presence of high swelling clay minerals such as
Montmorillonite. This has an expanding lattice when hydrated. Secondly, porosity
also depends on other factors such as shape, arrangement and degree of sorting of the
constituent particles, cementation and compaction. As illustrated in Section 5.2.1,
the Kenya samples were not highly compacted, so this contributed partly to the high
porosities. The range of water content for the different soil samples is an indication
of the difference in soil types.
Figure 6.1 shows points with low resistivity and chargeability values as indicated
in section 6.3.1. This is due to the presence of ions that raise the conductivity of
pore filling water. The conductivity of the water distributed within the soil pores
is increased by the ionization of the clay and surface conductance as discussed in
Section 2.3.2. The ions that are adsorbed by the clay surface may be desorbed in a
percentage depending on a particular clay mineral, the concentration of clay in water
and the particular cation involved in the desorption (Keller and Frischknecht, 1966).
Therefore, the conductivity of water in the pore structure will always be increased by
those ions supplied. This makes those soil samples to have low resistivities. This ob-
servation agrees with the data obtained from the drill hole geophysical and lithology
well logs presented by Draskovitis and Smith (1990) . This is an attempt to inves-
tigate correlations between IP/Resistivity parameters and reservoir properties such
as Clay content. The difference in range of chargeability values given by Spain and
Kenya samples is explained by the type of clay and the manner in which it is dis-
tributed within the pore structure. The IP response given will depend on those two
conditions in that if the clay occurs in small clumps in the pores, it produces a smaller
IP effect than when the clay is lining the pore structure uniformly.
The relationship between chargeability and water content within the pore struc-
tures of soil is based on the availability of ions for ion exchange, which result into an
IP response. The behavior between the chargeability of samples in regard to water
content in this research, manifests itself in two ways: There are those that give a
bell shaped chargeability curve and others a linear relationship (Section 6.3.2). The
79
6.5. Interpretation and discussion of results
bell shape type of behavior is explained by presence of fixed sites on the clay surface
that are responsible for active IP. When these sites are all occupied by water, the IP
response obtained is small, however, as the water gets decreased, the sites become
available for ion exchange which gives a marked increase in the response up to the
maximum. When the water decreases further, there is a decrease of mobile ions due
to the fact that they are now closely attached to the clay surface and as result the IP
response decreases as well.
For the linear relationship, it is because the right side of the curve (bell-shape curve)
is not reached .i.e water contained in the soil samples does not occupy all the fixed
sites. The observation in Figure 6.2 is in agreement with what Parkhomenko (1971)
noted. The same behavior was also observed by Ogilvy and Kuzmina (1972) in shaly
sands.
Another explanation is that with the introduction of water and mixing with ma-
terials inside the soils, the mobility of the ions is changed by increasing the viscous
drag as it moves through water (Keller and Frischknecht, 1966). This is a result of the
fact that many layers of water molecules are normally attached loosely to the walls
of the pore. When an ion migrates through a fine pore, the forces holding these layers
of water to the soil increase the viscosity, which in turn decrease the mobility of ions
travelling through the pore. There will be a tendency for ionic charge to pile up on the
upstream side of such constrictions, forming an anion concentrations on one side and
cation concentrations on the other side, therefore, when an electric current is applied,
there will be movement of ions but on withdrawal of the electric current, equilibrium
is attained through diffusion processes, which gives an IP response which decreases
with a decrease in the amount of water.
The linear increase of the IP response with water (Figure 6.3) has also been ob-
served by Fraser and Ward (1965) . This is in response to increased production of
membrane barriers because the IP effect depends partly on the degree to which the
pore structure in the soil is filled with water.
All samples from Kenya show a linear relationship between chargeability and
water content (left part of the active IP curve) whereas Spain samples are showing a
bell shape type of relationship . This is attributed first to the formation conditions of
these soils. The Kenya soils were formed as a result of intense weathering of volcanic
rocks under high temperature and rainfall. These type of soils have high presence of
active aluminium and Iron oxides that cause isomorphic substitution in clay mineral
structures hence giving rise to high CEC due to high amounts of exchangeable Al
ions. This is further depicted in the highest values of chargeabilities that are given
by these soils. Soils formed by weathering are the residual clay containing materials
rich in high swelling clay minerals (Gillot, 1968). Soils from Spain are formed under
different climatic conditions from Kenya, Continental Mediterarranean type and lack
80
of enough rainfall leads to calcareous layers to be shallow which impedes the produc-
tion of bivalent cations like Ca that are used for ion exchange in the inter layer clay
mineral structures, which process gives rise to IP response. It should, however, be
noted that the parent material effects soil formation and has particularly important
influence on the origin of the clay minerals. In that regard, the porosity and water
holding capacity of the parent material effect the concentration of solutions and it is
further responsible for what ions are available to go into solution (Gillot, 1968).
The observed trend in chargeability and CEC values with soils in Figure 6.4 at
a fixed clay content classes agrees with observations made by Parkhomenko (1971) .
CEC refers to the quantity of negative charges on clay surface in soil, thus the capac-
ity to attract positively charged ions (cations). It is by this analogy that the more the
ions that are available for attraction, the higher the IP response . It was shown by Sill
(1964) that as CEC increases, membrane polarization increases too. With high CECs,
the free mobility of ions is restricted locally and as a result, membrane blockage is
caused, thus blocking the flow of negative charge carriers on the application of an
electric field. However, when the electric field is eliminated, all the charges return to
equilibrium positions by a diffusion process, which gives IP effect proportional to the
number of exchangeable ions and available ions in the electrolyte. Montmorillonite
is one of the clay minerals with a high CEC value (Section 2.3.1)
The observation made in regard to granulometrical composition of Spain samples
is in agreement to what most investigators notably Keller and Frishknecht, 1966;
Ogilvy and Kuzmina,1972 found when considering a much wider range of grain size
parameters, however, they both worked with artificial samples. Samples from Kenya
(Figure 6.7) do not show any relationship between the chargeability and the grain
sizes indicating that in some natural samples, there are other variables that con-
tribute to the chargeabilities, an observation also made by Illiceto et al. (1982) when
working with fine sediments.
The decay curve analysis showed that A1 , A2 , A3 depend on water content and
soil type. This is in agreement with what Illiceto et al. (1982) observed when work-
ing with fine sediments. The use of ratios of Aj is, however, to remove the influence
caused by the on time inducing of the voltage and investigate the response due to
membrane barriers in the soil. Therefore, these ratios attempt to normalize fun-
damental discrepancies and enhance the response due to the property of the soils.
The use of ratios in this research was able to differentiate soil samples according to
their origin and give imperfect clusters of soils in regard the same CEC class. How-
ever, the ratios were not able to differentiate samples perfectly according to their
lithotype as observed by Illiceto et al. (1982) . This is due to the fact that though
Illiceto et al. (1982) worked with natural soil samples, they carried out specific tests
where they separated samples into discrete classes according to granulometric con-
81
6.5. Interpretation and discussion of results
stituents, which was made possible by natural processes that made sediments more
or less sorted depending on the horizon and locations sampled. This enabled them
to get materials with dominant lithotype. Whereas for this research, the soil sam-
ples were a mixture of grain sizes and there was no control over their constituents
particles, possibly contributing to the failure of identifying lithotypes by use of ratios.
The results show that IP depends on water content, cation exchange capacity and
clay content. The dependency of polarizability of soils upon their lithological compo-
sition and hydrologic properties favours the applicability of IP method in studying
the composition, hydrologic and strength properties of soils/rocks, which is important
for engineering–geologic investigations. Therefore, the IP technique can be used as
a parameter for establishing correlative relations between geophysical and strength
parameters of soils. The polarization parameter makes it possible to differentiate soil
with respect to the lithology and characterize its particle size distribution which is
crucial for engineering purposes. Therefore, it is essential that polarizability data be
compared with mineralogical and chemical analysis as well as porosity and other hy-
drologic and granulometrical properties. Such comparisons are of interest in estab-
lishing correlations between IP parameters and parameters related to engineering
geologic investigations.
It should be noted that IP techniques can not be applied independently of other
electrical, mineralogical–chemical analysis, porosity determination, compressive
strength, particle size analysis, remote sensing techniques and seismic methods for
engineering geologic studies. The comprehensive solution to the problem of soil
swelling as confronted by foundation engineers can be achieved by comparing the
parameters obtained from the analysis of various physical fields.
82
Chapter 7
Models
7.1 Introduction
The use of IP in this research is an attempt to find an in–situ information provider
for the detection/identification of clay minerals as regards their swelling potential in
soils. Having carried out laboratory measurements on a suite of near surface soil
samples, the goal was to attempt to obtain an empirical relationship, which can give
a quantitative estimation of the CEC from chargeability values, water content, resis-
tivity and clay content. However, this chapter attempts to describe the concepts of
various models.
7.2 Theoretical aspects of model development

Soils are inhomogeneous systems that consist of macropores filled with one or more
pore fluids and macrograins of different sizes with or without clay coating. Move-
ments of free and bound ions in the vicinity of the interface between ionic solution
and ion exchange competent (clay) particles cause electrical (induced) polarization
and to a lesser extent also electrolyte/grain interface.
Experimental investigations relating rock properties and electrical parameters
in regard to electrolytic polarization have been restricted to sedimentary rocks e.g
Archie, (1942) ; Waxman and Smits, 1968 and Knight and Nur, 1987 although stud-
ies related to microcracks involved igneous and metamorphic rocks as well as for
electrode polarization. For the whole spectrum of rocks, resistivity has been used.
There have been attempts to use IP measurements to estimate hydraulic permeabil-
ity as first envisaged by Vacquier et al. (1957) . In this regard, Oluronfemi and
Griffiths, 1985 described an empirical relation between time domain IP parameters
and permeability of the Sherwood sandstone formation.
83
7.2. Theoretical aspects of model development
The Cole–Cole dispersion is the spectrum of a model first proposed by Cole and
Cole (1941) to describe the complex dielectric behavior and has been used by Pelton et
al. (1978) to describe complex resistivity behavior of rocks. The Cole–Cole expression
for the impedance of the equivalent circuit is then given by Pelton et al. (1978) as:
1

Z(ω) = Ro 1 − m 1 − (7.1)
1 + (jωτ )c
Where Ro is the resistance that simulates the continuous pore paths, m is the charge-
ability of the rock, τ is the relaxation time constant, c is the frequency dependency, ω
√
is the angular frequency and j= −1. This is one of the most known and used models,
however, the ion trap model by Keller and Frishcknecht, 1966 is important especially
in this type of study.
Keller and Frischknecht, 1966 described a relation between induced polarization
and clay content. He began by approximating that in a rock containing clay, the
number of polarization centers (fixed clay anions) will be proportional to the density
of fixed charges on the rock frame work, a0 . This density of fixed charges was consid-
ered to be the ion exchange capacity of the rock expressed in chemical equivalents per
unit pore volume. The constant of proportionality between number of clay particles
and total exchange capacity would then depend on the type of clay and the manner
in which it was distributed in the rock. This is supported by the fact that if clay
occurs in small clumps in the pores, it will form fewer potential barriers than if the
clay lines the pore structure uniformly. It was also assumed that the polarization
was approximately proportional to the number of anions in solution which are free to
accumulate at the potential barriers, (a − a0 )/a. Using the cation exchange capacities
for the various clay minerals, an estimate of the clay content that would cause the
maximum amount of induced polarization to take place in a rock could be obtained.
It follows that the amount of polarization should be proportional to the number of
potential barriers formed by the clay and to the fraction of the anions free to collect
at such barriers:
a0 (a − a0 )
P olarization ∝ (7.2)
a2
a0 (a−a0 )
where a2
is the anion trap number.
The expression 7.2 models a rock when it is saturated with water containing dis-
solved salts .i.e having a concentration of ions in ppm and in which one of the three
minerals may occur, illite, montmorillonite or kaolinite.
Keller and Frischknecht, 1966 calculated curves showing the relationship be-
tween the fractional volume of clay within a reservoir rock and a factor proportional
to induced polarization (anion trap number). Figure 7.1 shows three clay minerals
that were considered separately, illite, montmorillonite and kaolinite. The calcula-
tions were made for a 20% porosity and an electrolyte salinity of 100 ppm. Maximum
84
Chapter 7. Models
polarization was observed for small amounts of montmorillonite clay in a rock, from
0.1 to 0.4 %, while in the case of kaolinite, which has a much lower exchange ca-
pacity than montmorillonite, maximum polarization was observed for clay content
between 3 to 12%. However, these figures were considered to be only qualitative. In
case a higher water salinity could be assumed, then larger amounts of clay would be
required to cause a maximum polarization.
Figure 7.1: Anion trap number against clay concentration (Keller, 1966)
It is upon this that the highly idealized model of a clay bearing rock is based be-
cause the induced polarization observed in such rocks correlates with the ion–trap
ratio defined in Equation 7.2. This ratio can be computed for a given rock and a given
water salinity, a, if the cation exchange capacity, a0 is known. Though the model has
been applied to rocks, it allows for the derivation of most of the essential features of
the IP phenomenon observed in the field and in the laboratory with unconsolidated
samples like soils. From expression 7.2, the chargeability (polarization) depends on
the amount of clay present in isolated selective zones and the cation exchange capac-
ity. Then the co-efficient of proportionality between the number of clay particles and
the total exchange capacity depends on the type of clay and the manner in which it is
distributed within the soil.
At this stage, it would be logical to fit a model that can in turn be used to get an
estimator for CEC. The Cole–Cole model would not be adequate in describing these
soils compared to the model by Keller and Frischkneit (1966). This model (the ion
trap model) is useful because it explains from a physicochemical perspective the bell
shaped response of most of the factors influencing IP response: CEC, Clay concentra-
85
7.2. Theoretical aspects of model development
tion, electrolyte concentration and water saturation.

According to the observations made in this research, the model that could fit or
describe the results would be a modification of the Keller model but an attempt to do
so was not possible because part of the data set was noisy due to errors that have been
mentioned in Section 5.9. The remaining part of the data set did not yield a good fit
for the model. Therefore, the CEC estimator was not obtained when non linear least
square fitting algorithm (Levenberg–Marquardt) was applied on this data. However,
the idea was to obtain a CEC estimator from IP/Resistivity measurements so as to
be able to obtain information that can be applied by foundation engineers as regards
the subsurface information before any engineering project could take foot. It has been
clearly shown in Section 6.3.3 that IP depends on CEC and from earlier discussion,
CEC has been shown to represent clay mineralogy, so this estimator would have been
an added advantage for the potential evaluation of this approach (detection of clay
minerals by IP technique).
86
Chapter 8
Conclusion and
Recommendation
To detect and map the effect of the presence of clay minerals, several attempts have
to be made in order to find ways that can be used. However this has been the first
time, the IP technique has been tried at ITC for this purpose and therefore it has
been a learning process. Through this attempt, the IP technique has showed promise
of being an effective tool in attempting to detect clay mineralogy though it requires
some improvements.
8.1 Conclusions
• It was observed that the subject of IP in regard to identification of clay minerals
is not in common use. So this has made this study a try and error way of
learning. But since the presence of clay particles significantly intensifies the
on set of induced potentials, then the potential of using the IP technique for
detecting the presence of clay minerals is viable.
• Experimental investigation in the laboratory led to a better understanding of

the fundamental relationship between soil properties and the electrical param-
eters (IP and Resistivity), however, their usage and application of their infor-
mation content needs to be improved.
• The task of interrelating variables like chargeability, resistivity, water content,

cation exchange capacity, grain size and clay content using graphical display
has been shown in this research. This inter relationship provided a better un-
derstanding of the IP technique and ways of approaching the subject.
87
8.2. Recommendations
• The results have demonstrated that the IP response is possible when soils con-
tain some water as opposed to being completely dry. The IP response depended
highly on the amount of water. The Kenya samples showed higher chargeability
values than Holland and Spain samples.
• A model of 3 exponentials fitted to the decay curves led to the use of ratios
of the Amplitude (co-efficients), which provided a clear distinction of samples
according to their origin but not according to their lithotype.
• Through this experimental study, all sources of errors and their magnitude
have been identified. Personal experience in carrying out laboratory experi-
ments, proper reading and recording of results has been acquired.
• IP/Resistivity measurements on soil samples have shown that characterization

of the subsurface for engineering purposes is possible but more studies and
experiments in this regard are required.
• Results presented in this research are of an academic nature in the sense that
they evaluate potential approaches to the IP technique rather than provide
ready to use applications.
8.2 Recommendations
The results that have been obtained, so far, have created confidence in the possible
use of the technique in clay swelling potential evaluation, however, the following
considerations have to be noted:
• In an attempt to improve the IP technique for clay mineral identification, due

attention has to paid on the sampling procedure. Special sample tubes should
be used in the sampling procedure and subsequently used for the IP measure-
ments in order to reduce the disturbance within the soil.
• The soil samples should specifically be prepared for one set of measurements e.g
water content and IP/Resistivity measurements. This is meant to keep polar-
ization properties of materials intact, for example, when membrane polariza-
tion materials, such as clay, are heated or dried and then water is subsequently
reintroduced into the pores, they lose most of their polarization properties.
• For any further research in this field of of laboratory IP, it is recommended to

use more than one charging time (T) in order to fully examine the IP character-
istics due to a wide variety of charging times (e.g T=4s, 8s, 16s)
88
Chapter 8. Conclusion and Recommendation
• For any future experiment, care should be taken with the non polarizable po-
tential electrodes. They should be assembled properly without any possible
dripping. Preferably, they should be factory made or locally assembled from
good porous materials. This would prevent the flow of copper sulphate solu-
tion into the soil, thus reducing the contamination of the samples. Secondly,
this would reduce the amount of work involved in filling these electrodes with
copper sulphate solution on a daily basis.
• In order to improve the signal to noise level, higher input voltages should be
used, considering the maximum input voltage of the receiver in use, however,
this should satisfy the requirement of linearity, therefore, sufficiently low cur-
rent densities should be used.
• When using the CTU sample holder for any other experiment, the springs
should first be oiled in order to allow easy movement of the current electrode
extensions.
• In order to evaluate how well these laboratory measurements represent in situ

IP response of soils, it is strongly recommended that for any further research on
this topic, these small scale laboratory measurements should be accompanied
by field measurements that are carried out at the same sampling sites.
89
8.2. Recommendations
90
Bibliography
Angoran, Y. and Madden, T. R. (1977). Induced polarization: a preliminary study of

its chemical basis. Geophysics, 42:788–803.
Archie, G. E. (1942). The electric resistivity log as an aid in determining some reser-
voir characteristics. Am. Inst. Mining Matall. Engineers Trans., 146:54–62.
Bertin, J. and Loeb, J. (1976). Experimental and Theoretical Aspects of Induced Po-
larisation, volume 1 of 1. Gebruder Borntraeger, Berlin.
Bioag (2001). http://bioag.byu.edu/aghort/282pres/claychem/s1d019.htm. Chemical

and colloidal properties of soil.
Bleil, D. F. (1953). Induced polarization, a method of geophysical prospecting. Geo-

physics, 18:636–661.
Boadu, F. K. and Seabrook, B. (2000). Estimating hydraulic conductivity and porosity

of soils from spectral electrical response measurements. JEEG, 5(4):1–9.
Bolt, G. H. and Bruggenwert, M. G. (1978). Soil Chemistry, Part A, Basic elements.

Elsevier, Newyork.
Borner, F., Gruhne, M., and Schon, J. (1993). Contamination indications derived
from electrical properties in the low frequency range. Geophysical Prospecting,
41:83–98.
Borner, F. D. and Schon, J. H. (1995). Low frequency complex conductivity measure-

ments of microcrack properties. Surveys in Geophysics, 16:121–136.
Borner, F. D., Schopper, J. R., and Weller, A. (1996). Evaluation of transport and stor-
age properties in the soil and groundwater zone from the induced polarisation
measurements. Geophysical Prospecting, 44(4):583.
Cohen, M. H. (1981). Scale invariance of the low–frequency electrical properties of

inhomogeneous materials. Geophysics, 46:1057–1059.
91
BIBLIOGRAPHY
Cole, K. S. and Cole, R. H. (1941). Dispersion and adsorption in dielectrics. Alternat-

ing current field. J. Chem. Phys., 9:341.
Dakhnov, V. N. (1962). Geophysical well logging. Q. Colo. Sch. of mines.
Dankhazi, G. (1993). A new principled approach to induced polarization in porous

rocks. The Log Analyst.
De Bakker, H. (1978). Major soils and soil regions in the Netherlands. Dr. W. Junk
B. V. Publishers.
Dizon-Bacatio, C. R. (1992). A study on mapping unit composition and map evalu-

ation in the Antequera area, Malaga, Spain. Msc, ITC, Enschede, The Nether-
lands.
Draper, N. and Smith, H. (1981). Applied regression analysis. Wiley, New York.
Draskovits, P. and Smith, B. D. (1990). Induced polarisation surveys applied to eval-

uation of groundwater resources, Pannonian Basin, Hungary. In Ward, S. H.,
editor, INDUCED POLARISATION: Applications and Case histories, volume 4.
Society of Exploration Geophysicists.
Edge, A. B. and Laby, T. H. (1931). The principles and practice of geophysical

prospecting. Technical report, Imperial geophysical Experimental Survey, Aus-
tralia.
Fairburn, W. A. (1963). Geology of the North Machakos–Thika area. Technical report,

Geological Survey, Kenya.
Franklin, J. A. and Vogler, U. W. (1979). Suggested methods for determining water

content, porosity, density, absorption and related properties and swelling and
slake–durability index properties. Int. J. Rock Mech. Min. Sci. and Geomech,
16:141–156.
Fraser, D. C. and Ward, S. H. (1965). Electrical properties of clay contaminated sand-

stone. Geophysics, 30:1233–1234.
Gillot, J. E. (1968). Clay in Engineering Geology. Elsevier Publishing Company,

Amsterdam London New York.
Grahame, D. C. (1952). Mathematical theory of faradaic admittance. J. Electrochem.

Soc., 99:370.
Griffiths, D. H. and King, R. F. (1965). Applied Geophyiscs for Engineers and Geolo-
gists. Pergamon Press, London.
92
BIBLIOGRAPHY
Grim, R. E. (1968). Clay mineralogy. McGraw-Hill, New York.
Head, K. H. (1992). Manual of Soil Laboratory Testing, volume 1. Pentech Press,

London.
Henkel, J. H. and Collins, T. C. (1961). Induced polarization in electrolyte saturated

earth plugs. Geophysics, 26:205–210.
Illiceto, V., Santarato, G., and Veronese, S. (1982). An approach to the identification
of fine sediments by induced polarization laboratory measurements. Geophysical
prospecting, 30:331–347.
IRIS (1998). ELREC–6 IRIS Receiver. User Manual.
Keller, G. V. and Frischknecht, F. C. (1966). Electrical Methods in Geophysical

Prospecting, volume 10. Pergamon Press Inc.
Klein, J. D. and Sill, W. R. (1982). Electrical properties of artificial clay-bearing sand-

stone. Geophysics, 47(11):1593–1605.
Knight, R. and Nur, A. (1987). The dielectric constant of sandstones. Geophysics,

52:644–654.
Magellan (1992). http://www.unctad.org. 1992 Magellan Geographix SM Santa Bar-

bara.
Marquardt, D. W. (1963). An algorithm for least squares estimation of non–linear pa-

rameters. Journal of the Society of Industrial and Applied Mathematics, 2:431–
441.
Marshall, D. J. and Madden, T. R. (1959). Induced polarisation, a study of its causes.

Geophysics, 24:790–816.
Mazac, O., Kelly, W. E., and Landa, I. (1985). A hydrogeophysical model for relations
between electrical and hydraulic properties of aquifers. Journal of Hydrology,
79:1–9.
Meijer, H. (1994). Compact Geography of the Netherlands. The Information and

documentation center for the Geography of the Netherlands, Utrecht.
Mitchel, J. K. (1993). Fundamentals of soil behaviour. John Wiley and Sons, New
York, second edition.
Mohamed, S. S. (1970). Induced polarisation, a method to study water collecting

properties of rocks. Geophysical Prospecting, 18:654–665.
93
BIBLIOGRAPHY
Nelson, P. H., Hansen, W. H., and Sweeney, M. J. (1982). Induced–polarization re-

sponse of zeolitic conglomerate and carbanaceous siltstone. Geophysics, 47:71–
88.
Ogilvy, A. A. and Kuzmina, E. (1972). Hydrogeologic and engineering-geologic possi-

bilities for employing the method of induced potentials. Geophysics, 37(5):839–
861.
Olhoeft, G. R. (1985). Low frequency electrical properties. Geophysics, 50:2492–2450.
Olorunfemi, M. O. and Griffiths, D. H. (1985). A laboratory investigation of the in-

duced polarization of the Triassic Sherwood sandstone of Lancashire and its hy-
drogeological applications. Geophysical Prospecting, 33:110–127.
Parasnis, D. S. (1966). Mining Geophysics. Elsevier Publishing Company, Amster-

dam.
Parasnis, D. S. (1973). Mining Geophysics. Elsevier Publishing Company, Amsterdam

London New York.
Parkhomenko, E. I. (1971). Electrification phenomena in rocks. Plenum Press, New

York.
Pelton, W. H., Ward, S. H., Hallof, P. G., and Nelson, P. H. (1978). Mineral discrim-
ination and removal of inductive coupling with multifrequency IP. Geophysics,
43:588–609.
Reeuwijk, L. P. V. (1995). Procedures for soil analysis. Technical report, International

Soil Reference and Information Center.
Reynolds, J. M. (1997). An introduction to Applied and Environmental Geophysics.

John Wiley and Sons.
Roussel, J. (1962). Etude sur modeles reduit des phenomenes de polarisation provo-
quee. Annales de Geophysique, 18:360–371.
Roy, K. K. and Elliot, H. (1980). Model studies on some aspects of resistivity and
membrane polarisation behaviour over a layer Earth. Geophysical prospecting,
28:759–775.
Schlumberger, C. (1920). Etude sur la prospection electrique du sous–sol. Gauthier–

Villars, Paris.
Schon, J. H. (1986). Physical Properties of Rocks; Fundamentals and Principles of

Petrophysics. Elsevier Science Limited, Amsterdam.
94
BIBLIOGRAPHY
Scintrex (1981). CTU–2: Physical Property Testing System. User Manual.
Sherman, D. M. (1997). http://mineral.gly.bris.ac.uk/mineralogy/clayminerals.html.

Phyllosilicates I–The clay minerals.
Siegel, H. O. (1959). Mathematical formulation of type curves for induced polariza-

tion. Geophysics, 24(3):546–565.
Sill, W. R. (1964). Induced polarization in clay bearing sandstones and effects of oil
saturation. Inst. of Geoph. and Plan. Physics, Univ. of California, San Diego.
SPSS. (1998). SPSS Base 8.0: Application Guide. Marketing Department, SPSS Inc.,
Chicago.
Sternberg, B. K. and Oehler, D. Z. (1990). Induced polarisation in hydrocarbon sur-

veys: Arkoma basin case histories. In Ward, S. H., editor, INDUCED POLARI-
SATION: Applications and Case histories, volume 4. Society of Exploration Geo-
physicists.
Sumner, J. S. (1976). Principles of Induced Polarisation for Geophysical Exploration.

Elsevier Scientific Publishing Company, Amsterdam.
Sumner, J. S. and Zonge, K. L. (1980). Induced polarisation for exploration

geophysics–short course. South African Geophysical Association, The Univer-
sity Of Witwatersrand, Johannesburg, South Africa.
Telford, W. M., Geldart, L. P., Sheriff, R. E., and Keys, D. S. (1976). Applied Geo-
physics. Cambridge University Press.
Towel, J. N., Anderson, R. G., Pelton, W. H., Olhoeft, G. R., and LaBrecque, D.
(1985). Direct detection of hydrocarbon contaminants using induced polarisa-
tion method. SEG meeting, pages 145–147.
Vacquier, V., Holmes, C. R., Kintzinger, P. R., and Lavergne, M. (1957). Prospecting
for groundwater by induced electrical polarisation. Geophysics, 23:660–687.
Van der Meer, F. (1999). Can we map swelling clays with remote sensing? and dis-
cussion by Dr. Jean Roy (p233–234). International journal of Applied Earth Ob-
servation and Geoinformation, 1:27–35.
Vanhala, H. (1997). Mapping oil contaminated sand and till with the spectral induced
polarization (SIP) method. Geophysical prospecting, 45:303–326.
Vanhala, H. and Peltoniemi, M. (1992). Spectral ip studies of Finnish ore prospects.

Geophysics, 57:1545–1555.
95
BIBLIOGRAPHY
Vanhala, H. and Soininen, H. (1995). Laboratory technique for measurement of spec-

tral induced polarization response of soil samples. Geophysical prospecting, 43.
Vinegar, H. J. (1984). Induced polarization of shaly sands. Geophysics, 49(8):1267–

1287.
Wait, J. R. (1958). Discussions on a theoretical study of induced electrical polariza-

tion. Geophysics, 23:144–154.
Wait, J. R. (1959a). Overvoltage research and geophysical application. In Interna-

tional series of monographs on earth sciences, volume 4, pages 22–83. Pergomon
press, Oxford, England.
Wait, J. R. (1959b). A phenomenological theory of overvoltage for metallic particles.

In Overvoltage research and geophysical application, volume 4, pages 22–28. Per-
gomon press, Oxford, England.
Wait, J. R. (1959c). The variable–frequency method. In Overvoltage research and geo-

physical application, volume 4, pages 29–49. Pergomon press, Oxford, England.
Waxman, M. and Smits, L. J. M. (1968). Electrical conductivities in oil–bearing shaly

sands. Soc. Pet. Eng. J, 243:107–122.
Waxman, M. H. and Thomas, E. C. (1974). Electrical conductivities in shaly sands. J.

Petr. Tech., 257:213–225.
Weller, A. and Borner, F. D. (1996). Measurements of spectral induced polarization

for environmental purposes. Environmental Geology, 27:329–334.
Worrall, W. E. (1968). Clays, their nature, origin and general properties. Maclaren
and Sons, London, first edition.
Worthington, P. A. and Collar, F. A. (1984). Relevance of induced polarisation to

quantitative formation evaluation. Marine and Petroleum Geology, 1.
Zonge, K. L., Sauck, W. A., and Sumner, J. S. (1972). Comparison of time frequency
and phase measurements in induced polarisation. Geophysical Prospecting,
20:626–648.
96
Appendix A
3.5
3
Chargeability(mV/V)
2.5
1.5
4 8 12 16 20 24
Water content (%)
Figure 1: Chargeability against water content for sample 1
97
3.2
2.8
2.4
1.6
1.2
10 15 20 25 30 35 40
Water content (%)
3
Chargeability(mV/V)
12 16 20 24 28
Water content (%)
2
Chargeability(mV/V)
8 12 16 20 24 28
Water content (%)
98
Appendix . Appendix A
1.6
1.2
0.8
0.4
14 16 18 20 22 24
Water content (%)
1.2
0.8
0.4
4 8 12 16 20 24
Water content (%)
2.5
2
1.5
0.5
8 12 16 20 24 28
Water content (%)
99
1.2
0.8
0.6
0.4
0.2
35.5 36 36.5 37 37.5 38

Water content (%)
6
8 12 16 20 24 28
Water content (%)
4
Chargeability(mV/V)
29 30 31 32 33
Water content (%)
100
2.5
1.5
0.5
8 12 16 20 24
Water content (%)
3
8 12 16 20 24
Water content (%)
10
8
20 24 28 32 36 40
Water content (%)
101
8
4
20 25 30 35 40 45
Water content (%)
6
20 24 28 32 36
water content (%)
3
Chargeability(mV/V)
8 12 16 20 24 28
Water content (%)
102
16
12
Chargeability(mV/V)
16 20 24 28 32
Water content (%)
3
Chargeability(mV/V)
12 16 20 24 28
Water Content (%)
103
3
2.5
2
1.5
0.5
30 32 34 36 38 40
Water content (%)
14
12
10
30 40 50 60 70
Water content (%)
1.6
1.2
0.8
0.4
16 20 24 28 32 36
Water content (%)
104
20
16
Chargeability(mV/V)
12
30 35 40 45 50 55 60
Water content (%)
6
Chargeability(mV/V)
25 30 35 40 45 50
Water content (%)
20
16
12
30 40 50 60 70
Water content (%)
105
16
12
8
40 50 60 70 80
Water content (%)
20
16
V/V)
12
Chargeability(m
40 45 50 55 60 65 70
Water content (%)
25
20
Chargeability(mV/V)
15
10
50 60 70 80 90
Water content (%)
106
16
12
Chargeability (mV/V) 8
52 56 60 64 68 72 76
Water content (%)
10
8
20 25 30 35 40 45
Water content (%)
16
12
20 30 40 50 60
Water content (%)
107
20
16
Chargeability(mV/V)
12
30 40 50 60 70 80
Water content (%)
CEC classes
2-10 me q/100g (Sp ain )
2-10 me q/100g (Ken ya )
11-15 me q/100g (Sp ain )
1 1-1 5 meq /100 (Kenya)
3 1 6-2 2 meq /100g (Spa in)
1 6-2 2 meq /100g (Ken ya )
2 3-26 meq/ 100g (Spain)
2 3-2 6 meq /100g (Ken ya )
2 7-30 meq/ 100g (Spain)
2 7-30 meq/ 100g (Kenya)

2
31 -45 meq/1 00g (Spain)
3 1-45 meq/ 100g (Ken ya)

A1
0 20 40 60 80 100
Water content (%)
1000
Figure 32: A1 against water content

100
Re sistivity (Oh m.m)
10
CEC classes
25

1
3
20

Cha rge ab ility (mV/V)
15
2 16-22 meq/100g (Spain)

10
5
A2
1
0
0 31-45 meq/100g (Kenya)
-1
0 20 40 60 80 100
Water content (%)
1000

100
Resistivity (Ohm.m)
10
108 25
20
15
10
CEC classes
2-10 meq/ 100g (Kenya)
2.5 11-15 meq/100g (Kenya)

2

1.5

A3
1 31-45 meq/100g (Spain)
0.5
-0.5
0 20 40 60 80 100
Water content (%)
1000

100
Resistivity (Ohm.m)
10
25
20
15
10 CEC classes
5 2-10 meq/100g (Spain)
5 2-10 meq/100g (Kenya)

40

3
A1/A2

2
0 20 40 60 80 100
Water content (%)
1000
Figure 35: Ratio A1/A2 against water content

100
Resistivity (Ohm.m)
10
25
1
109
20
15
10
CEC classes
6 2-10 meq/100g (Spain)
4 23-26 meq/100g (Spain)
A1/A3 27-30 meq/100g (Kenya)
0 20 40 60 80 100
Water content (%)
1000

100
Resistivity (Ohm.m)
CEC classes

5
10
4 16-22 meq/100g (Spain)
25 16-22 meq/100g (Kenya)

1
20 27-30 meq/100g (Spain)

3
A2/A1

C hargeability (mV/V)
15 31-45 meq/100g (Kenya)
10
51
0 20 40 60 80 100
Water content (%)
1000

100
Resistivity (Ohm.m)
CEC classes

2
10 16-22 meq/100g (Spain)

1.6

1.2
25
A3/A1

1
0.8 31-45 meq/100g (Kenya)
20
0.4
15
0 20 40 60 80 100
Water content (%)
10
1000
5
100
Resistivi ty (Oh m.m)
110 10
25
20
i ty (mV/V)
15
CEC classes

3 2-10 meq/100g (Kenya)

2

A3/A2

1
-1
0 20 40 60 80 100
Water content (%)
1000

100
Resistivity (Ohm.m)
10
25
20
15
10
111
112
Appendix B
113
Table 1: Soil sample description 1
Soil Easting Northing Altitude Depth Sand Silt Clay CEC Porosity
sample (m) (cm) (>0.05mm) (%) (0.05-0.002mm) (%) (<0.002m) (%) (meq/100g) (%)
1 344031 4106293 429 0-15 23.4 18.8 57.7 27.0 34
2 348375 4103048 397 0-10 16.8 21.8 61.4 27.6 54
3 340931 4112753 472 0-30 20.2 33.9 46.0 19.0 45
4 336625 4113242 431 0-15 6.9 27.7 65.4 24.0 46
5 346680 4098593 377 0-27 22.1 35.6 42.3 14.0 41
6 347177 4098651 378 0-25 34.3 21.2 44.5 13.0 39
7 345970 4096777 389 0-10 50.0 13.0 37.0 14.4 38
8 342670 4105921 411 0-18 41.8 13.6 44.6 18.9 42
9 340750 4106908 423 0-10 3.7 28.4 67.9 28.1 57
10 339034 4112378 480 0-10 3.7 24.7 71.6 45.0 51
11 345749 4097693 389 0-35 12.6 23.6 27.7 8.0 39
12 343763 4103538 411 0-10 39.0 19.1 41.8 5.0 51
13 343700 4103799 424 0-9 29.9 18.2 51.8 23.0 56
14 342784 4102307 430 0-23 13.7 11.7 38.3 15.0 42
15 345898 4100861 396 0-10 53.8 13.0 33.2 13.1 38
16 343137 4103236 412 0-10 24.2 27.0 48.8 23.4 52
17 347665 4098906 393 0-10 32.2 19.0 48.8 16.8 43
18 340595 4107443 429 0-10 12.9 29.9 57.2 25.0 55
19 344998 4099693 397 0-10 25.9 10.3 63.7 24.4 45
20 349800 4101000 397 0-15 16.7 32.8 50.4 21.0 53
21 344081 4103485 434 0-18 39.8 11.3 27.4 12.0 42
22 345223 4097321 383 0-25 3.9 25.1 57.3 29.0 43
23 344711 4102189 432 0-18 50.2 8.9 41.0 20.0 42
24 348664 4106126 456 0-10 33.9 20.0 46.1 25.0 54
25 275749 9868317 1519 18.3 40.6 41.1 21.1 63
26 298966 9864174 1508 35.2 44.2 20.6 11.5 64
27 243056 477543 38 67
28 234166 472713 58
29 227323 467414 72
30 243830 461320 28 54
31 267127 9866234 1610 13.0 15.5 71.5 12.2 59
32 316690 9860366 1590 41.3 40.1 18.5 4.8 56
33 305381 9878366 1464 24.4 4.9 70.7 19.8 63
34 298956 9862549 1525 24.5 48.1 27.4 17.6 65
35 267342 9851174 1661 14.1 28.4 57.4 36.8 63
36 309676 9845664 1563 43.5 4.3 52.3 4.3 60
37 284203 9856308 1540 21.0 23.0 56.0 25.6 73
38 243495 9855571 1969 10.8 3.6 85.6 8.1 62
39 271639 9843357 1583 32.6 25.5 41.9 9.7 56
40 242698 9843620 1882 10.0 19.0 71.0 33.5 65
41 319892 9835487 1322 54.0 5.2 40.8 2.1 54
42 301416 9857980 1519 46.6 17.9 35.5 23.0 57
43 282637 9881266 1557 17.3 29.0 53.7 28.8 66
114
Appendix . Appendix B
Table 2: Soil sample description 2
Soil Parent Landscape Soil Parent Landscape

sample material sample material
1 Lacustrine sediments lacustrine plain 20 Alluvium alluvial plain
2 Lacustrine sediments lacustrine plain 21 Marls & Gypsum hilland
3 Colluvium piedmont 22 Colluvium hilland
4 Lacustrine marls lacustrine plain 23 Colluvium hilland
5 Alluvium alluvial plain 24 Colluvium piedmont
6 Colluvium glacis 25 Residual soil lowland
7 Colluvium footslope 26 Residual soil undulating
8 Lacustrine deposits piedmont 31 Residual soil undulating lowland
9 Lacustrine sediments lacustrine plain 32 Residual soil foothill
10 Colluvium piedmont 33 Residual/Transported foothill
11 Bios sands hilland 34 Residual soil undulating
12 Colluvum piedmont 35 Residual soil lowland
13 Colluvium piedmont 36 Residual soil hillslope
14 Colluvium piedmont 37 Residual soil lowland
15 Bios sands piedmont 38 Residual/Colluvial hillslope
16 Bios sands piedmont 39 Residual soil undulating lowland
17 Colluvium hilland 40 Residual/Transported foothill
18 Lacustrine sediments lacustrine terrace 41 Residual soil flat hillslope
19 Bios sands piedmont 42 Residual soil lowland
43 Residual soil tophill
115
Table 3: Geometric constant (K) values
Soil Spacing between K Soil Spacing between K

sample potential elctrodes value sample potential elctrodes value
(cm) (cm)
1 1.5 0.3 31 1.5 0.3
2 1.7 0.2 32 1.5 0.3
3 1.6 0.3 33 1.5 0.3
4 1.6 0.3 34 1.8 0.2
5 1.6 0.3 35 1.8 0.2
6 1.7 0.3 36 1.7 0.2
7 1.5 0.3 37 1.8 0.2
8 1.8 0.2 38 1.6 0.3
9 1.5 0.3 39 1.5 0.3
10 1.6 0.3 40 1.8 0.2
11 1.5 0.3 41 1.7 0.2
12 1.5 0.3 42 1.7 0.2
13 1.6 0.3 43 1.7 0.2
14 1.5 0.3
15 1.7 0.2
16 1.6 0.3
17 1.6 0.3
18 1.6 0.3
19 1.7 0.2
20 1.7 0.2
21 1.7 0.2
22 1.7 0.2
23 1.7 0.2
24 1.7 0.2
25 1.8 0.2
26 1.8 0.2
27 1.7 0.2
28 1.7 0.2
29 1.7 0.2
30 1.7 0.2
116
Table 4: IP/Resistivity measurements reliability
Soil Comment
sample
1 Okay but shrunk on drying (Aluminium foil used)
2 Sample tube diagonally cut, no proper electricl contact, shrunk (Aluminium foil used)
3 No proper contact on drying (Aluminium foil used)
5 Okay
8 Overgrew and shrunk, no proper electrical contact on drying (aluminium foil used)
9 Overgrew and shrunk, no proper electrical contact on drying, irregular surface
10 Sample tube diagonally cut (Aluminium foil and old 2.50 guilder coins)
11 Sample tube diagonally cut (Aluminium foil used)
12 Okay
17 Sample tube diagonally cut (Aluminium foil used and old guilder coins)
20 No proper contact on drying (Aluminium foil used)
21 Okay
25 Overgrew and shrunk, no proper electrical contact on drying (Aluminium foil used)
27 Sandy, lost water fast and was spilling out of the tube
28 Sandy, lost water fast and was spilling out of the tube
29 No proper electrical contact on drying
30 sandy, lost water fast and was spilling out of the tube
31 Sample tube diagonally cut, overgrew and shrunk (Aluminium foil used)
32 Okay
33 Overgrew and shrunk, no proper electrical contact on drying
36 Okay
117
Table 5: Causes for the unreliable determination of Aj and τj constants
Soil Reason
sample
1 Aluminium (Al) foil used, current density inhomogeneity
2 Al foil used, current density inhomogeneity, sample tube diagonally cut
3 All were determined
4 Al foil used, sample cohesion, current density inhomogeinity
6 Al foil used, current density inhomogeneity
7 Al foil used, current density inhomogeinity, sample cohesion
8 Irregular surface, sample cohesion
10 Al foil and 2 guilder coins used, current inhomogeinity
11 Current inhomogeinity
12 Current density homogeinity
13 Not included because of non linear current regulation
14 Sample tube diagonally cut, no poper contact
15 Al foil used, current density inhomogeinity
18 Sample tube diagonally cut, current density inhomogeinity
19 Al foil used, current density inhomogenity
21 Unregular distribution of current inside the sample
23 Al foil used, current inhomogeinity
24 Al foil used, current density homogeinity
25 Sample cohesion, Al foil used, current density inhomogeinity
26 Localized current source
27 Irregular distribution of current in the sample.
31 Al foil used, sample tube diagonally cut, unregular distribution of current in the sample
32 Al foil used, sample tube diagonally cut, unregular distribution of current in the sample
36 Unregular distribution of current inside the sample
118
Table 6: Calibration of the transmitter
119

Kiberu PDF

Uploaded by

Copyright:

Available Formats

Kiberu PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kiberu PDF

Uploaded by

Copyright:

Available Formats

Induced polarization and

Thesis submitted to the International Institute for Geo-information Science

Degree Assessment Board

INTERNATIONAL INSTITUTE FOR GEO - INFORMATION SCIENCE AND EARTH OBSERVATION

ENSCHEDE , THE NETHERLANDS

Time–domain induced polarization (IP) and resistivity laboratory

2 Background to the research 9

3 Physicochemical basis of the Induced Polarization method 19

3.3 Electrode impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4 Study areas and soil samples 33

5 Data acquisition methodology 41

5.8 Test measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

6 Experimental data and processing 61

8 Conclusion and Recommendation 87

6.1 Water content (%) values obtained with sample 1 (Spain) . . . . . . . . 61

1 Soil sample description 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

1.1 Research methodology flow chart . . . . . . . . . . . . . . . . . . . . . . 8

2.1 Position of the IP amongst other geophysical techniques . . . . . . . . 13

4.1 Location of study area (Spain) (Dizon-Bacatio, 1992) . . . . . . . . . . . 34

5.10 IRIS Elrec–6 receiver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

6.1 A plot of chargeability against resistivity . . . . . . . . . . . . . . . . . 63

7.1 Anion trap number against clay concentration (Keller, 1966) . . . . . . 85

1 Chargeability against water content for sample 1 . . . . . . . . . . . . 97

15 Chargeability against water content for sample 20 . . . . . . . . . . . . 102

lithological composition and hydrogeological properties favors the application of the

This study is based on laboratory induced polarization and resistivity measure-

1.2 Problem definition

1.3 Objectives of the research

• Observe the inter-relationship between variables such as IP response, resistiv-

• Attempt to identify all sources of error encountered in the experiment in an

1.4 Research Questions

4. Is there any relationship between the granulometrical composition of soils and

1.6 Sources of Data and Information

• Data documentation and presentation

– MSword, Grapher, MsPowerpoint,and LaTex

1.7 Research methodology

From the Clay

Particle Size INDUCED POLARIZATION WATER CONTENT

ANALYSIS OF THESE DATA SETS

Figure 1.1: Research methodology flow chart

Background to the research

2.2 The use of geophysical techniques: electric

V (t) = A1e−t/T 1 + A2e−t/T 2 + A3e−t/T 3 (2.1)

range (1mHZ to 10kHz) (Olhoeft, 1985). Using Induced polarization measurements,

2.2.1 Position of the IP technique among other

POTENTIAL FIELD SEISMIC TECHNIQUES ELECTRICAL TECHINIQUES NUCLEAR TECHNIQUES OPTICAL

INDUCTIVE GALVANIC Gamma ray Neutron

TIME DOMAIN EM FREQUENCY DOMAIN EM

TELLURIC RESISTIVITY MISE-A-LA- IP MIP SP COMPLEX RESISTIVITY

Figure 2.1: Position of the IP amongst other geophysical techniques

2.3 General characteristics of clays

Figure 2.2: The silicon tetrahedron (after Bioag, 2001)

Figure 2.3: The Aluminium octahedron (after Bioag, 2001)

ditions. If an Al tetrahedra substitutes for a Si tetrahedra in the tetrahedral layer,

2.3.1 Formation and types of clay minerals

3. Weathering of clays to form other types of clay

Montmorillonite → Kaolinite + SiOx

Figure 2.4: 1:1 type clay-kaolinite (Sherman, 1997)

Figure 2.5: 2:1 type clay-montmorillonite (Sherman, 1997)