Migration of Fines in Porous Media

MIGRATION OF FINES IN POROUS MEDIA
Theory and Applications of Transport in Porous Media
Series Editor:
Jacob Bear, Technion - Israel Institute of Technology, Haifa, Israel
Volume 12
The titles published in this series are listed at the end of this volume.
Migrations of Fines
in Porous Media
by
Kartic C. Khilar
Department of Chemical Engineering,
Indian Institute of Technology,
Bombay, India
and
H. Scott Fogler
University of Michigan,
Ann Arbor, Michigan, U.S.A.
Springer-Science+Business Media, B.Y.

Library of Congress Cataloging-in-Publication Data
ISBN 978-90-481-5115-8 ISBN 978-94-015-9074-7 (eBook)

DOI 10.1007/978-94-015-9074-7
Printed on acid-free paper
All Rights Reserved

© 1998 Springer Science+Business Media Dordrecht and copyright holders as
specified on appropriate pages within.
Originally published by Kluwer Academic Publishers in 1998.
Softcover reprint of the hardcover 1st edition 1998
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
to:
Students, Staff and Faculty of the
lIT, Bombay, Mumbai
TABLE OF CONTENTS
PREFACE xi
1 PRACTICAL CONSEQUENCES OF RELEASE AND

MIGRATION OF FINES IN POROUS MEDIA 1
1.1 Migration of Fine Particles in Porous Media: An Introduction. 1
1.2 Practical Consequences in Petroleum Engineering . 2
1.3 Importance in Geotechnical Engineering. 3
1.4 Effects in Environmental Engineering 5
1.5 Importance in Chemical Engineering .. 7
2 CHARACTERIZATION OF PORE SPACE AND FINES 9

2.1 Pore Structure . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Pore Constriction and Chamber Size Distributions. 11
2.1.2 Pore Connectivity . . . . 13
2.1.3 Porosity and Permeability . . . . . . . . 13
2.2 Pore Surface . . . . . . . . . . . . . . . . . . . 14
2.2.1 Surface Characterization of Pore Surface 15
2.2.2 Colloidal Characterization of Pore Surface 15
2.2.3 Chemical Characterization of Pore Surface 17
2.3 F i n e s . . . . . . . . . . . . . . . . . . . . 19
2.3.1 Geometrical Characterization Fines 19
2.3.2 Colloidal Characterization of Fines 23
2.3.3 Chemical Characterization of Fines 26
3 COLLOIDALLY INDUCED RELEASE OF FINES IN

POROUS MEDIA 29
3.1 The Statics of the Release Process . . . . 29
3.1.1 Electric double layer repulsion. . 31
3.1.2 London-van der Waals attraction 33
3.1.3 Born repulsion . . . . . . . . . . 36
viii TABLE OF CONTENTS
3.1.4 AB (Acid-Base) Interaction 36

3.1.5 Hydrodynamic Forces . . . 37
3.1.6 Total energy of interactions 38
3.1.7 Condition for the release of fines: The concept of
critical salt concentration, CSC. . . . . . . . . . . " 43
3.1.8 Critical total ionic strength (CTIS) for mixed salt system 50
3.1.9 Some important effects on the energy of interaction . 51
3.2 The Dynamics (Rate) of the Release Process . . . . . . . . " 54
3.2.1 Theoretical models for the release of Brownian fines . 55
3.2.2 Some considerations on the release of non-Brownian
fines . . . . . . . . . . . . . . . . . . . . . . . .. 58
3.2.3 An empirical equation for the rate of release of fines . 59
4 HYDRODYNAMICALLY INDUCED RELEASE OF FINES IN

POROUS MEDIA 63
4.1 The Statics of the Release Process. . . . . . . . . . . . . .. 63
4.1.1 Mechanism of hydrodynamic detachment . . . . . " 64
4.1.2 Critical hydrodynamic stress and critical flow velocity 67
4.2 The Rate (Dynamics) of the Release Process. . . . . . . . . . 69
5 ENTRAPMENT OR PIPING OF FINES DURING MIGRATION 73

5.1 Analysis of Factors Affecting Entrapment or Piping of Fines 73
5.1.1 Pore structure . . . . 73
5.1.2 Size offines . . . . . . . 74
5.1.3 Concentration of fines . . 75
5.1.4 Hydrodynamic conditions 78
5.1.5 Colloidal conditions . . . 78
5.2 Mathematical Models for Entrapment or Piping of Fines 79
5.3 Application to the Phenomenon of Soil Erosion . . . 84
5.4 Application to Sand Filtration . . . . . . . . . . . . 87
5.5 Application to Water sensitivity of Berea Sandstone. 88
6 MATHEMATICAL MODELS FOR PERMEABILITY

REDUCTIONS DUE TO MIGRATION OF FINES 91
6.1 The Release and Capture Mechanism . . . . . . . . 91
6.2 Rate of Equations for the Release of Fine Particles . 94
6.3 Rate Equations for the Entrapment/Capture of Fine Particles 95
6.4 Mass and Population Balance Equations for Fine Particles at
Different Sites . . . . . . . . . . . . . . . . . . . . . . . .. 99
TABLE OF CONTENTS ix
6.5 Correlation between Entrapment and Permeability Reduction . 101

6.6 Solution Procedures, Results and Comparisons with
Experimental Measurements . . . . . . . . . . . . . . . 102
6.6.1 The model of Gruesbeck and Collins for hydro-
dynamically induced migration offines . . . . . . . . 103
6.6.2 The model ofKhilar and Fogler for colloidally induced
migration of fines . . . . . . . . . . . . . . .. . 104
6.6.3 The model of Sharma and Yortsos for colloidally
induced migration of fines . . . . . . . . . . .. . 109
7 USE OF NETWORK MODELS FOR PREDICTION OF PERME-

ABILITY REDUCTION DUE TO FINES ENTRAPMENT 113
7.1 Need for network models . . . . . . . . . . . . . . . . · 113
7.2 Network Models . . . . . . . . . . . . . . . . . . . . . · 114
7.2.1 Network Construction and Lattice Arrangements · 114
7.2.2 Fluid transport through the network . . . . . . . · 117
7.3 Use of network models for studying particle capture . . . · 117
7.4 Application of network models for prediction of permeability
reduction due to straining dominated fines entrapment . . 119
7.4.1 Network construction . . . . . . . . . . . . . . 119
7.4.2 Calculation of flow distribution in the network . 120
7.4.3 Particle movement within a network. . . . . . . 122
7.4.4 Effect of network size . . . . . . . . . . . . . . 124
7.4.5 Ability of network models to account for polydisperse
particles and pores . . . . . . . . . . . . . . . .. . 125
7.4.6 Comparison of model predictions with experiments . 125
7.5 Accounting for particle deposition in network models . 128
7.6 Improved network models for prediction of
permeability reduction . . . . . . . . . . . . . . . . · 130
8 METHODS TO PREVENT THE RELEASE OF FINES 133

8.1 Enhancement of Force of Adherence/Attachment . . . . 133
8.2 Reduction of Force of Detachment . . . . . . . . . . . . 136
8.3 Minimization of Fines Release: Enhancing the Attachment
Forces and Reducing the Detachment Forces . . . . . 138
9 SOIL POLLUTION DUE TO MIGRATION OF FINES 141

9.1 Facilitated Contaminants Transport due to Migration of Fines . 141
x TABLE OF CONTENTS
9.1.1 Sorption of Hydrophobic Contaminants on Fines in

Soil Masses . . . . . . . . . . . . . . . . . . . . . . 142
9.1.2 Modeling of Transport of Contaminants Facilitated by
Migration of Fine Particles. . . . . . . . . 144
9.2 Migration of Biocolloida1 Contaminants. . . . . . . 148
9.2.1 Characteristics of Biocollidal Contaminants . 148
9.2.2 Modeling the Migration of Biocolloidal Fines 149
REFERENCES 155
INDEX 169
PREFACE
This is the first book entirely on the topic of Migration of Fine Particles in
Porous Media. There are two purposes for the use of this book. First, the book
is intended to serve as a comprehensive monograph for scientists and engineers
concerned with problems of erosion, pollution and plugging due to migration
of fines in porous media. Second, the book is recommended to be used as a
reference book for courses offered at senior or graduate level on the topics of
flow through porous media, soil erosion and pollution, or formation damage.
The migration of fine particles in porous media is an engineering concern
in oil production, soil erosion, ground water pollution and in the operation of
filter beds. As a result, the topic has been studied by researchers working in
a number of disciplines. These studies in different disciplines are conducted,
by and large, independently and hence there is some repetition and perhaps
more importantly there is a lack of uniformity and coherence. These studies,
nevertheless, complement each other. To illustrate the point, consider for
example the migration of fine particles induced by hydrodynamic forces.
Studies are reported in geotechnical literature related to soil erosion as well as
studies reported in petroleum engineering literature related to the decline in oil
production. These studies even though reported with different terminologies
and with different perspectives do complement each other in terms of the fact
that once the particles are released, they are either entrapped to cause decline
in permeability or transported with the flow causing phenomenon such as soil
erosion. With these ideas in mind our primary objective is to bring together
these studies and develop an unified theory and understanding on this topic.
The book is organized in a systematic manner. Chapter 1 presents the
practical consequences of migration offine particles in porous media related to
different engineering disciplines. Chapter 2 begins our unified treatment of the
phenomenon by characterizing both the porous media and the fine particles.
Although, this chapter focuses on natural porous media such as rocks and soil
and the indigenous fine particles present in these media, the techniques and the
approaches discussed are general in nature.
xi
xii PREFACE
The mechanism of release is probably the most important basic process in

the phenomenon and therefore two chapters, Chapter 3 and 4 are devoted to
it. Chapter 3 presents both the statics and dynamics aspects of the release of
fine particles due to colloidal forces (colloidally induced release). Chapter 4
discusses the same phenomenon for the case of release of particles induced by
hydrodynamic forces (hydrodynamically induced release).
Chapters 5, 6 and 7 focus on the mathematical modeling of the migration
of fine particles. Chapter 5, discusses in a quantative manner whether, fine
particles will get entrapped (plugging) in or move through (piping) the porous
media. Criteria are presented to predict whether piping or plugging would
occur along with practical applications such as soil erosion and sand filtration.
In Chapter 6, macroscopic balance equations are presented to model the
permeability reductions due to migration of fine particles. Both the mass and
population balance equations are presented along with their solutions and
applications. Chapter 7 uses network modeling approaches and presents the
permeability reduction models primarily due to size exclusions.
The prevention of the migration of fine particles is an important practical
consideration and this subject is discussed in Chapter 8. This chapter presents
the various prevention techniques (known as clay stabilization techniques)
developed in recent years in an unified manner.
Chapter 9, the last chapter, discusses soil pollution due to migration of
fine particles. Both, the facilitated transport of organic contaminants and the
migration ofbiocolloidal contaminants are discussed in terms of mathematical
models.
We thank our publisher for showing patience when we failed to meet the
deadlines and her unfailing co-operation. We also wish to thank the Kluwer
editors for making valuable suggestions to the contents of the book. KCK is
indebted to the porous media research group of Prof. Scott Fogler, at the dept.
of chemical engineering, the University of Michigan, for providing the milieu
to learn and work on the topic both during the periods of graduate studies
and visiting assignments. We are indeed thankful to Mr. S.T. Jambigi, for
meticulously word-processing the manuscript. We also thank Ms. Ingrid Ward
for word-processing a part of the manuscript. We thank Dr. V. Ramchandran
for scanning all the figures. Two past members of the porous media group,
Dr. Ravi Vaidya and Dr. K.K. Mohan who have made research contributions
on the topic have also contributed to the making of this book in many ways.
We are thankful to them. We thank Mrs. Jaya Khilar for proof-reading the
pre-final version of this manuscript. Some of our colleagues and students both
PREFACE xiii
at lIT Bombay and at the University of Michigan have indirectly contributed

to the completion of this book. We wish to thank them all. Finally HSF thanks
his wife, Janet for supporting yet one more book-writing project. KCK thanks
his parents and his wife for their support and particularly his sons, Kunal and
Mrinal for being considerate during the writing this book.
KCK HSF
Mumbai Ann Arbor
CHAPTER 1
PRACTICAL CONSEQUENCES OF RELEASE AND

MIGRATION OF FINES IN POROUS MEDIA
1.1 Migration of Fine Particles in Porous Media: An Introduction
Fines or fine particles are small particles that are present in porous media.
The migration of fine particles in porous media is a challenging problem
of both scientific and industrial importance. By migration of fine particles,
here, we mean the entire sequence of occurrences of release or detachment
of fine particles present in the porous media, their motion with the flow, and
finally their capture at some pore sites or their migration out of the porous
medium. Many applications can be found in fields such as of Petroleum,
Geotechnical, Chemical, Environmental and Hydraulic engineering. The
practical consequences of the migration of fines can be either beneficial or
adverse in nature and will be discussed briefly in this chapter.
This monograph will focus on natural porous media, such as rocks and soil
mass and the fine particles present in them. The fine particles can be inorganic,
organic, or biological in origin. These fines generally possess the colloidal
characteristics of having a size of the order 1 /Lm and an electric surface
charge. During the flow of a permeating liquid through a porous medium, fine
particles attached to pore surfaces are released or detached under certain sets of
conditions. Colloidal and hydrodynamic forces are known to be responsible for
the detachment of these fine particles. Once the fines are released, the particles
move with the permeating liquid until they get retained at other locations in the
porous medium or exit the porous medium. The sites that retain the particles are
generally the pore constrictions, crevices, caverns, and regular pore surfaces.
Depending on whether the fines get captured at pore constriction sites or
whether the fines move out, two different sets of consequences occur. In the
first case, the porous medium can become plugged resulting in reduced flow
through the porous medium. In the second case, the porous medium may erode,
and thereby increase the porosity, resulting in structural failure in the porous
1
2 CHAPTER 1
medium. When the fines migrate over large distances, the distribution of the
size and content of the fines in the porous medium is altered. Both desired and
undesired consequences arise out of this migration phenomenon. Therefore,
while in some cases, techniques are designed to induce the migration of fines,
in other cases, techniques are developed to prevent the migration.
1.2 Practical Consequences in Petroleum Engineering
A widely recognized consequence of migration of fines in sandstone oil

reservoirs is known as water sensitivity of sandstones. Water sensitivity
can cause a drastic reduction in permeability resulting in severe decline
in oil production (Monaghan et aI., 1959; Moore, 1960; Hewitt, 1963;
Mungan, 1965). The problem manifests itself when relatively fresh water
contacts the formation sandstone. This phenomenon can be elucidated by
core flood experiment with Berea sandstone core, known as a water shock
experiment (Khilar and Fogler, 1983).
In a typical water shock experiment, a Berea sandstone core of 2.5 cm in
diameter and 5 cm in length is first vacuum saturated with a 0.5 M NaCl
solution and then placed in a core holder of a core-flood set up (Khilar, 1981).
Next a 0.5 M NaCl solution is passed through the core in the axial direction
for a period of time well beyond the time the flow is stabilized. Next the flow
through the core is abruptly switched from salt water to fresh water, and the
pressure drop across the core is monitored as a function of time. From the
pressure drop, one can determine the permeability K. These data are reported
in terms of the fraction of original permeability (KlKo) as a function of pore
volumes of fresh water passed through the core. Figure 1.2.1 presents a typical
set of such data. We observe from the Figure 1.2.1 that the overall permeability
of the core drops drastically after only one or two pore volumes of fresh water
have entered the system. Similar observations have been made when the flow
is abruptly switched from a high salt concentration solution to a lower salt
concentration solution (Khilar, 1981). Analysis of the effluent collected during
the permeability decline showed the presence of fine clay particles suggesting
that the migration of fines plays a role in this reduction (Khilar, 1981; Gray
and Rex, 1966). Similar experiments have been conducted to study water
sensitivity phenomenon in Hopeman sandstone (Lever and Dawe, 1984).
Water sensitivity can occur in sandstone reservoirs in oil field operations such
as drilling, work-over, acidization, steam and water flooding and enhanced
oil recovery. It can also occur when the water from the natural spring or
water aquifer permeates into the oil formation zone as well as in other in-situ
PRACTICAL CONSEQUENCES OF MIGRATION OF FINES 3
1. 00 ~--------I r-~-------------,
0.80
3°10 Noel Solution Fresh wo~er

0-60
KIKo
0-40
0.20
0·0 2 4 6 50 52 54 56 58 60
Pore volume
Figure 1.2.1. Permeability reduction in water sensitivity of sandstones (Khilar

ana Fogler, 1983).
operations such as solution mining, and hydraulic fracturing etc. It may be

noted here that plugging or entrapment of fines is preferred in some cases such
as flow profile modification in reservoirs and the containment of wastes in land
fills having clay barriers.
1.3 Importance in Geotechnical Engineering
The migration of fine particles present in both natural and compacted soil can
cause internal and surface soil erosions. These types of soil erosions have
been known to cause serious problems leading to the failure of earthen dams
and roads (Aitchinson and Wood, 1965; Parker and Jenne, 1967; Sherard et
aI., 1972). The fine particles which are primarily clayey fines are released by
colloidal and hydrodynamic forces generated by the seepage flow of water and
then these fines are carried away with the seepage stream, causing soil erosion.
This phenomenon is also known in geotechnical literature as "piping" which
leads to tunnel erosion.
Figure 1.3.1 shows a schematic illustration of "piping" and "tunnel" erosion
across a dispersive clay layer. The potential for this problem to occur is
particularly acute in solids containing dispersive clays. Piping and tunnel
erosion are usually caused from a leak resulting from a construction defect.
In some instances, however, piping can occur without being initiated by the
4 CHAPTER 1
Figure 1.3.1. A schematic illustration of piping and tunnel erosion across a

dispersive clay layer (Sherard et aI., 1972).
presence of a leak. In these cases, it begins by surface erosion of soil at the

seepage exit face and then slowly propagates upstream.
The phenomenon of piping has received considerable attention (Sherard et
aI., 1972; Arulanandan et aI., 1975; Hjeldness and Lavania, 1980; Arulanandan
and Perry, 1983; Sherard et aI., 1984; Khilar et al., 1985) and is discussed
further in a rather detailed manner in Chapter -5. Criteria based on the soil
characteristics and soil permeability and evaluation tests have emerged from
these studies.
Soil erosion by piping takes place in a relatively unconsolidated subsurface
soil. However, it is a secondary erosion process compared to that induced
mechanically by rain and wind. In many soil erosions, such as channel
and stream bank erosion and surfacial rainfall erosion, the phenomenon of
migration offines can augment the erosion process.
The extent of the erosion process is strongly dependent on the soil
characteristics, particularly relating to the texture, the organic matter content,
the structure, and the permeability (Goldman et aI., 1986). The texture of a
soil mass is related to the sizes and the proportions of the sand, silt and clay
particles making up a particular soil. Higher clay content soils can be of a
dispersive nature, and hence soil erosion due to migration of fine particles
may occur. On the other hand, higher clay content soils can be thought of as
sticky and hence will not greatly affect other kinds of erosions. Likewise, loose
structure and higher penneability soils are more susceptible to soil erosion due
to migration offines. In another geotechnical problem, the decrease in hydraulic
conductivity in seawater-freshwater interface in coastal aquifers is caused by
the migration of fine particles (Goldenberg et aI., 1983). This problem appears
to have some resemblance to the problem of water sensitivity in sandstones.
It can be noted here that the piping of released fine particles is desired in some
operations such as regeneration of filter beds or in acidization of petroleum
reservoirs.
1.4 Effects in Environmental Engineering
The migration offine particles in ground soil can cause soil and ground water
pollution. There can be two general mechanisms. In one, the fines themselves
are pollutants such as organic waste particulates and! or micro-organisms by
which pollution occurs. These fines, when detached from the sites to which
they are adhered, move with the seepage flow until they re-adhere at other
sites, thus spreading pollutants. In the second mechanism, the migrating fine
particles act as carriers of pollutants that adsorb on to their surface.
Pollution or contamination of soils can also occur through the accidental
release of materials on the surface or through direct introduction of
contaminants into the subsurface and by the presence of micro-organisms.
Contaminants in the soil can migrate by attachment to colloidal fines, such as
clay particles or to biocolloidal fines, such as bacteria and viruses (Kaplan et
aI.,1993).
The movement of colloidal fines such as clay particles, hydrophobic organic
compounds, as well as biocolloidal fines such as bacteria and viruses, has
been shown to cause groundwater pollution (McDowel-Boyer et aI., 1986).
Mobile colloidal fines in ground water can adsorb organic and inorganic
contaminants and stabilize them in the moving stream. Hence mobile colloids
increase the amount of contaminants that flow of ground water can transport
(McCarthy and Zachara, 1989; Corapcioglu and Jiang, 1993). Figure 1.4.1
shows a schematic diagram ofthe movement of contaminants through colloidal
fines in an aquifer medium. Failure to account for this mode of transport can
lead to serious underestimates of the distances that contaminants will migrate.
For example, at a disposal site at Los Alamos, radioactive Plutonium and
Americium disposed of at a liquid seepage site, migrated up to 30 m. This
distance is much higher than the predicted migration of a few millimeters
based on laboratory measurements where migration of colloidal fines was not
present. At another site at Los Alamos, contaminants were found to migrate
6 CHAPTER I
Contam nant movement

In groundwater
(8) T'MrJ)hase sYSlem
Mobile
phase
t
.~ _ __
(groundwaler)
(b) Three-phase system
Figure 1.4.1. A schematic diagram of movement of contaminants (McCarthy

and Zachara, 1989).
260 Code
o Electro deposition primer
% 240
"0 D. Styrene butadiene lote.
:::: 220 o Human albumin
'; 200 • Gelatin I@) 70'C)
..... o
g. 180
~. 160 o
::>
o
:;:: 140
.,
+-
~ 120
...2
100
,... 80
:; 60
40
20
0
1 2345678910 20 30 4050 80
Concentration, wt '%:. solids
Figure 1.5.1. Flux vs. concentration plot (McCabe et at, 1993).
over a mile. Furthermore, the migrating contaminants were shown to be present

as colloidal fines by ultrafiltration analysis (McCarthy and Zachara, 1989).
1.5 Importance in Chemical Engineering
Unconsolidated porous media are widely used in many chemical engineering

operations and processes such as deep bed filtration and washing of filter cakes
etc. In designing of these operations, consideration is given to prevent the
occurrence of entrapment or plugging of the bed. For some operations, such as
regeneration of deep bed filters, a complete washout of the captured particles
is required.
In the operation of deep bed filters, the suspension concentration is usually
kept low (Svarvsky, 1985) to prevent plugging of the filter bed during the
filtration operations as well as allowing a reasonably complete washout of
fines during the regeneration of the bed. The adverse effects of plugging
normally occurs at higher concentration of fine particles. Likewise in the use
of membranes in microfiltration, the plugging and/or fouling of membrane
may be caused partly due to migration of fine particles into the membrane.
Figure 1.5.1 shows the flux as a function of concentration for sub-micronic fine
particles. We observe that the flux decreases as the concentration of the fines in
suspension increases. At concentrations higher than a threshold concentration,
the flux goes to zero, indicating the possible plugging of the membrane. A
similar observation has also been reported in protein microfiltration (Herreo
8 CHAPTER I
et. al., 1997).

In cake filtration, fine particles in the suspension are captured in the cake
layer through which the filtration takes place. In some cases, particularly, for a
wide size distribution of fine particles present in the suspension, migration of
the finer particles present in the suspension may lead to plugging in the cake
layer, thereby adversely affecting the filtration process. Similar occurences can
also surface during washing of cake build-up.
CHAPTER 2
CHARACTERIZATION OF PORE SPACE AND FINES
Knowledge of the size, shape, and other characteristics of pores and of the
fine particles is essential to the understanding of the various processes that
occur in the release and migration of fines in porous media. This knowledge
can be obtained by pertinently characterizing the fines and the porous media.
A detailed characterization of this complex system is extremely difficult and
will not be attempted in this chapter. Instead, we intend to discuss only the
most relevant aspects of migration, such as the size of fines, size of pore
constrictions, surface charge of fines and porous surface, and the interactions
between the permeating liquid and the porous media. Furthermore, in order to
keep the task manageable, we confine our attention to migratory fines in two
important classes of porous media: sandstones and soil. The procedures and
the other aspects of this characterization, nevertheless, can be applied to fines
migration in a number of other porous media such as ground water flows.
2.1 Pore Structure
A porous material/medium contains voids (free of solids) dispersed in a solid

matrix. At least a part of this void space must be continuous to allow the fluid
to permeate from one side of the medium to the other. Porous media may
be considered equivalent to systems in which the solid and void phases are
randomly dispersed in such a way that both phases form continuous conducting
paths through the medium. It is the void phase or space that is important to our
subject of migration offines in porous media. Void space is generally known as
the pore space as it is known to consist of randomly distributed pores of various
shapes and sizes(Scheidegger, 1974). By a "pore ", we generally mean some
sort of a conduit of some shape such as cylindrical although an exact description
is difficult. The pores are either interconnected to form a continuous phase or
they are non-interconnected and form dead-end or blind pores (Scheidegger,
1974; Dullien, 1979a). The interconnected pores form the effective pore space
9
10 CHAPTER 2
Cubic RnombohNral
c..bic Rhombohedra I
Figure 2.1.1. Two types of packing of spherical particles and corresponding

pore spaces (Collins, 1990).
through which the transport of matter can take place, and the migration of
fines by and large occurs in these interconnected pores. The porous media,
such as sandstone, and soil mass, in which migration of fines generally occurs,
are formed by deposition and compaction of grains. Unconsolidated porous
media such as packed beds, sand filters, etc., are also made of packing of grains
and other materials. Both these unconsolidated and natural porous media can
be roughly approximated as a network of pore chambers, i.e., pore bodies
connected through pore throats (i.e. pore constrictions) (Lymberopoulos and
Payatakes, 1992). The concept of pore chambers and pore constrictions can
be better appreciated by visualizing a simple packing of rigid spherical grains.
For example, consider the most stable rhombohedral packing of spheres which
has void fraction, ¢ (ratio of void volume to total volume) equal to 0.2595. The
pore spaces shown in Figure 2.1.1 can be approximated as consisting of pore
chambers and pore constrictions that are connected in more than one direction.
We can characterize the pore space or pore structure geometrically by
two simple microscopic parameters: pore chamber size distribution and pore
constriction size distribution. It is to be noted here that there are several
CHARACTERIZATION OF PORE SPACE AND FINES 11
other geometric models proposed for the structure of the porous media,
which include spatialy periodic models, network models and Bethe lattice
models, etc. (Adler, 1992; Sahimi,1995a). Topological parameters such as
chamber to constriction coordination number are also useful to the analysis
of the phenomenon of migration of fines in porous media. Other topological
microscopic parameters and their determination and measurement techniques
are a subject of current research (Lymberopoulos and Payatakes, 1992;
Sahimi, 1995b).
2.1.1 PORE CONSTRICTION AND CHAMBER SIZE DISTRIBUTIONS
The pore constrictions, also known as pore throats or pore necks, are the
narrowest segments of pores while the pore bodies or pore chambers are the
widest segments of the pores. Figure 2.1.2 shows SEM (Scanning Electron
Micrograph) pictures of thin sections of some porous materials. The black
and white patches or areas in this figure indicate pore space and solid space
respectively. Figure 2.1.3 shows optical micrograph of a soil mass in which
white patches are the pores. We observe from these micrographs that the pores
have wider and narrower sections and that they are of different sizes.
Some data exist on the distributions of the size of pore constrictions and
chambers for porous media such as sandstone and soil mass. A sample of such
data is presented in Table 2.1.1. We observe from this table that the constriction
size is, in general, of few microns and the chamber size is, in general, an order
of magnitude higher than that of constriction.
While the pore constriction size distribution can be measured by some well-
developed techniques, the pore chamber distribution is difficult to measure.
The most popular methods for measuring pore constriction size distributions
are mercury porosimetry and optical methods. Descriptions of these and other
techniques, including the sorption method for sizes below 10 /lm, are presented
in a number of treatises on porous media (Dullien, 1979a; Scheidegger, 1974;
Collins, 1990) . Mercury intrusion porosimetry is an established technique,
but unfortunately it can only yield pore entry size distribution (Dullien,
1979a). The optical methods in combination with serial sectioning can
be used to obtain the chamber size distribution. Knowing the chamber
size distribution, the constriction size distribution can be determined from
the mercury intrusion-retraction curves (Tsakiroglou and Payatakes, 1990;
Tsakiroglou and Payatakes, 1991; Lymberopoulos and Payatakes, 1992). Other
methods such as small angle X-ray scatterings (SAXS) as well as nuclear
magnetic resonance (NMR) have also been used to determine the pore size
12 CHAPTER 2
Figure 2.1.2. Electron micrographs of some porous materials (Dullien, 1979a).
Figure 2.1.3. Optical micrograph of a soil sample (Biswas and Mukhetjee,

1994a).
TABLE 2.1.1. Pore constriction and chamber sizes of some sandstones
Sandstone Pore constriction Pore chamber Reference

size (pm) size (/Lm)
Berea 0.5 to 5.0 5 to 50 Dullien and Dhawan (1974)
Boise 39-74 Ioannidis and Chatzis (1993)

Bartseville 19-44 Ioannidis and Chatzis (1993)
STl 33-120 33-250 Lymberpolous and

Payatakes (1992)
distribution in rocks (Sahimi, 1995b).
2.1.2 PORE CONNECTIVITY
The connectivity of a pore network is generally characterized by a coordination

number which represents the number of independent paths through which pores
are connected to each other. The connectivity of pore space can be precisely
characterized by Betti numbers (Sahimi,1995b). Serial tomography has been
applied to determine the coordination number of porous rocks and sintered
granular materials (Yanuka et aI., 1984; Lin and Slattery, 1982). An average
coordination number representing the connectivity between constrictions to
chamber is an important parameter. In other words, we would like to know,
on an average, how many pore constrictions are connected to a pore chamber.
Likewise, there can be a coordination number which represents an average
number of pore chambers connected to a pore constriction. A reasonable
estimate of an average coordination number for sandstones is between 4 to
8 (Lin and Slattery, 1982; Yanuka et aI., 1984). The larger the coordination
number, the less likely that all constrictions will get plugged. A knowledge of
pore connectivity is particularly useful in network modeling of porous media.
This is discussed in Chapter 7. The concept of connectivity has also been used
to classify the porous media (Bear and Bachmat, 1990).
14 CHAPTER 2
2.1.3 POROSITY AND PERMEABILITY
The macroscopic models for the migration offines in porous media (presented
in Chapters 5 and 6) are based on two important macroscopic parameters of
porous media: porosity and permeability.
"Porosity", also known as void fraction, is the fraction of the total volume
of the porous medium that is occupied by void space or pores. The porosity
of naturally consolidated porous media varies from a low value of 0.05 to a
high value of 0.45. There are a number of standard techniques to measure the
porosity (Scheidegger, 1974) one of which is mercury injection. Scheideggar
(1974) presents a list of representative values of porosity of a number of porous
materials.
"Permeability" of an isotropic porous medium is the proportionality constant
in the local relationship between the pressure gradient and the seepage flow
velocity vector. This relationship was discovered experimentally by Darcy.
Darcy's law is given as
K
v=--\lP' (2.1.1)
- JL-
Here K is the permeability of the porous medium, and JL is the viscosity of the
permeating fluid. This linear law has been rationalized theoretically (Slattery,
1972; Greenkom, 1983). A brief but excellent discussion on Darcy's law and
the concept of permeability is given elsewhere (Barenblatt et aI., 1990).
The dimensions of permeability are (length?, and the practical unit of
permeability is Darcy, with 1 Darcy = 0.987 JLm2 . A porous medium is said to
have a permeability equal to 1 Darcy if a pressure drop of 1 atm across a cube
of a length of 1 cm produces a flow rate 1 cm3 S-1 of a fluid with a viscosity of
1 cPo Permeability can be measured by standard set-ups such as a falling head
permeater (Dullien, 1979a). The range of permeabilities for porous media,
where migration of fines usually occur, is typically from 1 mD to 1000 mD.
2.2 Pore Surface
The pore surface plays an important role in the release and capture offines. The
nature of attachment of fines to the pore surface, the nature ofliquid film present
next to the pore surface, the physical, chemical and physiochemical interactions
between the pore surface and the permeating liquid, etc., are dependent on
the several characteristics of the pore surface. Three major categories of
characterization, will be discussed: surface, colloidal, and chemical.
2.2.1 SURFACE CHARACTERIZATION OF PORE SURFACE
The wettability and roughness of the pore surface can influence the process
of release and capture of fine particles. The surface roughness is believed to
affect the adhesion between the pore surface and the fine particles as well
as the detachment process (Sharma et aI., 1992; Das et aI., 1994). Surface
roughness influences the interaction energy and hence the forces of adhesion
(Swanton, 1995). Das et al. (1994) have shown that the surface roughness
is responsible for generating restraining torque during the detachment of
particles by hydrodynamics forces. This detachment process will be discussed
in Chapter 4. In general, the pore surfaces in many natural porous media
in which migration of fines occurs are rough as can be seen with a scanning
electron microscope or an atomic force microscope. Fractal concepts have been
used to characterize the surface roughness of the pore surface (Sahimi,1995a).
The wettability of the pore surface is defined as the tendency of one fluid
to spread on or adhere to the solid surface in presence of other immiscible
fluids (Craig, 1971). It is known to greatly affect most capillary hydrodynamic
phenomena involving oil and water flow in porous media. Natural rocks or,
for that matter, solids in general can be water-wet, or oil-wet or can have
fractional wetting characteristics (Anderson, 1986a; Dullien, 1979b). By water-
wet we mean that there is a tendency for water to occupy the small pores and
to contact the majority of the rock surface. Oil-wet can be defined in an
analogous manner. Natural rocks of fractional wettability have part of the
surface water-wet and part oil-wet. A special type of fractional wettability is
the mixed wettability - rocks having mixed wettability contain oil-wet surfaces
in larger pores and water-wet surfaces in smaller pores (Anderson, 1986a).
Wettability can be measured quantitatively using techniques such as contact
angle, the Amott method and the USBM method (Anderson, 1986b). Almost
all clean sedimentary rocks are strongly water-wet and, in most cases, contain
deposited species on the surface that change the wettability. The wettability
characteristics of pore surface are particularly important in migration of fines
caused due to water sensitivity phenomenon. Fines adhering to strongly water-
wet pore surfaces are more likely to be released due to the contact with water.
16 CHAPTER 2
60~~~--~-T--~-r--r-~-.
50 UNBAKED BEREA SANDSTONE

CRUSHED OXIDE
40 o 0.1 M CaCI 2
_ 30
>
..§. 20
.-J
« 10
f=
~ 0
b
a.. -to
«
W-
N
20
-30
-40
-50
-602 3 4 5 6 7 8 9 10 11
pH
Figure 2.2.1. Variation of zeta potential of crushed Berea sandstone with pH

(Sharma et aI., 1987).
2.2.2 COLLOIDAL CHARACTERIZATION OF PORE SURFACE
Colloidal characterization refers to surface charge or the streaming potential

of the fine-free porous medium. This information is essential in the estimation
of forces between the pore wall and the attached fine particles.
The streaming potential (Es) is the electric potential across a porous medium
that develops when an electrolyte flows through the porous medium under
steady applied pressure difference, b..P. The relationship between the zeta
potential (0 of the pore wall and the streaming potential (Es) can be derived
assuming that the mobile part of the double layer contributes to the current
flow (streaming current) (Hiemenz, 1986) and is given as
c(b..P
(2.2.1)
Es = J-L [kb - (2k s )j R] .
Here E is the pennittivity of the liquid solution in C2 J- I m- I , kb is the specific

bulk conductance of the liquid solution in 0- 1 m- I , ks is the specific surface
conductance of the pore surface in 0 -I, and R is the pore radius in m. The
zeta potential is the potential near the solid-liquid interface and is described
in the next section. The streaming potential Es can be easily measured (Shaw,
1969).
The zeta potential and surface charge density of powdered Berea sandstone
samples have been reported (Kia et aI., 1987a; Shannaet aI., 1987; Shanna and
Yen, 1984). Figure 2.2.1 shows the zeta potential of grains of Berea sandstone
in 0.10 M CaCh solution at different pH values. We observe from this figure
that zeta potential is strongly dependent on pH and the potential changes from
positive to negative values at higher pH.
Even at a relatively strong ionic strength ofCaCh, the zeta potential remains
negative for pH above 3. Kia et aI. (1987a) have also measured the zeta
potential of fine-free Berea sandstone powder at different pH as a function of
concentration of sodium chloride solution.
Fine-free Berea sandstone samples were obtained by exposing a very thin
slice of sandstone to an alternate flow of salt solution and water many
times to release and washout the fines. Figure 2.2.2 shows that the values
of zeta potential vary from 0 to -10mV at high concentrations of sodium
chloride solution and at low pH. The zeta potential increases to -40 mV at low
concentrations and at high pH.
Figure 2.2.3 shows the surface charge per unit weight of sample of Berea
sandstone as a function of different concentrations of KCI. The surface charge
remains positive up to a particular pH value, and then it becomes increasingly
negative with increasing pH. These variations are typical of those found in
sandstones and have been satisfactorily explained by using an adsorption-
surface complex model (Shanna and Yen, 1984; Mohan, 1996a).
The surface charge density can also be estimated from cation exchange
or anion exchange capacity data (van Olphen, 1977). It should be noted that
usually fines are good ion exchangers, therefore, in order to obtain infonnation
of pore surface, the exchange capacity measurements must be conducted with
porous materials that are free of fines.
2.2.3 CHEMICAL CHARACTERIZATION OF PORE SURFACE
Chemical characterization of pore surfaces focuses primarily on the

mineralogical composition and can be identified by several well-established
techniques such as X-ray diffraction, electron diffraction, and differential
18 CHAPTER 2
pH
-
-60 A 9.5
... 8.6
-80
~ 0 6.5 >
....'",;
+ 4.5 -60.§
E -40 (jj
"!- ;::
E -40 ~
oS 0
a..
.~ -20 <II
~ -20 ;
0 -+-+-+-+ N
:::< -+--+
0.0 0.02 0.04 0.06 0.08 0.1

Salt Concentration (gmoi/lit.)
Figure 2.2.2. Mobility and zeta potentials of sandstones in NaCI salt solutions
at different pH (Kia et al., 1987a).
10
UNBAKED BEREA SANDSTONE I pzc' 7.81
aiel"> O~------.>-=---.,'- " ' . - - - - -
-~
-IO'--_~_~_.J'__~ ..........~__'___'_ _~~_

o 5 10 14
pH-
Figure 2.2.3. Surface charge vs pH plot for the unbaked Berea sandstone
sample (Sharma and Yen, 1984).
thennal analysis.
The mineralogical composition of fine-free sandstone pore materials such
as Berea sandstone is quartz (~83% by weight), feldspars (~4% by weight)
and carbonates, namely calcite and siderite (3% by weight) (Khilar, 1981).
The inorganic constituents of soil mass contain primarily the quartz,
feldspars, micas, silica, calcium carbonates, and free oxides of silicon, iron
and aluminum. The organic matter which constitutes between 0.40 to 10% by
weight contains organic compounds of plant and animal origin decomposed
over the years, known as soil humus. The micro-organism contents, known as
soil microflora constitute only 0.01 to 4 % by weight of the soil (Biswas and
Mukherjee, 1994b).
2.3 Fines
By fines, here we mean the small discrete solid or solid-like particles

indigenously present in natural porous media which can be mobilized by
means of colloidal, hydrodynamic, and other forces. These fines are tenned
as migratory fines. The origin of most of these fines is linked to porous
almino-silicate materials that constitute the bulk of the natural porous media.
Primary silicates are fonned through natural processes which produce a second
generation of inorganic minerals known as secondary silicates, oxides of iron,
aluminum and silicon as well as carbonates and other salts. These secondary
silicates and oxides fonn fine discrete particles that are either adhered to
the pore material or are attached by means of a coating and cementation.
In addition to these fines, which can be tenned as inorganic fines, there
can be fines of organic and biological materials. The biological fines consist
primarily of the micro-organisms present in the porous media while the organic
fines are comprised of organic matter. The migration of these types of fine
particles is the subject of this book. While the focus is on inorganic fines
due to their preponderance, biological and organic fines are also discussed.
The migratory fines in porous media such as sandstones and soil mass, are
essentially clay minerals (crystalline layer silicates). In addition, soil masses
also contain other kinds of fines of inorganic oxides, organic materials and
micro-organisms. In this section, we shall discuss the characterization of three
different aspects important to analysis of migration of fines : geometrical,
colloidal, and chemical.
20 CHAPTER 2
2.3.1 GEOMETRICAL CHARACTERIZATION FINES
Two geometrical parameters of fine particles, the shape and size, playa crucial
role in determining whether or not plugging will occur. Therefore information
of the geometrical aspects is important to the analysis of migration offines in
porous media. The migratory fine particles are generally clay particles such as
kaolinite, montmorillonite, illite and mica as well as fine sand particles, organic
and biological moities. In general, these particles are not of regular shape, and
defining a size becomes difficult. Yet, we must use a relevant measure of size
for our analysis. Because the suspended particles normally orient their larger
dimension generally along the direction of the flow, a relevant size in our case
is the size in the direction normal to the longer dimension. Another approach
would be to have a complete 3-dimensional geometrical characterization which
is difficult and may not be really necessary.
The shape of the fines can be determined by scanning electron microscopy.
The most common shapes of migratory clayey fines are those of clay minerals
as shown in Figures 2.3.1 to 2.3.4. Figure 2.3.5 presents a scanning electron
micrograph offine particles eluted from Berea sandstone during a water shock
experiment.
We observe from these figures, that the shapes can be described as platelets,
blades, needles and flakes. The organic and biological migrating fines as well
as the fine sand particles are not as irregular as the clayey fines. None of the
migrating fines can be really considered spherical in shape.
In general the the largest dimension as well as the other dimensions of the
migratory fines varies from 0.1 to 10 /Lm. The size of the largest dimension of
the migratory fines in Berea sandstone, for example, was found to vary from
0.1 to 5 /Lm (Khilar, 1981). Likewise, the size of the migratory fines in soil
mass is usually below 5 /Lm. Seldom are migratory fines larger than 10 /Lm.
The size of the fines can be measured by means of a number of standard
techniques, such as microscopy, dynamic light scattering and the Coulter
counter method. Microscopy is generally favoured as it gives size and shape
in a somewhat direct manner whereas the other techniques give sizes derived
from the data dependent on other properties.
The first step in the characterization of migratory fines is to collect in
sufficient amount from the porous medium under consideration, and this
procedure usually is not an easy task. For some porous media, such as water-
sensitive sandstones and eroding soil masses, the fines are commonly collected
from the effluent samples during the core-flooding experiments. The samples
are then filtered with a 0.10 /Lm filter paper to separate the fine particles. For
Figure 2.3.1. Electron micrograph of kaolinite (Grim, 1968).
Figure 2.3.2. Electron micrograph of illite (Grim, 1968).

22 CHAPTER 2
Figure 2.3.3. Electron micrograph ofsaponite (Grim, 1968).
Figure 2.3.4. Electron micrograph ofattapulgite (Grim, 1968).

Figure 2.3.5. Electron micrographs of clay particles eluted from Berea

sandstone (Khilar and Fogler, 1983).
less permeable porous media, fines remain inside. In these cases, thin-section
analysis may give information on fines. The size in a suspended state may
differ from that in a dry state.
2.3 .2 COLLOIDAL CHARACTERIZATION OF FINES
Colloidal characterization concerns itself with the physical size and the surface
charge of the colloidal materials. Size characterization has been discussed in the
preceeding section. In this section we shall discuss the surface electric charge.
Fine particles, in general, carry a surface charge, the origin of which depends on
how the fines were formed. For suspended fine particles, the physicochemical
24 CHAPTER 2
interactions between the fine particles and the suspending solution also govern
the surface electric charge.
Clayey fines can acquire a surface charge by three different mechanisms:
isomorphic substitution, broken bonds - particularly at the edges - and
protonations and deprotonations of surface hydroxyls (Grim, 1968; Schechter,
1992). The surface charge on the organic and biological fines in solution is
generally acquired through the adsorption of specific ions from the solutions
and dissociation of functional groups present on the surface. Whatever may
be the origin, it is a fact that most migratory fines in solution carry a charge,
and hence colloidal forces playa dominant role in the processes of release,
migration, and capture of these fines. In order to estimate the magnitude of
these colloidal forces, it is necessary to determine the surface charge of the
fines.
The surface charge resulting from lattice substitution and from broken bonds
can be determined by measuring the ion-exchange capacity (van Olphen, 1977;
Helffrich, 1962). A negatively charged surface can exchange its portable
cations present on the surface for cations present in the solution. The amount
exchanged at equilibrium is known as the cation exchange capacity, i.e. CEC.
Likewise, we can define anion exchange capacity, AEC, of positively charged
fines. Cation exchange capacity can be measured by standard techniques such
as pH titration and NHt or Ba++ saturation techniques (Helffrich, 1962;
Grim, 1968). Knowing the exchange capacity, which is usually measured
in terms of meqgm~l, and the surface area, the surface charge density can
be estimated in Cm~2 (van Olphen, 1977). The cation exchange capacity
of several clayey minerals is presented by Grim (Grim, 1968). For migratory
clayey fines such as kaoline and illite, the CEC is the order of3 to 15 meq gm ~ 1 .
The more relevant and useful method of surface charge characterization offines
is, however, the zeta potential measurements offines at various ionic conditions
of the permeating solution.
The zeta potential, ( of a fine particle can be measured by the technique
of electrophoresis. Here, the velocity of a suspended charged particle in an
applied electric field is measured. The concept of zeta potential is linked with
this electrophoretic motion of the particles. The potential of the electric field
generated due to charge of the particle at the shear plane (the plane at which the
velocity begins to decrease) is taken as the zeta potential. The zeta potential is
the potential at the "effective" solid-liquid interface and can be conveniently
measured by commercial zeta meters. Detailed discussions on zeta potential
and its measurement techniques can be found elsewhere (Hunter, 1981; Shaw,
-50 • Co-clay + No-cloy - -70

"in 0 Co-kaolinite 0 No-kaolinite - -GO ~
<: -40 >
E - -50..§.
l-30 ~~~
E
- -40 ~~
-: -20 r- ~---<>---"""'-<>---~----=l- -30 ~
.~- - -20 ::,
:g::E -10 -
...IL ~ ..., -10
Q)
N
o
Q02 0.04 0.06 0.08 0.10
Molar Salt Concentration
Figure 2.3.6. Variation of zeta potential of fines with salt concentration. (Kia
et aI., 1987b).
1969). Recently, it is shown that corrections are necessary if the particles

carry a non-uniform surface charge as well as for the possible polarization
of the double layers(Fair and Anderson, 1989; Solomentsev et aI., 1993).
The polarization effects resulting from the transport of counterions within the
double layer tend to reduce the electrophoretic mobility. A non-uniform surface
charge on the particle makes the electrophoretic motion complex. There is a
rotation of the particle, and the direction of the particle motion is not necessarily
in the direction ofthe applied field. The correction for the non-uniform surface
charge, is particularly important for the clayey fines. Kaolinite, for example,
is known to have two different surface charges - flat surface charge and edge
surface charge.
Kia et al. (1987b) measured the zeta potential offines eluted from thin Berea
sandstone cores during water sensitivity experiments. Figure 2.3.6 shows the
values of the zeta potential as a function of salt concentration. We observe from
this figure that the zeta potential is negative at the neutral pH and approaches
a constant value at higher salt concentration. We further observe that the zeta
potential of the fines in calcium chloride solution is significantly lower than that
in the sodium chloride solution, possibly because of the stronger adsorption
of Ca2+ ions on the clay surface as compared to that of Na+ ions. Similar
quantitative and qualitative results are obtained for Berea sandstone clay
particles indicating that the migrating fines in Berea sandstone are essentially
kaolinite particles.
Like sandstones, the migratory fines in soils are, in general, clay particles.
26 CHAPTER 2
The zeta potential of some clay minerals such as kaolinite have been studied
(Williams and Williams, 1978; Kia et aI., 1987a). Kaolinite particles are known
to carry a negative surface charge on the faces and a positive charge on the
edges in a certain pH range (van Olphen, 1977). The face zeta potential varies
from approximately -35 mV to -25 mV at a pH of 6.5 in the range of sodium
chloride concentration between 10-4 and 10-2 M. In this range and at this
pH, the edge zeta potential varies approximately from 10mV to 5 mV (Kia
et aI., 1987a). The potentials, however, vary with pH. At pH 3 or lower, both
potentials are positive. Corrections for the non-uniform surface charge were
not made in the calculations for the above potentials.
Organic and biocolloidal fines particularly present in underground soil may
also carry a charge. Zeta potentials of these fines are expected to have the same
qualitative variations with salt concentration and pH as the clayey fines.
2.3.3 CHEMICAL CHARACTERIZATION OF FINES
Chemical characterization of fines implies the elemental and mineral

compositional analysis of migratory fines in porous media. In addition
to classical chemical methods, some standard techniques such as X-ray
diffraction, electron diffraction, and differential thermal analysis can be used.
Migratory clayey fines consist primarily of kaolinite, illite, montmorillonite,
and chlorite particles of authigenic origin (Schechter, 1992). The chemical
composition of these clay minerals is given in Table 2.3.1. We observe from
this table that silica, Si02 and alumina, Ah03 are the major minerals.
TABLE 2.3.1. Range in chemical composition of migratory clays
Clay Si02 Al 2 03 Fe203 ri0 2 CaO MgO K20 Na20

n
~
Kaolinite 40-49 35-40 0-13 0-2 0-0.8 0-0.50 0-0.10 0-0.20 ~
n
..,
tTl
Illite 50-56 18-31 2-5 0-0.8 0-2 1-4 4-7 0-1
~
~
Chlorite 31-38 18-20 35-38
~
o'Tl
Montmorillonite 45-55 19-50 0-3 0-2 ."
~
[/)
~
n
tTl
~
'Tl
Z
tTl
[/)
N
-.I
CHAPTER 3
COLLOIDALLY INDUCED RELEASE OF FINES IN

POROUS MEDIA
The release of fines from a pore surface is the first step in the phenomenon
of migration of fines in porous media. A comprehensive understanding of
this release process is necessary (1) to design treatments either to prevent
or to induce/trigger the migration of fines, and (2) to analyze and develop a
mathematical model to gain insights into the phenomenon. In general, two
major types of forces are responsible for the release of fines: colloidal and
hydrodynamic. This chapter focuses on colloidally induced release while the
next chapter focuses on hydrodynamically induced release.
The release of fines is treated in a manner analogous to chemical reaction
kinetics; that is, two complementary aspects - thermodynamics and kinetics-
are required for a complete description of the process. We have addressed two
similar important issues: 1. what initiates the release of fines (Statics) and 2.
what factors affect the rate of release (Dynamics).
3.1 The Statics of the Release Process
The statics of the release process will be analysed for a fine particle adhering
to a pore surface using potential energies of London-van der Waals attraction
and electrical double layer repulsion. This single-particle analysis has been
tested experimentally and has been shown to satisfactorily describe conditions
necessary for release (Khilar and Fogler, 1984; Khilar et aI., 1990). In this
approach, the total energy of interactions between a fine particle and the pore
surface on which the particle adheres needs to be determined.
Figure 3.1.1, shows fine particles adhering to the surface of a pore chamber
through which a permeating liquid is flowing. The distance of separation
between a fine particle and the pore surface may be very small (order of
10- 1 nm), and therefore we expect both short-range and medium-range forces
between macroscopic bodies to act. In addition, the fine particle will be
29
30 CHAPTER 3
Figure 3.1.1. A conceptual picture of a system consisting of fine particles

adhering to pore surface (Khilar and Fogler, 1987).
o
Figure 3.1.2. Sphere-plate and plate-plate conceptualization of fine-pore
surface system.
subjected to hydrodynamic forces of the flowing fluid.

Hence, there are several contributions to the total energy of interactions, and
if the net is repulsive, the fine particle may be released. Our objective is to
identify these different energies of interactions and discuss them in relevance
to migration offines. Furthermore, we intend to select appropriate expressions
for these energies to carry out a quantitative analysis.
Considering the fact that the equations for determining the various energies
COLLOIDALLY INDUCED RELEASE OF FINES 31
of interaction are only available for simple geometries, we need to idealize

our system. Such idealization can be achieved using the knowledge gained
through characterization of the fine-pore surface system - a topic discussed
in Chapter 2. The size offine particle is in the order of 1 /-Lm while that of pore
chamber is in the order of 10 /-Lm. We therefore neglect any curvature effects
and idealize the pore surface to be a flat plate of infinity extent. The shape
of the fines is irregular and can be of globular, platelet and needle type. For
the sake of obtaining equations for various energies of interactions, we shall
idealize the fine-pore surface system as a sphere-plate system or plate-plate
system as shown in Figure 3.1.2.
3.1.1 ELECTRIC DOUBLE LAYER REPULSION
We have seen in Chapter 1, that fines are released when the ionic strength of the
permeating solution is decreased. At this favourable condition for release, the
electric double layers formed at the pore surface and around the fine particle
expand and overlap. Such overlap of electric double layers of like-charge
surfaces gives rise to repulsive energy of interactions V DLR which tend to
expel the fine particles from the surfaces.
The double layer energy of interaction can be obtained by first
solving Poisson-Boltzman equation with appropriate boundary conditions to
determine the potential profile and then calculating the free energy change
occurred by bringing the two double layers from infinite separation to a small
separation. Therefore, to obtain a suitable expression for V DLR we need to
specify the geometry, the boundary conditions, and the type of electrolyte
(symmetrical or unsymmetrical) and make various other approximations. For
the sake of simplicity, we shall use a sphere-plate geometry. With regards
to boundary conditions, we must decide whether the pore and fine particle
surfaces maintain constant surface charge density, or constant potential, or
whether one of the surfaces (say fine particle) keeps its charge density constant
while the other surface keeps its potential constant (mixed case). We shall
present equations for all the three cases even though they are of the same order
of magnitude at moderately high potentials (Kar et aI., 1973). It is also further
shown that, in general, the constant potential case gives the lowest value for
VDLR while the constant charge case gives the highest value. Most equations
developed in literature, particularly those based on linerization of Poisson-
Boltzman equation, can be used for both symmetrical and unsymmetrical
electrolytes as well as for mixtures so long as the appropriate Debye lengths,
~-l are calculated properly (Israelachvili, 1992a).
32 CHAPTER 3
Equations for overlapping double-layer potential for a sphere-plate system

are given below.
Case 1: Constant potential case (Hogg et aI., 1966):
1 + exp( -~h)
VOLR = (wp/4)[2~ol~o21n (1 (
- exp -~h
))
(3.1.1)
Case 2: Constant charge case (Wiese and Healy, 1970):
1 + exp( - ~h) )
V OLR = (wp/4)[2~ol~o21n ( l-exp(-~h)
- (~51 + ~52) In (1- exp( -2~h))l· (3.1.2)
Case 3: Mixed case, surface 1 is at constant potential and surface 2 is at constant

charge density (Kar et aI., 1973)
(3.1.3)
Here, ap =particle radius, in m; h =distance of separation, in m; E = dielectric

constant, in J- 1 C2 m-1.' W ~ = ("'.t nf!e
t
2zt2/Ek B T)I/2 , in m-1., e = electron
charge, 1.60 x 10- 19 C;
kB = Boltzman constant, 1.38 x 10-23 JK- 1; zi = valence ofith ion; ni =
number concentration of the ith ion far away from the surface, in number m- 3;
T = absolute temperature in K.
The above equations present closed-form analytical expressions for VOLR
for three different boundary conditions. These equations are valid for low to
moderate potentials « 60 mV), when the thickness of double layer is smaller
than the fine particle size, ~ap ~ 1. Fortunately, these conditions prevail in most
situations of migration offines in porous media. The potentials ~Ol and ~02 can
be replaced by the measured values of the zeta potentials. Recently, it has been
noted that zeta potential sometimes underestimates the surface potential, and
surface potential estimated from surface complex - double layer models can
be used (Ryan and Gschwend, 1994). Equations (3.1.1) to (3.1.3) present three
simple expressions for VDLR out of the many proposed in literature (Hunter,
1989a). The double layer around a clay particle is complex and therefore
several models have been proposed to calculate the double layer repulsion
(Mohan, 1996a). For most purposes, this simple constant charge or constant
potential equation such as given in Equations (3.1.1) to (3.1.3) can be adaquate.
3.1.2 LONDON-VAN DER WAALS ATTRACTION
The London-van der Waals energy of interaction between two similar bodies is
usually attractive, and therefore it contributes to the strength of the attachment
between a fine particle and the pore surface. It is electrostatic in nature and
originates due to the interactions between permanent and/or oscillating dipoles
of atoms. This weak interaction energy between a pair of atoms when integrated
over a pair of microscopic bodies becomes substantial. Furthermore it decays
slowly and acts over distances as large as 10 nm.
In order to obtain a suitable expression for the London-van der Waals energy
we need to adopt a geometry of the system. From the discussion presented in
the preceding section, it is reasonable to assume the sphere-plate geometry.
The London-van der Waals energy of interaction is given as (Hunter, 1989b)
VLVA =
Am [2(1 + H)
--6- H(2+H) +In 2+H
(H)] ' (3.1.4)
where H = h / ap and A 132 are the Hamaker constant between a pore surface
(1) and a fine particle (2) separated by an aqueous medium (3). The negative
sign accounts for the fact that it is an attractive potential. Clearly, the major task
in calculating the VLVA (LVA stands for London-van der Waals) is to obtain,
an accurate value for the Hamaker constant, Am. The Hamaker constant for
two identical bodies in a vacuum (All or A22 etc.) is a function of molecular
properties such as polarizability, dipole moment, etc., and has been obtained
experimentally for many materials. These experimental values by and large
agree with theoretical calculations (Israelachvili, 1992b). The typical values
for solids or liquids are about 10- 19 J.
Simple approaches are also available to calculate the composite Hamaker
constant Am from the values of All,A22 and A33 (Hunter, 1989b). An
34 CHAPTER 3
expression for A132 is given as
1/2 - A33
A132:::::' ( Al1 1/2) (1/2 1/2) .
A22 - A33 (3.1.5)
We note from the Equation (3.1.5) that both Al1 and A22 must be greater
than A33 for A132 to be positive. Let us estimate A132 for a pore surface, (1) a
clayey fine particle, (2) and water, (3) system. The Hamaker constant for pore
surface (A l1 ) and for clayey fine particle (A22) are not available. We therefore
make a reasonable assumption that both Al1 and A22 can be taken as that of
fused quartz. The value for Al1 :::::' A22 is then approximately 6 x 10-20 J
(lsraelachvili, 1992b). The Hamaker constant for water (A33) is approximately
4 x 10- 20 J (lsraelachvili, 1992b). Using Equation (3.1.5), we obtain A132
equal to 2 x 10-21 J which is somewhat low. Values as high as 6 x 10- 21 J
have been used in calculations relating to the release of fines (Khilar et aI.,
1990). When the distance of separation is 10 nm or longer, the time it takes for
the electromagnetic waves to reach the particle/wall must be accounted for. A
suitable expression for this retarded interaction is given as (Ho and Higuchi,
1968)
A132 [
V LVA = --6- H +1aH2 ] for H ~ 1 and h < 15 nm (3.1.6)
and
VLVA
A132 [ 0.49
= --6- NRtdH2 -
2.17
15N2 H3
0.59
+ 35NRtd
3 H4
1 (3.1.7)
Rtd
for H ~ 1 and h > 15 nm.
Here, H = h/ap ; NRtd = (27ra p )/ Ae; a = 1.76NRtd and Ae is the

electromagnetic wave length. In the present system the size of fines, ap , is
ofthe order of 1 11m and the separation distance, h, is of the order of hydrated
size of the counter ion, say about 1 nm and hence H = h / ap ~ 1. Furthermore,
the retardation parameter NRtd is about 102 and therefore one can neglect the
effects of retardation on VLVA in the present case of release offines in porous
media. It may be mentioned here that the current trend is to calculate Hamaker
constant using methods based on Lifshitz theory (lsraelachvili, 1992b). An
appropriate expression is given as
(3.1.8)
where El, E2 and E3 are the static dielectric constants of bodies, (1) and (2)
and the medium, (3) respectively, and E(iv) are the values of E at imaginary
frequencies. While the first term in Equation (3.1.8) gives the sum of the
contributions of Keesom and Debye interactions, the second term represents
the first term of an infinite series that gives the London dispersion contribution.
The other terms can be neglected in most cases (Israelachvili, 1992b).
Considering the second term, we need functinal forms for E( iv) for all three
materials to compute the London contribution. Mohanty and Ninham (1976)
have presented a simple functional form which can be further simplified to
the case when London dispersion primarily results from electronic excitations
(Israelachvili, 1992b)
where n is the refractive index in the visible region, and Ve is the main
electronic absorption frequency in the UV. This form can be used for all
three materials provided they are non-conducting. Further, if we assume the
absorption frequencies of all three materials as the same at the same V e , the
second term (London contribution) can be obtained by integraton and the
Hamaker constant, A132 is approximately given as
Al32 ~ ~kBT (El

4 El
- E3) ( 3hv
+ E3 8V2
e)
(3.1.9)
The values of the various parameters are: h = 6.63 X 10- 34 J s; Ve = 2 to

5 X 10 15 s-l; kB = 1.38 X 10-23 JK- 1 •
With the knowledge ofthe static dielectric constants and refractive indices of
36 CHAPTER 3
all three materials, Am can be estimated by using Equation (3.1.9). For further
details and discussions one is referred to elsewhere (Israelachvili, 1992b).
3.1.3 BORN REPULSION
The Born repulsive potential is a short range potential resulting from the
overlap of electron clouds as the particles approach the point of contact. Its
value is very sensitive to structural details of the surfaces as well as the
intervening medium. In case of a solid-liquid-solid system as ours, the Born
repulsive potential is dependent on the surface roughness and the structure of
liquid lattice and is difficult to estimate. Feke et al. (1984) have presented a
formulation for Born-type short-range repulsion. Using the repulsive part of
the Lennard-Jones potential and adopting the Hamaker procedure of pairwise
additivity and integration, they have developed an equation for Born repulsive
potential for two identical spheres as a function of center-to-center distance.
We shall, however, adopt an expression developed by Ruckenstein and Prieve
using a similar approach for a sphere-plate system (Ruckenstein and Prieve,
1976a).
(3.1.10)
Here A 132 is the Hamaker constant and (J is the atomic collision diameter in
Lennard-Jones potential.
The typical value for (J is about 0.5 nm. Calculations show VBR can be
neglected compared to other potentials such as VDLR and VLVA for distance of
separation, h > 1 nm. Furthermore, the first term in the parentheses is usually
negligible because h / ap = H is much less than 1.
3.1.4 AB (ACID-BASE) INTERACTION
Acid-base (AB) interaction at solid-liquid interfaces is one among a number

of important non-DLVO (Derjaguin-Landau-Verwey-Overbeek) short-range
interactions (Christenson, 1988). This interaction may contribute substantially
to the adhesion energy between two phases (Fowkes, 1987; Vrbanac and Berg,
1991) as well as govern, in some cases, the stability of colloidal systems (van
Oss, 1994). An equation for the energy due to AB interaction is given as
(Chedda et aI., 1992)
(3.1.11)
Here VAB = hydrophobic interaction, in J; ho = minimum equilibrium distance,

in m; A =decay length of the liquid molecules, in m.
Note, the negative sign implies hydrophobic attraction, and the positive sign
implies hydrophilic repulsion. This equation is valid for a solid surface-liquid
system. When two similar solid surfaces with an intervening liquid system are
considered, the AB interaction energy is expected to be reduced as compared
to that for a surface-liquid system. For further discussion on AB interaction,
one is referred to elsewhere (van Oss, 1994).
VAB can be determined by a method suggested by Xu and Yoon (Xu and
Yoon, 1989). A value of2900 kBT has been proposed for highly hydrophobic
solids. The pore surface may not be hydrophobic in nature. Most fines of clayey
and other inorganic origin also may be more hydrophilic in nature. Organic and
some biological fines may, however, be hydrophobic. In the present system,
it is expected that VAs may be of small value of few tens to few hundreds of
kBT. Taking water as the liquid, reasonable values for ho and A are 0.5 nm
and 1.0 nm respectively.
3.1.5 HYDRODYNAMIC FORCES
The effects ofthe hydrodynamic forces on the release ofparticles will, however,
be dealt with in a rather detailed manner in the next chapter. In order to complete
the listing of possible contributions to the total interaction energy of the system,
an approximate method to calculate the potential due to hydrodynamic forces
is presented here.
Our objective is to obtain a simple expression for the potential due to the lift
force as a function of separation distance which can be added to double layer
and van der Waals potentials. In porous media, the flow is usually laminar.
Laminar flow over a fine particle (sphere) situated on a pore surface (plate)
does not generate any lift force (force perpendicular to the direction of flow)
(O'Neill, 1968). For a rotating sphere, however, there exists a lift force. Ifwe
assume that the lift force remains virtually unaltered over small distances, we
38 CHAPTER 3
obtain
(3.1.12)
Here R is the radius of the pore and p is the density of the liquid.
In case of turbulent flows, which are unlikely, we can use the result of
Cleaver and Yates and again assume that the force remains virtually unaltered
over small distances, to obtain (Cleaver and Yates, 1973)
2a v*)3
VHR = O.076pv 2 ( -----;- h. (3.1.13)
Here v* is the friction velocity and v is the kinematic viscosity.
3.1.6 TOTAL ENERGY OF INTERACTIONS
In the preceding sections, we have identified and obtained equations for the
various contributions to the energy of interactions between a pore surface and a
fine particle. We shall now obtain an equation for the total energy ofinteractions
VT by simply algebraically adding all the contributions. Negative and positive
signs represent the attractive and repulsive energy respectively. It is implicitly
assumed that the various different types of interactions are independent of one
another.
Before adding up the individual components of energy of interaction, we
shall rewrite them in the following manner to indicate the relevant parameters.
(3.1.14)
(3.1.15)
(3.1.16)
(3.1.17)
(3.1.18)
We observe from Equations (3.1.14) to (3.1.18) that all five components

of interaction energies are dependent on the distance of separation, hand
with the exception AB energy, the other four are dependent on the size of
the fine particle ap . The AB interaction energy is strongly dependent on the
hydrophobic/hydrophilic energy of interaction, VA's.
The magnitudes of both London-van der Waals and Born interaction energy
are primarily determined by the values of the Hamaker constant. The interaction
energy due to the electric double layer repulsion is strongly influenced by the
surface charge and potential and the salt concentration or the ionic strength of
the permeating liquid. The surface potentials are immeasurable quantities, and
therefore these are usually replaced by the measured zeta potentials. It should,
however, be mentioned that recently attempts have been made to determine
potential by charge regulation models and it is believed that this potential can
better represent surface potential than the zeta potential (Mohan, 1996a). The
hydrodynamic lift force is, as expected, dependent on the velocity and the
density of the liquid.
The total interaction energy of the system consisting of a fine particle
(sphere) and a pore surface (plate) is given as
(3.l.l9)
The important parameters such as particle size, salt concentration, Hamaker

constant, etc., are shown in parentheses. The VT can vary from negative values
(attractive) to positive values (repulsive) depending on the relative magnitude
of the various contributions.
Let us now look at the values of these contributions to VT at a particular set
of condition. The values of parameters used in the calculations are: h = I nm;
ap = 1 /Lm; C s = 0.1 M; 'If!OI = -30 mY; 'If!02 = -50 mY; AI32 = 6 x 10- 21 J;
VA's = -10kBT; v = 0.1 ms-I. Table 3.l.l presents these values.
Some useful observations can be made from this table. The Born repulsive
energy is relatively very small at h = 1 nm. It appears that the energy due
to hydrodynamic lift force can be neglected for small particles. It should,
however, be noted that the hydrodynamic shear can be important to the release
at high velocity. Cerda (1987;1988) has also shown that the hydrodynamic
potential can be neglected since it is only important for large particles at high
flow velocity. It needs to be retained however if we would like to obtain a
correct estimate of the primary minimum.
40 CHAPTER 3
TABLE 3.1.1. Magnitudes of various contributions

to VT at a particular set of condition.
Energy of interaction Magnitude (J)
V: R 5.2 x 10- 19
VOLR 7.9 X 10- 19
VLVA -9.9 X 10- 19
VBR 7.4 X 10- 23
VAB -2.5 X 10- 20
VHR 2.1 X 10- 23
A weak AB interaction is assumed, and hence the value of VA'B is about

two orders of magnitude smaller than that for a strong acid-base system. Weak
AB interaction is likely for a clay-water system. No measurements on VA'B for
systems pertinent to the migration of fines in natural porous media is available
in literature. With the assumption of a weak interaction, the value is about one to
two orders of magnitude less than the London-van der Waals and double-layer
contributions. The two major contributions to VT are; DLVO contributions -
London-van der Waals attraction and double layer repulsion.
Figure 3.1.3 shows the general qualitative variations of VT with h at three
greatly differing salt concentrations. While the plots A and C are at high and low
salt concentrations respectively, the plot B is at moderate salt concentration. We
found that, at very close separation, VT takes on large positive value as a result
of strong Born repUlsion which is independent of salt concentration. The Born
repulsion decays very rapidly as h goes beyond the collision diameter iJ", and at
these distances the slowly decaying London-van der Waals and exponentially
decaying AB repulsion/attraction begin to dominate. The extent of domination,
however, depends on the magnitude of a slowly varying double-layer repulsion
which is a strong function of the salt concentration. At high salt concentrations,
double-layer repulsion is weak, and hence VT remains attractive and decays
slowly after going through a minimum (plot A). This minimum is known as
primary minimum. At very low salt concentrations, the double layer repulsion
is strong and hence VT never becomes attractive. The presence of London-van
der Waals attraction and its relatively slow decrease as compared to double
layer gives rise to a maximum in VT (plot C). At moderate concentrations,
A High salt conc.

+ S Moderate salt cone.
C Low salt cone.
e
!
0
......
u
c
0
'-
2.-.....
cl-
.- en
-01- ~
...........
...... >
0
...... ---
III
III
>-
01
<li l-
e <lI
e
0 <li
III
C
<lI
E
0 Distance of
sepa ratio n, h ---.
Figure 3.1.3. Typical plots ofVT/kBT vs h.
both primary minimum and maximum occur. Often at these concentrations, a

secondary minimum may occur, resulting from the slowly varying London-van
der Waals attraction.
:rhe attachment of a fine particle to a pore surface can be modeled as if the
particle is located at the primary minimum as shown in Figure 3.1.3. At lower
salt concentrations the depth of the primary minimum is significantly decreased
and becomes small enough for the particle to be very loosely adhered. This
concept will be dealt with in the next section.
Numerous total interaction energy plots at various values of one or more
parameters (shown in Equation (3.1.19)), have been published in literature for
many systems. Keeping the theme of the chapter in focus , we shall show five
sets of (VT / k B T) vs h plots to demonstrate the colloidal effects in the release
of fines in Figures 3.1.4 to 3.1.8. Calculations have been done for koalinite-
42 CHAPTER 3
10
NoCI
8
6 -t3
A=2.6 x 10 erg
~ 8=5Ao
0
v 4
N
"
>- 2
...J
::!
~ 0
w
b
Cl. -2
...J
~
f:? -4
-6
-8
o 10 20 30 40 50
h, DISTANCE OF SEPARATION (Ao)
Figure 3.1.4. Total interaction energy plots for kaoline-pore surface system
in NaCl solution (Khilar and Fogler, 1987).
pore wall system in Berea sandstone. AB interaction is not considered in

these calculations and different equations have been used for the double-layer
repulsion. Neverless, the qualitative features remain intact and the elucidation
of the phenomenon can be achieved using these figures.
Figures 3.1.4 to 3.1.6 are constructed for kaolinite-pore wall system in Berea
sandstone for three different salt solutions of NaCl, CsCI and CaCh (Khilar
and Fogler, 1987). We found that while the nature of the plots remains virtually
unaltered, the magnitude is strongly dependent on the type of the salt solution.
Figure 3.1.7 is constructed for the same system at different pH of the solution
(Vaidya and Fogler, 1990). We observe from this figure that the interaction
energy, VT, becomes increasingly repulsive (positive) with an increase in pH
CsCI
6 -13
A =2.6 It 10 erg
8=5Ao
4
0.005M
~ 2
0
~
.... 0
>-
~
-I
ct
-2
t=
z
w -4
b
a..
-I -6
~
~
-8
-10
-12
0 10 20 30 40 50
h. DISTANCE OF SEPARATION (Ao)
in CsCI solution (Khilar and Fogler, 1987).
owing to the change in surface (zeta) potential. Finally, Figure 3.1.8 shows the
plots for a mixture ofNaCI and CaCh salt solutions at different values of the
total ionic strengths and at a fixed percentage of calcium ions (Khilar et aI.,
1990). We observe from this figure that the total ionic strength in a mixed salt
solution is similar to the salt concentration in single salt solution.
Even though these figures are for a specific system, it is believed that the
various trends and features shown are in qualitative agreement with those for
other fine-pore surface systems as well as systems consisting of organic and
biological fines. The values are, however, expected to be different.
44 CHAPTER 3
6
-13
A=2.6xI0 erg
4 8=5AO
~ O.OOOlM
-'" 0 ----O.OOlM---
0
V
N
"- 0.1M
->
~
-2
0
'"§ -4
II>
;£
-0 -6
~
-8
-10
-12
0 10 20 30 40 50
h, Distance of Separation Of}
in CaCh solution (Khilar and Fogler, 1987).
3.1.7 CONDITION FOR THE RELEASE OF FINES: THE CONCEPT OF

CRITICAL SALT CONCENTRATION, CSC.
Equipped with the knowledge of the nature of the total interaction energy
(VT/kBT vs. h) plot, we shall now discuss an approach to delineate the
colloidal conditions under which the release of fines may occur. Early
work concerning the colloidal conditions necessary for the release of fine
particles was inspired by the success of DLVO theory in describing the
stability for colloidal suspension. Based on the concept of critical flocculation
concentration, CFC, Khilar and Fogler (1984) proposed that the release offine
particles may begin at a specific salt concentration at which both total energy
and total force acting on the particle are zero,
400
300
1\.0
10.0
9.0
pit. 1.0
.n
-700 HCI"'.ker CClnskln'" 2.4 ,,10 ...,_
-300+--~~---.--~-~-,--~---.--~----.
O.DO O.O~ 0.10 0.15 0.10 O.2~ 0.30 6 D.l!. 0."0 0.45 0.50
Distance of separotion. (If 10 cm.)
at aifferent pH (Vaidya and Fogler, 1990).
f-o 30
~
E-
Calcium Percentage = 10
;>
-....
<';I
.~
t::
~
20
<t)
10
~
t::
0
.....
u
0
<';I
-
s-
.....
CI)
·10
t:;
'";;i
0
r- ·20
0,0 0.2 0.4 0,6 0.8 1.0
x 200A
Distance of Separalion. h
Fi{f!1re 3.1.8. Total interaction energy plots for different ionic strengths and
10 Yo calcium in solution (Khilar et aI., 1990).
46 CHAPTER 3
Mathematically, the approximate CSC can be stated as
Vr =0, (3.1.20)
and
_ dVr =0 (3.1.21 )
dh .
Figure 3.1.9 shows the variations of Vr as a function h at this specific

salt concentration. It is important to note that there can be a maximum of
Vr, which effectively acts as an energy barrier since the particle located at
the primary minimum has to climb over the energy hump to be separated
(released). Assuming the energy barrier AE to be small (AE < tok B T), it is
argued that the Brownian motion (for particle of size < IJ.lm) andlorthe flow-
induced motion on the particle, would assist in the cross-over of the energy
barrier. Below this specific salt concentration, the Vr would be repulsive, and
hence the conditions are favorable for release. This specific concentration was
referred to as approximate critical salt concentration, CSC below which the
particles adhering to the pore wall may be released (Khilar and Fogler, 1984;
Vaidya, 1991). However, it has been shown by Mohan (1996b) that the energy
barrier AE could be substantial (several times that of kBT), and as a result the
particle release may not take place. Based on their work on water sensitivity of
Stevens sandstone, Mohan (1996b) proposed a better technique to determine
the critical salt concentration. In this technique, the energy barrier AE is plotted
as a function of the salt concentration. The critical salt concentration, CSC is
determined as the concentration at which AE vanishes. At this condition, the
particles are neither adhered to the pore wall nor prevented from release by an
energy barrier.
Khilar and Fogler experimentally found the existence of such a critical
salt concentration (Khilar and Fogler, 1984). The salt concentration in the
solution permeating through a core of Berea sandstone at constant flow rate
was decreased in a stepwise manner. The permeability of the sandstone was
continuously recorded, and the effluent collected at each step-decrease in salt
concentration was analyzed for the presence of fine particles. Figure 3.1.10
shows the decrease in permeability they observed (a manifestation of release
of fines as confirmed by corresponding effluent analysis) from a step decrease
in NaCI solution.
We observe from this figure that the permeability drops only when the salt
I-
>
>-'"
0'1
c
~
OJ +.
OJ
c
0
...... 0
u
0
\...
OJ
......
C
0
......
0
I- CTIS(or CSC)
Distance of separation,h
Figure 3.1.9. Interaction potentials showing the approximate critical ionic
strength (Khilar et at, 1990).
BEREA
t" Dio., t"Length,
FLOW RATE' \oOcc/h.
1.o·,..<><O-O-C><>C>-O<Jo()oOoo()oOoO<><:>o()o()oOoC>Q----------,
0.8
0.6
k/kr
0.4[1~ Ie IE IE IE
a.. 1Q. I~ Ig: I~ I~
0.2
80 - 10
I0 -
10
I0 -
10
0 _
10 10
I 10 - I 0 -
o~lq~tq~t~~I~~lq~1
rt)Z!!2zI IOZ i","ZI"'"ZI"'"ZI
°0~~40~-8~0--12~O--16~O~2~O~O~2~40~~28~0~3~2~O~3~60~4~00
PORE VOLUME
Figure 3.1.10. Critical salt concentration in water sensitivity (Khilar and

Fogler, 1984).
48 CHAPTER 3
concentration is decreased below the CSC (4250 ppm 9:! 0.07 M ). In some
instances, there may be a small decrease in permeability of sandstone above the
CSC due to the electro-kinetic interactions of pore surfaces with permeating
liquid (Khilar, 1983).
Observations similar to CSC have been reported earlier, and these are
tabulated in Table 3.1.2. We observe from this table that the CSC of a single
salt system depends on :
1. the valency of the cation,
2. the specific characteristics of cation (Na+ vs Cs+),
3. the pH of the solution, and
4. the type of the porous medium (pore surface to which particles are
adhered).
Predictions of CSC based on the approach developed by (Khilar and
Fogler, 1984) and (Mohan, 1996b) give excellent qualitative and reasonable
quantitative agreements with measurements (Khilar et aI., 1990; Vaidya and
Fogler, 1990; Khilar and Fogler, 1984; Kia et aI., 1987a; Kia et aI., 1987b;
Mohan, 1996b). From the Table 3.1.2, we observe for Berea sandstone, the
CSC ofNaCl, CsCI and CaCh are 0.07 M, 0.006 M and less than 0.0001 M
respectively. Such large differences in CSC have been addressed by Kia et aI.
( 1987a) who showed that the differences in CSC can be satisfactorily explained
in terms of the changes in the zeta potentials resulting from changes in the
valence and specific characteristics of cations (Kia et at, 1987b). The zeta
potential of the kaolinite--CaCh system is (-10 mY) as compared (-30 mY)
for the kaolinite - NaCI system. Likewise, the zeta potential of the cesium
system is also significantly lower than that of the sodium system. As a result,
for calcium and cesium systems, lower concentrations are necessary for the
electric double layer repulsive force to be sufficiently strong to counter-balance
the adhesive forces.
It is interesting to note that the relationship between the CSC and the nature
of cation is similar to that between the critical floculation concentration CFC
and the nature of the counter-ion. The difference in CSCs for the same salt
system but for a different porous medium can also be explained in terms ofthe
differences in surface potentials of the pore surfaces as well as the presence of
other interacting processes which will be discussed in the next section.
The CSC is also found to be dependent on the temperature. Khilar and
Fogler have measured CSC of the NaCl salt system for Berea sandstone at
temperatures of 0, 30, and 60°C as 0.056, 0.07 and 0.09 M respectively
(Khilar and Fogler, 1984). Such weak effects of temperature can also be
TABLE 3.1.2. Critical salt concentration for single salt systems.
Authors Porous Media Salt CSC (M) (pH)
Quirk and Packed bed of soil NaCI 0.25 (5.2)

Schofield (1955) Clay content - 19 % KCI 0.067 (5.2)
Kaolin - 40 % MgCl2 0.001 (5.4)
Illite - 40 % CaCh 0.0003 (5.4)
Rowel, Payne Packed bed of soil

and Ahmed Clay content - 22 %
(1969) Kaolin - 10 - 15 % NaCl O.l (-)
Illite - 75 - 80 %
Montmori-
llonite - 5 - 10 %
Hardcastle and Bed of compacted soil

Mitchel (1976) Clay content - 15 %
Illite - 100 % NaCl 0.05
Kolakowski and Packed bed of glass NaN0 3 0.20 (11.5)

Matijevic (1978) beads Ca(N03h 0.0001 (11.5)
Chromium oxide Co(dipY)3(Cl04 h 0.00001 (1l.5)
Khilar and NaturuallY consolidated NaCI 0.070

Fogler (1984) sandstone LiCl 0.068
Clay content - 8 % KCl 0.044
Kaoline - 88 % N~Cl 0.013 (8.0)
Illite - 12 % CsCI 0.0006
Temperature = 303 K CaCh < 0.0001
Velocity = 19.2 cmlh MgCb < 0.0001
Kia, Fogler Same system as NaCl 0.03 (8.5
and Reed (1987) used by Khilar to to
and Fogler (1984) 0.004 9.5)
50 CHAPTER 3
satisfactorily explained by accounting for the variations of the double layer

repulsion resulting from the change in temperature. Khilar (1981) has further
observed that velocity as high as 1.57 x 10- 3 m S-1 in Berea sandstone has
negligible effect on the CSc.
3.1.8 CRITICAL TOTAL IONIC STRENGTH (CTIS) FOR MIXED SALT

SYSTEM
The concept of critical salt concentration has been extended to a mixed salt
system ofNaCI and CaCh solution (Khilar et al., 1990). Previous studies have
shown that the number of Ca2+ ions present in the solution is important in
preventing the release of fines. These studies have further concluded that the
amount ofCa2+ must be greater than a minimum (Quirk and Schofield, 1966;
Jones, 1964; Kia et al., 1987b). The surface coverage ofCa2+ ions on the fines
is more important as it reduces the zeta potential significantly. In a mixed salt
system ofNaCI and CaCh, it is understood that the surface coverage of Ca2+
ions, will depend on two parameters relating to the ionic conditions of the
solution. These two parameters are the total ionic strength and the percentage
of calcium present in the solution. In the case of the NaCI and CaCh system,
these two parameters are written as,
I = MNa + 3Mca,
and
CP = (MN::~~CJ x 100, (3.1.22)
where I is the total ionic strength, CP is the calcium molar percentage and
MNa and Mca are the concentration of sodium and calcium ions in moles
per litre respectively. The generalized expressions of these two parameters are
given below.
1
1=-
2
L n· z·0
~
2
~
(3.1.23)
and
iP =( Zi
I: n i
~iZt.) x 100, (3.1.24)
TABLE 3.1.3. CTIS for Berea

sandstone and soil with clay at different
calcium percentage.
CP cns (M) cns (M)

(sandstone) (soil)
0 0.071 0.05
5 0.025
10 0.005 0.017
15 « 0.001 0.011
20 0.009
where Zi is the valence of the ith ion and ni is its bulk concentration.
In analogy with the CSC, the critical total ionic strength is defined as the total
ionic strength of a mixed salt system below which the fines may be released.
Khilar et aI. (1990) have experimentally measured CTIS of the NaCl and CaCh
system at different percentages of calcium for Berea sandstone. Their results
are presented in Table 3.1.3.
We observe from this table that CTIS is strongly dependent on the percentage
of calcium, CPo We further observe that if the CP is greater than 15, the release
of fines in Berea sandstone can be prevented, which agrees with the previous
studies. The values of CTIS are higher for the soil system shown, and this
can be partly attributed to the presence of illite particles which carry a higher
electrical charge on their surface than that of kaolinite particles.
With relevant modifications accounting for more than one counter-ion, the
approach discussed earlier has been successfully used to describe the influence
of calcium percentage on CTIS (Khilar et aI., 1990). In the expression for
Debye--Huckel, parameter ni is replaced by I, and another suitable expression
for VDLR was used in the calculations.
3.1.9 SOME IMPORTANT EFFECTS ON THE ENERGY OF

INTERACTION
There are some physico-chemical processes that generally occur during the
flow of a permeating solution in a consolidated porous medium. Two such
processes can have significant influences on the total interaction energy, VT.
52 CHAPTER 3
These are the ion exchange between the permeating solution and the pore
surface, and the fine particles adhering to the pore surface and the dissolution
of materials such as calcite to produce polycations in the solution. We shall
discuss these two processes in some detail. In addition, we shall discuss one
recently studied effect that alters the CSC significantly. It is found that the
swelling clays present in sandstones can induce the release of migratory clay
particles resulting from a swelling-related jump in interlayer spacing, referred
as micro quake (Mohan and Fogler, 1997).
Ion exchange between the permeating solution and the solid surfaces
Solid surfaces present in consolidated porous media such as pore surface and
the surface of the fine particles are known to exchange ions with the solution.
A typical example of this kind of ion exchange process is the exchange of
Na+ ions for H+ ions from the fresh water in a water shock experiment on
Berea sandstone (Kia et aI., 1987a). We observe from Figure 3.1.11 that, the
pH remains unaltered at the value of9.5 during the salt water flow. However,
we further observe that the pH increases to a value of about 10.3 temporarily
after the flow of fresh water. This increase in pH is a result of the decrease
in H+ ions which have been exchanged for Na+ ions. Likewise, other cations
such as Ca2+ and AIH can also be exchanged t%r from the surface. Several
examples are also found in studies on flow in soil.
As a result of this ion exchange processes, the ionic strength of solution can
change in situ due to the exchange of ions of different valency. The electric
charge on the solid surface (pore and fine particles) can change, and the pH of
the permeating solution may change.
All these changes are likely to significantly alter the double layer interaction
energy, VDLR, as well as the acid-base interaction, VAB. Consequently the total
interaction energy, VT, changes and thereby changes the conditions (e.g. CSC)
for the release offine particles. From a knowledge of the chemical compositions
of the permeating solution and the solid surfaces, it is possible to identify the
ion exchange process and its effect on the energies of interactions.
Dissolution ofpolycation forming materials ofpore surfaces and fines

It is known that calcite in sandstones can dissolve in water or salt water
generating calcium ions (Schweich and Sardin, 1985). Polycations such as
AIH can also be leached away from fine clay particles such as kaolinite by
water (Swartzen-Allen and Matijevic, 1976).
The dissolution and/or leaching of polycations such as Ca2+, Mg2+,
AIH, and Fe H from the porous media alters the ionic conditions of the
..f'···••• \r
1.0.,----_-_-_-_-_-_-_-_"""W 10
1</1<1 I '--,, ______ _
0.8 I
SAL T FLOW I FRE SH WATER FLOW
0.8 ..;. ~
0.4 I
I
4
~
0.2 2
0.0.L..._""T"""_-r-_41_....,._--,~-L0
o 30 80 90 120 150
Pore-volume
Figure 3.1.11. Variation ofpenneability and effluent pH (Kia et aI., 1987a).
penneating solution. Furthennore, these polycations can either be adsorbed or

ion exchanged on to the surfaces of the pore and/or fine particle. Together these
two processes can alter significantly the VDLR and VAB components of Vr. In
most cases, VDLR decreases when polycations are present, thereby inhibiting
the release offines.
The effect of solution pH in the release of fines in Berea sandstone can also
be related to dissolution of calcite (Kia et aI., 1987a). It has been shown that
there is a critical pH value of2.6, below which fines are not released in Berea
sandstone (Leone and Scott, 1988; Kia et aI., 1987a). When the penneating
solution is below a pH of2.6, the zeta potential of clayey fine particles present
in Berea sandstone is very low due to the ion exchange with Ca2+ ions leached
from Berea at such low pH solution.
These considerations also apply to the release of fines in soil mass or other
porous media. The major task is, however, to identifY these processes that alter
the colloidal conditions ofthe penneating solution and to assess the magnitude
of changes in the various components of energy of interaction.
54 CHAPTER 3
TABLE 3.1.4. CSC of various salt solutions

in Stevens and Berea sandstones
Salt Stevens Berea

NaCl 0.50-0.25 M 0.07M
KCl 0.30-0.20M 0.03M
CaCh 0.3-0.20 M None
Presence of Swelling Clays

Clay minerals such as montmorillonite and smectite swell in water or dilute
aqueous salt solutions and are known as swelling clays. In a recent study Mohan
and Fogler (1997) have measured the CSC of Stevens sandstone containing
both swelling and non-swelling (migratory) clays. Table 3.1.4 presents the
CSC of three salt solutions for Stevens and Berea sandstones.
We observe from this Table that the CSC for Stevens is much higher than that
for Berea which does not contain any swelling clays. In a systematic study,
it is shown that swelling of smectite particles undergoes a transition from
crystalline regime to osmotic regime at 0.25 M NaCl. At this concentration
the interlayer separation increases by a discrete jump (>20 1\) and can be
visualized as a microquake (Mohan and Fogler, 1997). This swelling transition
and the consequent jump in interlayer spacing (micro quake) causes the particles
on the surface to be dislodged and released. Model calculations based on the
theory of critical salt concentration, described earlier, showed that the jump
occurs when the layers go from primary minimum to the secondary minimum
resulting from the lowering of salt concentration.
3.2 The Dynamics (Rate) of the Release Process
In the preceding section, we have discussed issues concerning why fines

adhering to pore surface are released due to the colloidal forces and under
what colloidal conditions of the permeating liquid, the release of fines may
occur. In this section, we shall discuss how rapidly the fines are released once
the conditions for release prevail during the flow.
A number of studies - both experimental as well as theoretical in nature -
have been conducted on the rate of colloidally induced release of fine particles
from model porous media. A major experimental study was undertaken by

Matijevic' and co-workers (Kuo and Matijevic, 1979; Kuo and Matijevic,
1980; Kallay et al., 1986; Kolakowski and Matijevic, 1979; Thomson et al.,
1984; Kallay and Matijevic, 1981). Using a packed-bed technique, they have
investigated the effect of various parameters in packed bed on the rates of
adhesion and removal of particles. The results will be discussed in appropriate
places in this section. Theoretical studies addressing the rate of release of fine
particles from surfaces have been presented by Dahneke (1975a; 1975b) and
by Ruckenstein and Prieve (1976a; 197 6b). These theories are based on the
Brownian diffusion of particles over an energy barrier between a fine particle
and a pore surface. Barouch et al. (1987) have addressed the unsteady nature of
the process by solving the full-time-dependent Kramer equation. However, an
analytical expression for the release rate coefficient, 0: has not been obtained.
These theories apply for the release of Brownian particles and, therefore, we
shall first discuss the kinetics of the release of Brownian fines (a p < 1 /-Lm)
and then consider the release of non-Brownian particles.
3.2.1 THEORETICAL MODELS FOR THE RELEASE OF BROWNIAN

FINES
Brownian diffusion over an energy barrier can be described by Fokker-Plank

equation, which is too complex to render a general solution. Under a quasi-
steady assumption, Dahneke has solved Fokker-Plank equation to determine
the kinetics of escape of particles from the surface (Dahneke, 197 5a).
Consider a flat surface to which particles are adhered or surface -borne (as
Dahneke has worded). Further consider that the particles are bound to the
surface by a potential energy barrier b..E = VT(h max ) - VT(hmin) where h max
and hmin are the distances at which VT is maximum and minimum respectively
(see Figure 3.1.3). With, n being the concentration of particles in the liquid
volume, the one-dimensional convective diffusion equation (Smoluchowski's
equation) can be written as
(3.2.1 )
where DB = Brownian diffusivity, in m2 s-l; kB = Boltzman's constant, in

J K- 1; T = absolute temperature, in K; h = perpendicular distance from the
surface, in m; v = fluid velocity near the surface, in m S-l .
56 CHAPTER 3
Note the friction factor has been replaced in tenns of DB using Einstein's
equation. The first tenn on the right hand side represents the flux due
to Brownian motion while the second tenn represents the convective flux
generated due to the interaction force. The left hand side is the material
derivative.
Dahneke (1 975b) made the following simplifications to obtain an analytical
solution of Equation (3.2.2). Fluid velocity, v near the surface, was assumed to
be small, and the energy barrier (VTmax - VTmin) was taken to be several times
(> lO)kBT. The latter assumption implies that the particle escape over a high
energy barrier would be small, and hence quasi-steady state assumption can
be made. He further assumed that VT(h) approaches -00 for large values of
h, even though it really goes to zero. Likewise, VT goes to +00 for h < hmin.
With these assumptions Dahneke solved Equation (3.2.1) to detennine the
diffusional escape of Brownian particles from a surface having concentration
of particles, N.1t is shown that the rate of release can be written as a first order
decay process as given below.
dN
-=-aN (3.2.2)
dt '
where a is the release coefficient.

According to Dahneke's solution, a is given as (Dahneke, 1975b)
a =
[h mox rX) exp [_ (VT(h max ) - VT(h)] dy dh, (3.2.3)

J~~ h kBT
where y is a dummy variable.

The assumptions of VT being infinitely positive at the minimum particle
separation from the surface and infinitely negative at large separation are not
realistic. Ruckenstein and Prieve (1976a;1976b), using a similar approach but
with a reasonable total interaction energy profile, also solved the quasi-steady
state Smoluchowski's convective diffusion equation. In their approach, they
define an interaction force boundary layer near the surface, outside of which
the particle-surface interactions can be neglected and inside which the fluid
motion can be neglected. Usually the thickness of this layer is smaller than
the diffusion boundary layer. Using a typical VT vs. h variation of having one
maximum at h max and two minima at hminl and hmin2,

Ruckenstein and Prieve
(1976b) showed that the net quasi-steady state diffusive flux of particles from
the surfaces can be represented by a reversible first order chemical reaction.
The expressions for the rate constants are given as (Ruckenstein and Prieve,
1976b)
(3.2.4)
and
(3.2.5)
In arriving at these expressions, Vr(h) has been expanded in a Taylor series

about h max and truncated after the second term. Furthermore, the Vrmax is
approximated by a sharp spike. The rmax and rmin2are the absolute values of
the second derivative ofVr at the maximum and primary minimum respectively
and are given as
rmax = - ( d;~J) Ih=hm~ ,
rmin2 = + ( d;~J) Ih=h min2 • (3.2.6)
The values for the release coefficient as given by Equations (3.2.3) to

(3.2.6) can be calculated from the knowledge of Vr and the dependency of
Brownain diffusivity with distance. A systematic study to test these theories
with measurements has not been reported yet. The difficulty of such a task
is well recognized when we consider the fact that Vr vs hand DB vs h
relationships are indeed difficult to obtain experimentally. Neverthless, these
theories give good qualitative agreements with measurements.
Ryan and Gschwend (1994) have shown that in the absence of an energy
barrier the release offine particles from the surface is most likely controlled by
the diffusion through the hydrodynamic boundary layer. This study is relevant
to the colloidally induced release of fines where the energy barrier may not be
58 CHAPTER 3
present. Using this approach, we have
(3.2.7)
where bb is the thickness of the boundary layer and DB is the Brownian

diffusivity.
Suitable equations for DB and bb can be used to calculate the release
coefficient, a. Equation (3.2.7) predicts that a is dependent on the flow
velocity of the permeating solution. Some qualitative comparisons between
the measurements and the predictions of the equations of a as presented can
be made. Matijevic and co-workers have reported the effect of a number of
parameters on the rates of detachment of particles (Kuo and Matijevic, 1979;
Kuo and Matijevic, 1980; Kolakowski and Matijevic, 1979; Thomson et al.,
1984; Kallay and Matijevic, 1981; Kallay et al., 1986). They have shown that
any change in one or more parameters of the permeating solution, such as pH
and ionic strength, cause an increase in the energy barrier i::lE, which in tum
decreases the rate of release. Scrutinizing Equations (3.2.3) and (3.2.6), we
observe that indeed an increase in i::lE will decrease the rate. Quantitatively
the release rate coefficient, a has been found to be between 10- 10 to 10- 2 8- 1 ,
which also turns out to be the typical range of a one obtains from these
equations. The influence of temperature likewise qualitatively agrees with
measurements, that is, a increases with the increase in temperature. The effect
of flow rate on a observed by Ryan and Gschward (1994), however, cannot be
predicted by the theories based on diffusion over an energy barrier.
3.2.2 SOME CONSIDERATIONS ON THE RELEASE OF

NON-BROWNIAN FINES
Non-Brownian fines are those particles for which the Brownian motion can be
neglected. These particles are generally of size greater than 1 j1m. Keeping
the same approach as for Brownian fines, one can write the unsteady particle
balance equation without the Brownian diffusion term
(3.2.8)
where f is the coefficient of friction and (8VT / 8h) is the force acting on the
particle.
In principle, one can solve this equation with the knowledge of VT as a

function ofh and tto obtain n(h, t). From the knowledge ofn(h, t), one can in
principle determine the amount of flux leaving the surface, which then yields
an estimate of the rate of release of non-Brownian fines.
Assuming the validity of this approach, we can further speculate the effects
of various colloidal parameters on the rate of release. Colloidal parameters
such as pH and ionic strength will affect the rate through their effect on VT.
3.2.3 AN EMPIRICAL EQUATION FOR THE RATE OF RELEASE OF

FINES
Based on the discussions of the theoretical and experimental studies on the rate
of release of fines from the surfaces presented above, it appears plausible that a
first order rate equation can be proposed to describe the kinetics of the release
process. As a matter of fact such an equation has already been proposed in
literature by Khilar and Fogler and by Matijevic and co-workers (Khilar, 1981;
Kuo and Matijevic, 1979; Kolakowski and Matijevic, 1979; Kuo and Matijevic,
1980). The proposed rate expression contains two parameters (a and 0"0) and
is given by
dO"
dt = -aO", with 0" = 0"0 at t = O. (3.2.9)
where 0" is the concentration of particles expressed in amount (mass or number)

per pore volume that remain adhered to the solid surface at any time, t.
The parameter a is the release coefficient, and it depends on a number of
physical and chemical factors. The dependency of a on various parameters
can be discerned from the theoretical discussions presented in Section 3.2.2.
Table 3.2.1 presents the major variations of a in a qualitative manner.
The other parameter, 0"0' represents the effective concentration of particles
initially present on the pore surface, taking into account that a fraction of
the initially present particles can be released due to colloidal forces. This
parameter, 0"0' depends primarily on the nature of attachment or adherence of
the fines to the pore surface. For example, in studies of particles adhesion and
removal in a packed bed, the history of adsorption process such as ageing time,
etc., may influence the parameter, 0"0' In case of sandstones and soil masses,
the nature of the nucleation and growth processes of the fines may influence,
0"0' In any case, 0"0 must be determined experimentally.
0\
o
TABLE 3.2.1. Variations of the release coefficient
Factors Nature of variation References
Increase in ionic strength Decreases Khilar and Fogler, 1983;

Kolakowski and Matijevic, 1979
Increases in valence of counter ions Decreases Khilar and Fogler, 1984;

Kolakowski and Matijevic, 1979
Increase in pH Increases Kolakowski and Matijevic, 1979;

Kuo and Matijevic 1979; Kia et at., 1987a
Increase in temperature Increases Khilar and Fogler, 1983
Increase in flow velocity Increases Khilar, 1981; Ryan and Gschward, 1994 w
~
This two-parameter empirical rate expression describes the complex process

of colloidally induced release of fines in porous media reasonably well.
CHAPTER 4
HYDRODYNAMICALLY INDUCED RELEASE OF

FINES IN POROUS MEDIA
The fine particles adhering to the pore surfaces of the porous media of
relatively higher porosity and permeability can be released or detached by
the hydrodynamic forces. Hydrodynamically induced release offines is found
to be a more common occurrence in loose soil embankments and in packed
beds than that in oil sandstone formations. There are two major reasons for
this: (a) the flow velocity encountered is higher and (b) the fines are usually of
large size and therefore experience a large drag force. We shall analyse these
aspects later in this chapter.
In this chapter, we shall discuss the hydrodynamic detachment/release of fine
particles in an analogous manner to that followed in Chapter 3 on colloidally
induced release of fines. First, we shall address hydrodynamic forces and
mechanisms (Statics) that cause the detachment, and then we shall discuss the
rates of release of fines (Dynamics).
4.1 The Statics ofthe Release Process
The statics of the release or detachment of fine particles induced by

hydrodynamic forces is more complex than that induced by colloidal forces
because hydrodynamic forces may act in more than one direction. Consequent
to this action, the particle may undergo different kinds of particle motion
such as rolling, sliding, and lift. In addition, the effects arising out of the
contact deformation of the particle and the surface may influence the release
mechanism.
Unfortunately, this particular topic has not drawn adequate attention and, as a
result, a comprehensive understanding is yet to be achieved. We first begin with
developing a basic understanding using an ideal system of a sphere adhering to
a flat surface. We shall limit our discussion to laminar flow regime as turbulence
seldom occurs in seepage flow. We shall then discuss the important concepts
63
64 CHAPTER 4
Figure 4.1.1. Shear flow over a particle attached to a surface.
of the critical hydrodynamic stress or the critical flow velocity.
4.1.1 MECHANISM OF HYDRODYNAMIC DETACHMENT
We consider a simple yet fairly representative system of a spherical particle

adhered to a flat surface by means of London-van der Waals force of attraction.
A shear plane of flow past the particle is shown in Figure 4.1.1.
We consider the velocity profile to be linear in a cylindrical pore as a
reasonable first approximation. The hydrodynamic forces acting on the particle
due to the flow has been determined by O'Neill (1968). His analysis showed
that there is no lift force acting on the particle (Fz = 0) and that a tangential
force acts on the particle in the direction of flow. It is given per unit area as
(4.1.1)
where J-l is the viscosity ofliquid, a p is the particle radius and Vy is the velocity
in the y-direction at a distance ap from the wall (z = ap).
Furthermore, a torque acts on the particle about the x-axis, and it is equal to
Tx = 7.55 7r J-l a~ v y. (4.1.2)
In case of turbulent flows, the mechanism of detachment of colloidal particles

from a flat substrate has been analysed by Cleaver and Yates (1973). They
have shown that a lift force may exist due to the unsteady nature of the viscous
HYDRODYNAMICALLY INDUCED RELEASE OF FINES 65
sub layer in the turbulent boundary layer, and it is given as
(4.1.3)
In this expression, 1/ is the kinematic viscosity and v* is the friction velocity.

We shall, however, consider the more likely situation of laminar flow and
consider the process of particle detachment. Particle motion can be generated
for possible detachment by the following three possible mechanisms:
1. a force in z-direction (lift force) causes the particle to detach;

2. the hydrodynamic force in y-direction slides the particle along the surface;
3. The torque about the x-axis causes the particle to roll along the surface.
Out of the three mechanisms, the first one is ruled out as there is no lift
force acting on the particle. As regards to the other two, Equations (4.4.1) and
(4.4.2) indicate that both mechanisms are probable. Viser (1976) postulated
that the tangential force, FH required to release the particle is proportional to
the adhesive force, FA with the result that
(4.1.4)
where 'Y is a proportionality constant and its value will depend on whether
the particle release mechanism is sliding or rolling. If the release mechanism
is sliding, then 'Y is the coefficient of static friction whereas for the rolling
mechanism'Y is evaluated through a torque balance (Hubbe, 1984). Sharma
and co-workers have carried out a systematic study to determine the release
mechanism of glass and polyethylene particles from the glass and mica surfaces
using a flow cell consisting two parallel microscope glass slides with a gap
of about 200±10 /Lm (Sharma et aI., 1992; Das et aI., 1994). Both glass and
polyethylene particles ranging from 10 to 40 /Lm in size were used. Particles
were made to detach under two situations: a drag force acting on the particle
in a flow cell and a centrifugal force acting on the particle when the cell is
kept in a centrifuge. If sliding is the mechanism, then the force required to
detach the particles in both situations would be the same. If, however, torque
required is the same, then rolling is the particle release mechanism. Their
experimental studies showed that the hydrodynamic torque required for the
release of particles in the flow experiment and the torque due to centrifugal
force required in the centrifuge experiment are approximately the same, thereby
66 CHAPTER 4
Figure 4.1.2. A schematic diagram of contact deformation of an elastic sphere

on a rigid surface.
indicating that the rolling is the release or detachment mechanism in their

experimental studies (Sharma et aI., 1992).
The fact that a critical amount of torque or force is required to detach the
particle indicates that a restraining torque or force is present in the detachment
process. Such is feasible either when the adhesive force deforms the particle or
the surface or both at the point of contact or when the particle begins to roll over
a rough surface. Figure 4.1.2 shows a schematic of this contact deformation
phenomenon for an ideal case of an elastic sphere located on a smooth rigid
surface.
Here, a is radius of the contact circle and the angle f3d is a measure of
deformation as shown in Figure 4.1.2. Derjaguin et ai. (1975) have derived
an expression for the deforming forces in terms of elastic and geometric
parameters.
A _ 4a~ E 3/ 2
F N - 3(1 _ (j~) f3d . (4.1.5)
Here F/J
is a deforming force and is equal to London-van der Waals force in
our case. E and (jp are the Young's modulus and Poisson ratio, respectively.
Based on this concept of contact deformation and the presence of roughness
on the surface, Das et ai. (1994) predicted that a critical hydrodynamic
force exists for the detachment of particles from the surface. This critical
hydrodynamic force / stress or equivalently critical flow velocity is an
analogous concept to critical salt concentration in colloidally induced fines
release and is discussed in detail in the next section.
4.1.2 CRITICAL HYDRODYNAMIC STRESS AND CRITICAL FLOW

VELOCITY
In the preceding section, theoretical considerations using an ideal system led

us to the conclusion that a critical shear stress or a critical flow velocity is
required to detach particles adhering to a surface.
In case of the release of particles in turbulent flow, Cleaver and Yates
(1973) have shown that there exists a minimum wall shear stress for release
and that the magnitude is dependent on the particle size. In cases of flow
through soil masses and sandstone formations, the existence of a critical shear
stress or a critical velocity has been found experimentally (Arulanandan et
aI., 1975; Gruesbeck and Collins, 1982). Arulanandan et ai. (1975) used a
rotating cylinder to measure the erosion rate of soil samples at different stresses.
The inner cylinder made of soil samples was held stationary while the outer
cylinder was rotated with eroding liquid placed in the annular space. Figure
4.1.3 presents erosion rate as a function of shear stress for two eroding liquids
of different salinity.
We observe from this figure that there exists a critical shear stress above
which the erosion begins. The magnitude of this critical shear stress depends
on a number of parameters, one of which is salinity.
Gruesbeck and Collins (1982) conducted core-flood experiments to study
the release offine particles from Berea sandstone due to hydrodynamic forces.
Figure 4.1.4 shows the permeability reduction as a function of interstitial
velocity through the core. Noting that permeability reduction is a manifestation
of release and migration offines, we likewise observe that there exists a critical
flow velocity above which the fine particles are released.
It should be mentioned here that the fine particles released consist of clay,
quartz and feldspar (Gruesbeck and Collins, 1982). Furthermore, the flow
velocities are higher. The influences of various parameters relating to the pore
surfaces, fine particles, and permeating liquid on the magnitude of critical
shear stress have not been systematically studied. Nevertheless, the studies of
Arulanandan et ai. (1975) have shown that the critical shear stress, Tc for the
erosion of soil mass depends on the type of clayey fines, the ionic strength
and composition of the permeating liquid and the pH and temperature of the
68 CHAPTER 4
Critical Shear
Stress AI
SAR"13.0
SAR (Pore Water)
Figure 4.1.3. Relationship between erosion rate and shear stress for different
concentration of eroding fluid - Yolo Loam, SAR "",35, (J' = 2.0 mmhos/cm,
pore Fluids = 0.02 N (Arulanandan et al., 1975) (notations and units as used
by Arulanandan et al., 1975).
30
25
z
'">=
g 20
fa
a:
5a; 15
~
! 10
~ ISOPAR M
;i<
AT CONNATE
2% KCI BRINE
SATURATION
O~~ __ ~ __ ~ __ ~ __ ~==~~ __ ~ __ ~~
o .02 .04 .06 .08 .10

.12 .14 .16 .18 .21
INTERSTITIAL VELOCITY. "I</Ii' CM/S
Figure 4.1.4. Permeability reduction vs interstitial velocity in Berea core:

diameter 3.81 cm, length 3.0 cm (Gruesbeck and Collins, 1982).
permeating liquid. In fact, these parameters are the same set of parameters that
influence the critical salt concentration, indicating that both colloidally induced
release and hydrodynamically induced release are phenomenologically similar.
The qualitative effects of clay type, the pH, and ionic strength of permeating
liquids on the critical shear stress are similar to those concerning the critical
salt concentration.
Further, if hydrodynamic lift were to be the main force in the release of
particles then exactly similar analysis can be conducted to describe critical
flow velocity as that has been done to describe the critical salt concentration. It
is believed that if the particles are irregular in shape and are at a small distance
from the pore wall, then the lift force at higher flow rate can be substantial and
can cause the release.
The critical shear stress, Tc has been used to classify soils based on their
erodibility characteristics for their use in constructing earthen embankments
such as dams (Arulanandan and Perry, 1983). It has been shown that if Tc for
a soil is <0.40 Pa, then soil can be considered as highly erodible. The typical
values of critical shear stress, Tc for soil masses range between 0.10 to 2.0 Pa.
Using model systems, Sharma and co-workers have also identified the
parameters affecting the release, particle size, particle material, pH and ionic
strength of the permeating liquid and the roughness of the adhering surface
(Sharma et aI., 1992; Das etaI., 1994). An increase in particle size and in rigidity
of particle material will decrease the critical hydrodynamic force while a rough
surface tends to increase it.
Hydrodynamically induced release of fines in sandstone formations has
not yet been studied systematically. The existence of a critical flow velocity
beyond which the fine particles are released has nevertheless been documented
(Gruesbeck and Collins, 1982; Muecke, 1979; Leone and Scott, 1988).
The typical critical flow velocities range from 0.01 to 0.1 m.s- 1. The
hydrodynamically induced release of fines in soils has been studied in the
context of colloid-facilitated transport of contaminants (Kaplan et aI., 1993).
This study has shown that flow rate influences the type, the size, and the rate
of release of fines.
4.2 The Rate (Dynamics) ofthe Release Process
In Section 3.2, theories for the rate of release of Brownian particles due to
diffusion over an energy barrier have been presented. In principle, a similar
approach can be applied to describe the release rate of Brownian particles
due to hydrodynamic forces. In most practical cases, however, the size of
70 CHAPTER 4
particles released due to hydrodynamic forces is larger than the Brownian size.
A theoretical study describing rate of release of particles from an adhering
surface due to hydrodynamic forces is that of Cleaver and Yates (1973). This
study, however, is applicable to turbulent flows, and the rate is determined by
the size and frequency of the turbulent bursts that occur in viscous sublayer.
For non-turbulent flows, empirical rate expressions have been developed
based on experimental measurements. Using soil erosion measurements in
a rotating cylinder type set-up, Arulanandan et ai. (1975) have shown that,
beyond a critical stress, the rate of erosion is proportional to the difference
between wall shear stress and the critical shear stress (see Figure 4.1.3).
Accordingly the rate of erosion r r can be expressed by
(4.2.l )
Here T w is the wall shear stress in Pa, Cth is the erosion rate coefficient in
kgs- 1 N- 1 and rr is the erosion rate in kgs- 1 m- 2 •
Like T e , the erosion rate coefficient, Cth also depends on the colloid-chemical
nature of the permeating fluid as well as the mineralogy of the soil. In fact, the
nature of dependency is so similar that Cth can be related to Te. Such an attempt
has been made by Arulanandan et al. (1980), and an empirical polynomial
relationship has been developed for natural cohesive soils for critical stress in
the range 0 < Te < 1 Pa. (Arulanandan et aI., 1980). This relationship is given
by
3 X 10- 8 - 17.1 x 1O- 8Te + 40.2 x 1O- 8T;

- 43.0 x 1O-8T~ + 17.5 x 1O- 8T:. (4.2.2)
Here the rate of erosion falls off rapidly in the range of (0 < Te < 0.5 Pa) and
then approaches a fairly constant value. This finding is significant because it
implies not only that soils with Te < 0.50 Pa are prone to erosion but also that
their erosion rates will be high. Conversely, for soils with Te > 1.0 Pa, the
release/erosion rates will be relatively low even after the critical shear stress
is exceeded. The values of Cth have also been determined from other types of
tests with natural cohesive soils. Table 4.2.1 presents the range of calculated
values of Cth for different flow erosion tests. We observe from this table that the
erosion rate coefficients, Cth vary to a great extent (by 5 orders of magnitudes)
from one type of flow test to another. We further observe that, Cth is relatively
lower for flows which are space constrained by macroscopic conditions and
TABLE 4.2.1. Comparison of calculated values of erosion rate coefficient, ah

from various types of flow tests (Gray et aI., 1985)
Types of flow test Range in calculated values References

of Gh (kgs- 1 N- 1)
Flume test 3-80 x 10- 8 Arulanandan et ai., 1980

Rotating cylinder 2-30 x 10- 9 Arulanandan et ai., 1975
Flow through slot 1-100 x 10- 10 Sanchez et ai., 1983
Flow through pores 1-2 x 10- 13 Khilar et ai., 1985
are higher for relatively unconstrained flows. Explanations based on the build
up of soil particles in permeating fluid and subsequent redeposition are not
completely satisfactory.
Gruesbeck and Collins (1982), based on their measurements using a glass
bead-sandpack, system proposed an expression for the rate of hydrodynamic
entrainment. The expression is given as
(4.2.3)
where (Y is the concentration of particles on the surface, in kgm- 3 ; v and Vc

are the flow velocity and critical flow velocity respectively, in m s-l and G~ is
the entrainment rate coefficient, in m -I, and r r is in kg m - 2 S-I .
CHAPTER 5
ENTRAPMENT OR PIPING OF FINES DURING

MIGRATION
In the previous chapters we have discussed the release offine particles from the
pore surface to the liquid stream flowing through the porous medium. These
fine particles while flowing with the liquid phase, can either readhere to the
pore surface or flow without capture or get entrapped at the pore constrictions.
The later occurrence leads to the plugging of the pore constrictions. In the
migration of fines, the later two occurrences are more common as the colloidal
and hydrodynamic conditions that bring about their release are not likely to
allow these particles to readhere back to the pore surface at the same conditions.
Therefore, during the migration of fines in porous media, the released fine
particles either get entrapped (entrapment of fines) or migrate without getting
captured (piping or washout of fines). Whether entrapment or piping occurs
will depend on parameters relating to the characteristics of the fine particles,
the permeating liquid (suspending medium) and the porous medium. This
chapter describes the factors affecting the entrapment or piping of fines. A
mathematical model to predict whether entrapment or piping occurs or not is
also discussed. Finally the model is applied to some important processes where
migration offines occurs.
5.1 Analysis of Factors Affecting Entrapment or Piping of Fines
Whether entrapment occurs or whether piping occurs is a function of several

parameters such as the pore structure, the size and the concentration of fines,
and the hydrodynamic and the colloidal conditions of the permeating liquid.
The occurrence of entrapment or piping is discussed in the light of these
parameters.
73
74 CHAPTERS
V\
bond
\
node bond node
Figure 5.1.1. Two dimensional networks of coordination number 4 and 6.
5.1.1 PORE STRUCTURE
Two important pore structure parameters that significantly affect the

entrapment or piping of fines are the pore constriction size distribution and
pore constriction to chamber interconnectivity. This interconnectivity can be
described in terms of coordination numbers; the number of pore constrictions
(bond) per pore chamber (node). Figure 5.1.1 shows two two-dimensional
networks with coordination numbers 4 and 6.
For porous media having small pore constriction size distributions and
low coordination numbers, the probability of entrapment of fine particles is
relatively high. Plugging resulting from entrapment is more likely to occur in
naturally consolidated porous media such as sandstones and rocks probably
because the constriction size is usually small (1 Mm) and the coordination
number is low (4 to 8). On the other hand, the relatively unconsolidated porous
media such as soil mass and packed bed are more susceptible to piping because
of their relatively large constriction size.
5.1.2 SIZE OF FINES
The size of fine particles or more pertinently, the ratio of size of fines to that of
pore constriction is a crucial parameter in determining whether entrapment or
piping would occur. If the fine size is equal to or greater than the size of the pore
constriction, then entrapment or plugging will inevitably take place. However,
ENTRAPMENT OR PIPING OF FINES 75
when the size of fines is much smaller than the constriction size, piping will
by and large occur. In reality, however, both fines and pore constrictions will
have respective size distributions. In such cases, Rege and Fogler (1987) have
shown that, the higher the overlap between the size distribution of fines and
that of pore constrictions, the higher is the extent of plugging and vice versa.
Table 5.1.1 presents some qualitative results on whether plugging, piping, or
other possible types of deposition would possibly occur as determined by the
ratio of the size of fines to size of the pore constrictions (Herzig et al., 1970;
Gruesbeck and Collins, 1982; Muecke, 1979). Single particle plugging would
occur when the fine size is comparable to that of the constriction size. Likewise
when the fine size is very small, piping would take place. When the ratio of
fine size to constriction size is, however, between 0.1 to 0.01, either piping or
multiparticle blocking can occur. Which of these occurs depends on conditions
such as the concentration of fines and flow rate. Retention of particles due to
surface deposition at sites other than the constriction sites also takes place.
5.1.3 CONCENTRATION OF FINES
When the ratio of size of fines to size of pore constriction is in the range
of 0.01 to 0.1 the concentration of fines will be a crucial parameter in
determining whether entrapment or piping occurs. To date, no studies have
been reported delineating the role of concentration on entrapment or piping.
There is, however, circumstantial evidence available from studies on deep bed
filtration, ultrafiltration, and flow of suspension in porous media that indicate
a role of the concentration of particulate phase.
Figures 5.1.2 and 5.1.3 show the decline in flux in ultrafiltration across a
membrane as a function of concentration of solid particles. We can observe
from the Figure 5.1.2 that beyond a solid concentration of 4 wt.%, the flux
declines, resulting from a continuous plugging of pores of the membrane. A
similar observation can be made as well from the Figure 5.1.3. We observe from
this figure that the decline in flux as a function of concentration of particulate
phase is dependent on the nature of the particulate phase.
In a recent study (Pandya et aI., 1998) using glass bead packs and aqueous
suspension of polystyrene latex particles, a critical particle concentration,
CPC was found to exist beyond which the bead pack gets plugged due
to mUltiparticle blocking. It is further shown that this critical particle
concentration, CPC, is strongly dependent on the ratio of bead size to particle
size. The CPC was found to increase from 0.35% (v/v) to 9% (v/v) when the
ratio bead size to particle size was increased from 12 to 40. (Bhuniya, 1996;
--.l
0\
TABLE 5.1.1. Dependence of plugging or piping on the ratio of size offines to size of pore constrictions
Size offines
Occurrence
Size of pore constrictions
2:1 Plugging due to blocking or size exclusion

(")
0.1 to 0.6 Plugging due to bridging and muitiparticle blocking
~
0.04 to 0.10 Plugging due to surface deposition, bridging, and muitiparticle blocking t:;l
i"
v.
0.01 to 0.04 Surface deposition. Multiparticle blocking mayor may not occur
less than 0.01 Piping

"
180
'r
"
160
140 P IN - 3.4 ot
"" ~
POUT - 0.8 0
flOW = 8.6 m H
120 TEMP ~ 35'C
I
'~
~
'"e 100
.......
..... ,~
tl..
I
)(
80
.....
:::;)
.... "
r".\
60
40
,~
,
20
I
0
C ? 3 4 6 8 10 15
CONCENTRA n ON % SOLI DS
20 .'
"
Figure 5.1.2. Typical flux-concentration relation for SBR latex suspension

(Zahka and Mir, 1977).
prllMf
..,t.,'." Ia I
"1. solids
Figure 5.1.3. Flux vs concentration plot (McCabe et aI., 1983).

78 CHAPTER 5
Pandya et al., 1998)

In deep bed filtration the concentration of the solid phase is usually small
(less than 1 wt.%) and, as a result, plugging does not occur. The particles are,
however, retained at cavern, crevice and surface sites; otherwise, the filter will
be ineffective. At higher concentrations, the filter beds have been found to
plug.
5.1.4 HYDRODYNAMIC CONDITIONS
Hydrodynamic forces acting on the released fine particles affect the plugging
phenomenon at the pore constrictions. If the flow velocity is small and the
particle density and size are large, then sedimentation force will affect the
trajectories of the particles, causing them to settle to the bottom wall of the
pore space. In most cases of migration of fines in porous media, the size of
fines is small and, therefore, the fines tend to follow the streamlines of the
flow. For very small size particles (less than 0.1 fIm), Brownian motion may
dominate the particle trajectory.
Small particles (say particle size an order of magnitude smaller than that
of constriction size) may form bridges at pore constrictions (Muecke, 1979)
similar to formation of dendrites in deep bed filtration (Payatakes et al., 1981).
The structures of these bridges as well as the rate of formation depend on the
hydrodynamic conditions of the flow. At higher flow rates, higher drag force
acting on the bridge will tend to break them apart while at lower flow rates,
the number of bridges formed may be greater.
The flow velocity can also indirectly affect the plugging of pores. The
rate of particle release is dependent on the flow velocity, and therefore the
concentration of fine particles in suspension will depend on the flow velocity.
The concentration of the fines in suspension, in turn, will affect the plugging
phenomenon.
5.1.5 COLLOIDAL CONDITIONS
Migrating fines in porous media carry an electrical charge on their surface.

Therefore, the colloidal conditions of the permeating liquids, i.e., the ionic
strength and the type of salt molecules, can cause flocculation or dispersion of
the fines in suspension. In either case, the plugging of the pore constrictions is
affected.
In hydrodynamically induced release of fines, the colloidal conditions of the
permeating liquid can be offloculating type, and the fine particles can form flocs
that increase their effective particle size. As a result, the probability of pore
plugging increases. Likewise, under a peptizing environment the dispersion
of flocs into individual fine particles will occur resulting in either, piping or
plugging.
In the colloidally induced release of fines, the colloidal environment of
the permeating liquid is usually of a peptizing/non-floculating type. Therefore,
floculation of the released particles does not occur and the fine particles remain
dispersed. In this case, an interesting effect relating to the rate of change in salt
concentration has been observed (Khilar et al., 1983). It has been found that
there exists a critical rate of salinity decrease, CRSD above which permeability
decline takes place indicating the occurrence of plugging and below which the
permeability decline does not take place indicating that either there is no release
or that piping occurs. It has been shown both experimentally and theoretically
that the rate of salinity decrease affects the pH of the permeating liquid resulting
from the ion-exchange between adsorbed Na+ on the fines surface and the
dissociated H+ ions in water (Vaidya, 1991). It is further shown that a slow
decrease in salinity causes a small increase in pH and, therefore, the zeta
potential of the fines remains small. As a result, the rate of release of the fines
becomes slow during the slow decrease in salinity experiments. The rate of
release of fines in turn determines the concentration of fines in suspension and
thereby indirectly influences the plugging at the pore constrictions.
5.2 Mathematical Models for Entrapment or Piping of Fines
From the above discussion, we note that there can be three regimes related to
the occurrence of piping or entrapment. In regime 1, the entrapment occurs
because the size of fines is comparable to that of pore constrictions. Regime 1
also includes small particles that form floes which then become trapped. In
regime 2, the piping takes place because the size offines is very small compared
to that of pore constrictions. Furthermore, if a peptizing environment exists,
the larger size floes may break down into smaller particles. In regime 3, either
entrapment or piping occurs. In this regime, the size of fines is small but not
small enough for piping to occur. Whether or not fine particles will plug the
pore constrictions will depend on the fluid velocity and the concentration of the
fine particles in the suspension. Therefore, the aim of the modeling studies is to
predict the concentration of fines in suspension as a function of time and space
in the porous media. A model discussed next has been develped to describe
soil erosion in earthen embankments. (Khilar et a!., 1985).
We consider a porous medium of length L in direction of flow and of
80 CHAPTER 5
Figure 5.2.1. A schematic diagram of a segmented cylindrical porous medium.
unit cross-sectional area as shown in Figure 5.2.1. We divide this into N

number of unit cells each having length (l = LIN). Each unit cell can be
thought of as a differential porous medium having porosity <Pi and permeability
Ki. Both Ki and <Pi are likely to vary during the process of entrapment
or piping. During piping, the porosity will increase due to loss of fine
particles, and consequently permeability will increase. On the other hand
during entrapment, the permeability drastically decreases due to the plugging
of the pore constrictions.
We further assume one-dimensional flow and develop a macroscopic
continuum model. The flow rate through the porous medium is governed by
the Darcy's law. Following are nomenclatures used in the model equations.
q =flow rate per unit cross-sectional area, m3 s-I m- 2 ;
C i - I = concentration of fines in the effluent fluid of (i - l)th unit cell, in
kgm- 3 ;
C i = concentration of fines in the effluent fluid of ith unit cell, kg m - 3 ;
rri =rate of release of fines in i th unit cell per volume of fluid, in kgm- 3 s-l;
rei = rate of capture of fines in i th unit cell at the pore constrictions,
kgm- 3 S-I;
<Pi =porosity of the ith unit cell;
Ki =permeability of the ith unit cell, in m2 ;
l = length of the unit cell, in m;
O"li = concentration offine particles present in the ith unit cell at any time t,
kgm- 3 ;
0"2i = concentration of fine particles captured at the pore constrictions in the
th
i unit cell, in kg m -3;
O:ci = the colloidal release rate coefficient in the ith unit cell, s -I;
O:hi =the hydrodynamic release rate constant in ith unit cell, kg N- 1 s-I;
(3i= the capture coefficient in the ith unit cell, in s -1 ;
Ti = shear stress at the wall inith unit cell, in Pa;
Tc = critical shear stress, in Pa;
Asi = wall surface area in the ith unit cell, m2 m- 3 ;

Pb =bulk density of the fine particles, in kgm- 3 ;
For the sake of simplicity, the volume used in expressing the concentration
of fines and the rates of release and capture etc., is the pore volume. Fine
particles are assumed to be uniformly suspended.
We write an unsteady mass balance for the fine particles in the suspension
in the ith unit cell, having unit cross-sectional area.
qCi - 1 qCi + rrJ¢i r cil¢i

(flow rate of particles) (flow rate of particles) (rate of release) (rate of capture)
m out
(5.2.1)
(rate of accumulation)
Equation (5.2.1) can be written for each unit cell and, therefore, we rewrite
noting I to be constant.
for i = 1 ... N, (5.2.2)
where N is the total number of segments. Furthermore, the initial conditions

are, that initially there are no particles in any unit cell and that the fluid entering
the first unit cell does not contain particles.
Ci =0 at t =0 (5.2.3)
and
Co = 0 t > O. (5.2.4)
For colloidally induced release, the rate expression can be written as

82 CHAPTERS
discussed in Chapter 3, Equation (3.2.9).
dO"Ii
rri = -Tt = O!ciO"Ii, (5.2.5)
with O"Ii = 0"10 at t = O. (5.2.6)
F or hydrodynamically induced release, the rate expression can be written as

discussed in Chapter 4.
dO"Ii
rri = -Tt = O!hi(Ti - Tc)Asi, (5.2.7)
and
O"Ii = 0"10 at t = 0, (5.2.8)
where Asi is the surface area per unit pore volume.

The rate of capture will be assumed to be dependent on the concentration of
fines (Groesbeck and Collins, 1982; Muecke, 1979; Khilar and Fogler, 1987).
We can write the rate expression of the capture of fine particles as
d0"2i
rci = - - = PiCi (5.2.9)
dt
and
0"2i = 0 at t = o. (5.2.10)
The capture coefficient P is expected to be dependent on the flow rate, the

concentration of fines. and other variables.
Equations (5.2.2) to (5.2.10) can be solved simultaneously to obtain C i as
a function of time provided both the flow rate per unit cross-sectional area, q,
and the porosity, rpi, are known as a functions of either time or 0" c. The velocity
or the flow rate per unit cross-sectional area, q, can be obtained from Darcy's
law as given below:
(5.2.11)
We note here that there are N number of unit cells oflength l in the direction
of flow. Furthermore, noting that the permeability of unit cells may be different
due to the differences in porosity and extent of plugging in different unit cells,
we write the overall permeability, K, as
N 1 N
K=L Ki ' (5.2.12)
~
Here Ki is the permeability of the ith unit cell.

The permeability of the ith unit cell, Ki is expected to be dependent on the
porosity, ¢i, or on the extent of capture, (}"2i, depending on whether the change
in permeability is due to piping or due to entrapment at pore constrictions. The
exact nature of this relationship is, however, extremely difficult to determine.
Hence, we write, in a general way
(5.2.13)
While (}"2i can be obtained from Equation (5.2.10), the porosity, ¢i, can be
determined by writing an unsteady mass balance equation for the fine particles
in solid phase, with an assumption that the change in porosity results from the
piping of particles.
An unsteady mass balance for fine particles in the solid phase can be written
after simplification, and is given as
(5.2.l4)
Here C s is the concentration of fines present on the pore wall, kgm- 3 . From
the knowledge of initial porosity, ¢o, we can also write
¢i = ¢o at t = 0 for i = 1, ... , N. (5.2.15)
We shall now discuss conditions that determine whether plugging or piping

will occur. Khilar et al. (1985) specified the conditions for plugging or piping
as a simple threshold type of condition. If in any unit cell, the concentration of
fines goes above a critical particle concentration, epe, then plugging occurs
due to pore bridging and convective jamming. If, however, the concentration
remains below epe, piping occurs. This simple criterion for plugging yields
reasonable agreement with observations (Khilar et aI., 1983; Khilar et al.,
1985).
The critical particle concentration, epe is expected to depend on (a) ratio
84 CHAPTERS
of fine size to pore constriction size, (b) the hydrodynamic conditions near the
pore constrictions and (c) topological parameters of porous medium.
5.3 Application to the Phenomenon of Soil Erosion
The phenomenon of soil erosion is described in Chapter 1. Here, we apply the

general model developed in the preceeding section to determine whether soil
erosion will occur or not.
We make the following simplifying and reasonable assumptions in order to
solve the general model (Khilar et at., 1985).
1. A capillary model is used to represent the geometry of the pore structure

of the soil mass.
2. The porosity increases due to the erosion, and the consequent changes in
the permeability and the flow rate are accounted for using the capillary
model.
3. The soil erosion is assumed to occur entirely due to hydrodynamic forces.
The shear stress at the wall is linearly related to the velocity (Tw = av).
4. To simplify the calculations further, only piping is considered. That is,
the concentration of particles is monitored until it reaches CPC. Beyond
CPC, soil erosion does not occur due to entrapment of particles at the
pore constrictions.
We rewrite Equation (5.2.2) on substituting for d¢ddt from

Equation (5.2.10) and for the release term from Equation (5.2.4) and setting
the term representing the rate of capture as zero
dCi
qCi- 1 - qCi + ochi(aq - Tc)Asil¢i = l¢iCi(ochi(aq - Tc))Asi + l¢iTt·
(5.3.1)
We note that Tw is replaced by aq. Using a simple capillary model consisting

of capillaries of diameter b, it can be readily shown that,
(5.3.2)
(5.3.3)
(5.3.4)
(5.3.5)
The qo and 00 are the flow rate and pore diameter at time, t = O. To determine
the pore diameter 0 at any time, t, we write a mass balance of eroding fine
particles for a pore,
(5.3.6)
where L is the length of the pore.

We also use the following dimensionless variables
D=-,
o (5.3.7)
00
Equations (5.3.1) and (5.3.6) are written in the dimensionless variables as
d~i (2
dO + D + N Fi -
NGi) ~i =
D ( N Fi -
NGi) + D 2 ~i -
D 1
fori = 1, .. . ,N (5.3.8)
and
dD 1 N
-dO = -"CNF"D
2N ~ t
-NG")
t .
(5.3.9)
t=!
After substitution for 00 and qo from the capillary model and Darcy's law
respectively, N Fi and N Gi are given as
(5.3.10)
and
N "= lJ.Lahi (~) (2¢0)3 / 2 (5.3.11)

Gt Cs I::lPj L Ko
Equations (5.3.8) to (5.3.9) can be solved simultaneously with initial

86 CHAPTER 5
10,--------------,
0-90
0.60
o.m
0.60
050
~.~
·0.40
030
0.20
0.10
Figure 5.3.1. Dimensionless clay particle concentration in the last segment

(1/J N) as a function of dimensionless time (0) for various values of mitial
permeability Ko (Khilar et ai., 1985).
conditions
1/Ji =0 and D = 1 at 0 = o.
Khilar et ai. (1985) solved these equations numerically to compute the
concentration of fine particles in unit cells as a function of time. The following
values for the model parameters have been used: <Po = 0.40; l = 0.005 m;
C s = 750 kgm- 3 ; Tc = 0.50 Pa; L = 0.20 m; t::.P/L = 2 x 106 Pam-I;
Cihi = 10- 6 kgN- I s-I.
These values are typical for seepage of water through a clayey soil.
Decreasing the length of the segment lower than 0.005 m yields similar
results, and hence the results obtained for l = 0.005 m are representative
in nature. The details concerning the choice of the parameter values are given
elsewhere (Khilar et al., 1985). It is sufficient to monitor concentration of fines
in the last segment, 1/J N, since the last segment will eventually have a particle
concentration higher than other segments due to the convective transport of
particles from other segments to the last segment.
Figure 5.3.1 shows the dimensionless concentration of fines in the last
segment, 1/J N, as a function of dimensionless time 0 for various values of the
initial permeabilities, K o , with other parameter values remaining the same.
We observe from these plots given in Figure 5.3.1 that the rate of rise in
concentration is slower in the case of more permeable soil masses. We therefore
conclude that piping of eroded fine particles is more likely to occur in earthen
embarkments of high permeability. Likewise the eroded particles are more
likely to get entrapped at the pore constrictions in earthen embarkments of
low permeability. Such results are in qualitative agreement with a number of
observations made over the years.
Some insights can be obtained from these modeling studies. The
permeability of soil mass Ko is a crucial parameter determining whether
piping or entrapment occurs. More permeable media are usually more porous
and hence the critical particle concentration is expected to be higher. On the
other hand, the concentration of fines is dependent on the flow rate through the
medium which is determined by the permeability. The concentration of fines
on the pore wall, C s and the erosion rate (0:hi and Tc) are also the critical factors
to determine whether entrapment or piping occurs. These parameters strictly
depend on the type of the soil mass, the compaction, and the other formation
conditions. It can be mentioned here that the piping or plugging of organic and
biological fine particles can also be dealt with using the approach presented in
this section.
5.4 Application to Sand Filtration
Sand bed filters have been used in treating waste water for many years. The bed
may consist of natural silica sand, crushed anthracite coal, crushed magnetite
and garnet sands. A detailed description of conventional sand bed filters can
be found elsewhere (Fair et al., 1968).
The model developed for delineating the conditions for the occurrence of
entrapment or piping can be applied to the design and operation of sand bed
filters particularly to prevent the bed from plugging both during the filtration
and regeneration or washing. We shall first consider the filtration operation. The
sand bed is used as a deep bed filter. The particulate matter must get collected
on to the sand grains and at the same time the bed must not get plugged at
any location. We know that plugging may take place if the concentration of
the particulate matter is above a critical particle concentration, CPC. While a
reliable theory to predict the critical particle concentration is not available, we
can get some estimates, from the knowledge of the ratio of particle size to grain
size. The higher the ratio of sizes, the higher the CPC will be. Typical values
of CPC vary between 1 to 10% (v/v). In other words, the volume percentage
of the particulate matter and its average size are of primary importance in
88 CHAPTER 5
selecting the grain size of the sand filters.

During washing, fine particles are dislodged from the granular bed sites
and go into suspension. In a way, the phenomenon is similar to soil erosion.
Therefore, the materials presented in the previous section can be directly
applied in this case, with some minor changes. The rate equation for detachment
or release of captured particles during washing is likely to be similar to the
entrainment rate, presented in Chapter 6. Sandbed, being an unconsolidated
porous media, Ergun type equation can also be used to describe the flow rate-
pressure drop relationship. Nevertheless, the approach will be similar even
with these different inter-parameter relationships.
5.5 Application to Water sensitivity of Berea Sandstone
We shall now apply the model developed in the preceeding section to water
sensitivity of Berea sandstone which is briefly described in Chapter 1.
Khilar et al. (1983) have experimentally obtained a critical rate of salinity
decrease, CRSD above which the drastic permeability reduction occurs as
a result of the entrapment of fine particles. In these experiments, the salt
concentration was reduced exponentially with the time (exp( -t IT)). The space
velocity, which is reciprocal of space time, T, is a measure of the rate of
salinity decrease. Figure 5.5.1 shows the final sandstone core permeability as
a function of the space velocity. The CRSD at two superficial velocities of 24
and 2.4 crn s-I are 0.4 and 1.0 hr- I respectively. The CRSD is also found to
depend strongly on the nature ofthe salt solution. This finding ofCRSD is very
significant in that one can selectively cause entrapment or piping by choosing
the rate of decrease in salinity either above or below CRSD. The fines are
released due to colloidal forces, and hence the Equation (5.2.6) is used. The
release coefficient O'ci is however assumed to depend on the salt concentration
in the following simple manner:
O'ci = 0 for Gsa < CSC
(5.5.1)
Here, Gsa is the salt concentration, CSC is the critical salt concentration, and
O'co is the maximum value of the release coefficient.
The solution of the Equation (5.2.2) is sought for the case in which
entrapment is virtually absent. Under this condition, we can assume that {3
Flow Rate
o 130 cc/hr
A 120 cc/hr
o 12 cc/hr
0.001 L-...L....JL...U..I.JJ..L'---'-''-'-'-.LLJ,.L;L-.-..I........L-'-1.LJ..W_''--_ _ _ _ _--'

0.01 0.10 1.0 10.0
SPACE VELOCITY (Hr- 1)
Figure 5.5.1. Variation of permeability reduction with rate of salinity decrease

(Khilar et al., 1983).
is equal to zero and that the permeability and porosity remain constant.
Using the following dimensionless variables e = tiT, Wi = Ci/lJlo, we
obtain the equation
(5.5.2)
for i = I, ... , N.
Here, T is the space time and is a measure of the rate of decrease in salinity
and T is residence time of a segment (T = l I v).
Equation (5.5.2) was solved simultaneously for N segments with condition
Wi = 0 at t = 0 and Wo = 0 for all t. As in the case of soil erosion, we focus
the concentration of particles in the last segment, W N.
e
Figure 5.5.2 presents WN as a function of at different values of space
velocity T- 1• We observe from the figure that the concentration of fines
increases sharply, goes through a peak and then decreases due to retardation of
the rates of release resulting from the depletion of fines on the pore surface. It
is assumed that if the peak value is above the CPC, entrapment would occur.
Taking a reasonable value of CPC for Berea sandstone, it can be seen from
this figure that when the values of CRSD are equal or to above 0.45 he 1,
90 CHAPTER 5
16
v' 25.6 em Ihr

14 CRSD: 0.43
o/p
10- 4 12
101----I-7"""'==--..----"......------~pc
8
S=O.72hr- 1
4
O.3Shr-'
2
O.18hr- 1
Figure 5.5.2. Dimensionless particle concentration at various space velocities

(Khilar et aI., 1983).
the concentration of particles in suspension goes above CPC implying that

entrapment occurs, leading to a drastic reduction in permeability.
This value of 0.45 hr- I agrees very well with the measured value. Likewise
at the small velocity of 2.4 cm S-I , the calculation for CRSD agrees with the
measurement. This study was focussed on determining CRSD, and no attempt
was made to calculate the small and slow reductions in permeability during
slow salinity decrease experiments. Vaidya (1991) rigorously modeled the
variations of the release coefficient resulting from slow decrease and further
used the release-capture model to be discussed in Chapter-6 to describe the
permeability decline in these cases. The release coefficient was assumed to be
dependent on the double-layer potential, which in tum is dependent on the pH
and ionic strength of the permeating liquid. The details of this approach can
be found elsewhere (Vaidya, 1991).
CHAPTER 6
MATHEMATICAL MODELS FOR PERMEABILITY

REDUCTIONS DUE TO MIGRATION OF FINES
In this chapter, we present mathematical models to describe the reductions

in permeability resulting from migration of fines in porous media. First, we
present the release and capture mechanisms followed by the model equations,
describing the processes. Finally the solution procedures are discussed and the
model predictions are compared with experimental observations.
6.1 The Release and Capture Mechanism
Mathematical models describing the drastic and rapid permeability decline

due to migration offines in porous media are based on the release and capture
mechanisms originally proposed by Khilar and Fogler (Khilar and Fogler,
1981; Khilar and Fogler, 1987). These mechanisms are concerned with the
release of fines particles from the pore wall of the pore chambers and the
capture of these released fine particles at the pore constrictions. The principles
can be elucidated by considering a unit cell that consists of a pore chamber
with two pore constrictions, one at the inlet of the chamber and the other
at the outlet, as shown in Figure 6.1.1. For simplicity, we do not show the
interconnectedness between the chamber and the pore constrictions.
We observe from Figure 6.1.1, that the fine particles are attached to the pore
wall. During the flow of liquid, these fine particles are released by colloidal
forces (as in water sensitivity of sandstones) or by hydrodynamic forces (as in
soil erosion) or by a combination of these forces. Once these fine particles are
released, they are entrained in liquid, where some particles may get entrapped
at the pore constrictions during the flow in the porous media.
Figure 6.1.2 shows the three sites of fine particles, namely on the pore wall
of the pore chamber, in suspension, and at pore constrictions. In this chapter,
we are, however, concerned with the reduction in permeability, and therefore
we shall assume that the conditions are favorable for entrapment.
91
92 CHAPTER 6
pore body
pore" constriction
Figure 6.1.1. A schematic diagram of pore-fine system before fresh water

flow (Khilar and Fogler, 1987).
Particles attached to pore wall

<::)
o Particles in suspension
C> Captured particles
Figure 6.1.2. A schematic diagram of pore-fine system after fresh water flow
(Khilar and Fogler, 1987).
MODELS FOR PERMEABILITY REDUCTIONS 93
The nomenclature for the concentrations of particles whether expressed in

mass or number of particles per unit of an appropriate volume or surface is as
follows. The concentration on the pore wall is a\, while that at the constriction
is a2, further, the concentration in suspension is C. All these concentrations are
expressed mass(gm) or number per unit pore volume. We assume that the fine
particles neither flocculate nor break into parts and hence they maintain their
identity as they get transported from one site to another. The drastic reduction
in permeability is strongly dependent on the quantity of particles captured at the
pore constrictions. Therefore, the primary objective of the modeling excercise
to describe the reduction in permeability resulting from the migration offines
is to write the balance equations for fine particles in such a manner that the
amount of particles retained at the pore constrictions can be determined.
Initially all releasable fines are on the pore wall i.e., a\ = alO, there are
no released fines in the suspension, i.e. C = 0, and there are no entrapped
fines at the pore constrictions, i.e., a2 = 0. The release of fines begins after
the salt concentration drops below the critical salt concentration, ese or when
the flow velocity is higher than a critical flow velocity, Vc. The released fines
remain dispersed in a peptizing environment offresh water and are carried with
the flow until these fines are captured at the pore constrictions by means of
surface deposition, and/or convective jamming or bridging or size exclusion.
As the constriction sites are plugged by released fines, the permeability is
reduced sequentially from the inlet to the outlet. This sequential reduction in
permeability has been experimentally observed as shown in Figure 6.1.3 in
Berea sandstone (Khilar and Fogler, 1981).
In this experiment, pressure tappings were used at the inlet face, at one-
third of the length and at two-third length of the core. The pressure drops
- i::lp\, across the inlet segment, b.P2 across the middle segment, and i::lP3
across the outlet segment - were measured as a function of pore volumes of
liquid sent through the core. We observe from Figure 6.1.3 that shortly after
the flow is changed from salt water to fresh water, b.P\ begins to increase
while i::lP2 and i::lP3 remain unaltered. We further observe that after a certain
period of time (say after 5 pore volumes of water have passed through), b.P3
begins to increase while b.Pj and b.P2 have almost increased to an asymptotic
limit. These observations indicate that the permeability reduces sequentially
along the length of the core. This occurrence results from the fact that fines
will only be released when the salt concentration drops below the esc. The
overall effect of this sequential reduction in permeability results in a moving
permeability front, which separates a zone in the rear whose permeability is
94 CHAPTER 6
100
90 o D.p,
"IIPz
[J AP3
80
70
60 lip, /lpz lip,

lip
(Psi)
50 -[[]]-
40
30
20
1O
6 7
Figure 6.1.3. Sequential reduction of permeability in water sensitivity of

Berea sandstones (Khilar and Fogler, 1981).
somewhat completely reduced and a zone in the front whose permeability

is unaffected. The reduction in permeability approaches an asymptotic limit
which depends on the size distribution offine particles and pores as well as on
the content of the fine particles in the sandstone.
The essential features o~ the model are the rate equation for the release of
fines, the rate equation for the capture of fines or balance on the open pores,
the balance on suspended fine particles, and the expression for reduction in
permeability resulting from entrapment or from the reduction in number of
open pores. The model formulation is presented in an unified manner based on
four modeling studies (Khilar and Fogler, 1983; Groesbeck and Collins, 1982;
Sharma and Yortsos, 1987c; Vaidya, 1991).
6.2 Rate of Equations for the Release of Fine Particles
The process ofthe release of fine particles from the pore wall due to colloidal
and hydrodynamic forces has been discussed in Chapters 3 and 4. It is noted
that a universal rate law for the release of fines does not exist. Nevertheless,
several rate equations have been proposed. We shall tabulate these equations
here, after discussing briefly two important features of these rate equations.
The two important features of these rate equations are that these are of first
order with respect to the concentration of fines on the pore wall and that they
have a threshold characteristic. That is, the fine particles are released only
beyond a certain value of a particular parameter. The threshold results because
a minimum change in parameter is required to cause an imbalance with the
the colloidal or hydrodynamic forces. The precise reason as to why the release
process is first order is not clearly understood. Experimental measurements
indirectly related to the release of fines, nevertheless, support the assumption
of a first-order release process. As discussed in Chapter 3, a theoretical rationale
for a first-order rate in the release of Brownian fine particles can be obtained
from the theory of Dahneke (l975a;1975b).
Table 6.2.1 presents the rate equations used by different investigators in
their studies relating to the release and migration of fines in porous media. As
mentioned earlier, these are first-order, and most of these equations have the
threshold characteristic.
6.3 Rate Equations for the Entrapment/Capture of Fine Particles
Entrapment of fine particles at the pore constrictions occurs primarily due to

surface deposition by direct interception, by mechanical bridging or convective
jamming, and by size exclusion.
Figure 6.3.1 shows the schematics of three types of entrapment at the pore
constrictions. We observe from this figure that while surface deposition offine
particles reduces the cross-section area of the constriction, both multiparticle
bridging and size exclusion types of entrapment result in blockage or closure
of the constrictions.
Khilar and Fogler (1983) suggested a simple rate expression for capture of
particles by direct interception:
Tc = {3C. (6.3.1)
Here, {3 is the capture coefficient.

Based on the concept of capture of particles by a single collector by means
of direct interception (Yao et al., 1971) and geometrically modeling the pore
constriction as the space between two spherical collectors, (see Figure 6.3.2)
it can be shown that the capture coefficient is given as(Khilar, 1981)
371" 2
{3 = SeNd v, (6.3.2)
\0
0'1
TABLE 6.2.1. Rate expressions for release of fine particles
Expressions Remarks Reference
1",. = (\'IT I Colloidally induced release Kolakwski and

of Brownian fine particles Matijevic, 1978.
in model packed columns.
1"r = C\'O"I Colloidally induced release Khilar and

0= 0 forC sa < CSC of clayey fines particles in . Fogler, 1983;
(\ = n forC sa > CSC Berea sandstone. Here, (t is a Vaidya, 1991.
function of C.,pH
and temperature.
n
::r:
l' r = 00"1 Diffusional particle transfer Sharma and >-
"1::1
-I
It = 2D tTl
(2 . 1:. )1/3 with no energy barrier. Levich's Yortsos, )<:I
II J)"II 0'>
x_l- solution for a capillary 1987a.
RI"
is averaged for a porous medium.
1"r = nO" I Diffusion limited flux through Ryan and

0= [(k[JT)/( 61fILa p )] 1/3 the boundary layer. Gschwend,
xu -4/ 3v 2/3 1994.
9
1",. = n(11 - 11c)al for v> VI' Hydrodynamically induced Gruesback and
'1",. = 0 for 11 < VI' release of fine particles in sandstones Collins, 1982; Khilar and Sarma, 1990
1",. = o.4(T, .. - T,.)al fi.lr T", > T,. Ilydrodynamically induced Arulanandan et aI., 1975;
'1",. = 0 for T", < T,. release offine particles in soil mass Khilar and Sarma, 1990
Note, (t has different units, each depending on the rate expression.

ap particle radius ag grain radius

Gsa salt concentration DB Brownian diffusivity
jp pore size distribution frequency lp length of the pore
Np = total number of pores
Pore body
~e particles
~re constriction
Surface deposition
Direction Fine particles

of
flow
Multiparticle bridging
A fine particle
Size exclusion
Figure 6.3.1. Conceptual diagrams of entrapment of fine particles at a the
pore constriction.
where N is the number of collectors per unit volume, d is the diameter of the
spherical collectors, e is the retention efficiency, and v is the average interstitial
velocity in the pore space. Experimentally it has been shown that f3 is directly
proportional to the flow velocity (Khilar, 1981).
This simple one-parameter expression, however, does not incorporate the
strong effect due to the mechanical bridging. In order to account for bridging,
Gruesbeck and Collins (1982) proposed a two-parameters rate expression as
98 CHAPTER 6
Spherical
collector
fine
particles
Figure 6.3.2. A conceptual diagram of deposition of fine particles in the space

between two spherical collectors.
given below:
rc = (~ + ba 2 ) V C, (6.3.3)
where f3 is the capture coefficient and b is an adjustable parameter. The term

Da2 vc gives the additional capture of fine particles due to bridging, which
increases with the amount of particles deposited.
A statistical approach has been used to determine the rate of pore closures
when size-exclusion dominates particle retention.
Here, we determine the probability of fine particles reaching pores of size
equal or less than the particle size. The rate of capture is governed by the
size distributions of fines, j and of pores, jp, by the flow velocity, v and by
the pore length lp. It has been shown by Rege and Fogler that the higher the
overlap between these two distributions, the higher is the pore closures due to
entrapment (Rege and Fogler, 1987). Using a statistical approach, Sharma and
Yortsos (1987a) have derived an expression for the rate of pore closures for
pores of size interval (rp and rp + drp) on the basis of the rate at which the
particles of size greater than r p are entrapped. It is shown that the rate of pore
closure of constriction size between r p and r p + dr p is given as
(6.3.4)
where P( r s) is the fraction of particles in the size interval r sand r s + drs

retained at each pore contriction at each step and is given as
(6.3.5)
6.4 Mass and Population Balance Equations for Fine Particles at

Different Sites
To determine the concentration of fine particles captured at the pore

constrictions, IJ2, we need to account for the fine particles at the three sites: on
the pore surface, in suspension, and at pore constrictions. Accounting for the
fine particles is done either by mass balances, (Khilar and Fogler, 1983) or by
population balances in which case the particle size distribution is incorporated
(Sharma and Yortsos, 1987a). First we shall write the mass balance equations.
Mass balance equations derived by Khilar and Fogler (1983), make use of the
following assumptions.
(I) The change in porosity is small.
(2) The variations in concentrations are only significant in axial direction. It
is a I-D problem.
(3) Brownanian diffusion offines is neglected.
The unsteady mass balance equation for the fine particles on the pore surface
equates the rate of depletion of fines to the rate of release:
00-1
at = -rr, (6.4.1)
with
at t = 0. (6.4.2)
Likewise, the unsteady mass balance equation for the fine particles 10
suspension can be written as
oC oC
at + v ox + rc - rr = 0, (6.4.3)
100 CHAPTER 6
with
C 0, at t = 0, x 2: 0, (6.4.4)
C 0, at x = 0, t 2: 0. (6.4.5)
An unsteady mass balance equation for the fine particles at the pore
constriction equates the rate of capture to the rate of accumulation:
(6.4.6)
with
at t = 0. (6.4.7)
Equations (6.4.1) to (6.4.6) with appropriate expressions for the rate of

release and capture can be solved simultaneously to obtain 0"2(X, t). More
information, however, is required to solve these equations. Specifically, we
need to know the following:
1. The nature of release (colloidally induced or hydrodynamically induced).
2. The release coefficient, 0:, and its dependence on the system variables.
In colloidally induced release, the local release coefficient varies with
salt concentration. In hydrodynamically induced release, the release
coefficient will vary with the flow velocity and the critical shear stress.
3. The capture coefficient, f3 and its dependence on the system variables,
particularly the flow rate.
4. The conditions for which the equations can be solved i.e., for constant
flow velocity or for constant pressure drop. In a mixed case, the equations
in principle can also be solved.
Accounting for fine particles can also be achieved by means of a population
balance approach. Sharma and Yortsos (1987a;1987b;1987c) have applied a
population balance approach to the problem of migration offines. We shall use
the notations already introduced in this book to write the original equations.
(6.4.8)
a(Cjc)
at + v a(Cjc)
ax + r c j c - j _
rr r -
°, (6.4.9)
(6.4.10)
(6.4.11)
Here, the concentrations 0"1, C and 0"2 are expressed in number of particles per
unit pore volume and hence represent the quantities integrated over the entire
size range. The size distributions fl' fe, and 12 are of particles on the pore
surface, in suspension, and at pore constrictions respectively. Equation (6.4.11)
represents the rate of pore closures. In this equation, Np is total number of pores,
fp is the pore size distribution and Tpe is the rate of pore closures.
Equations (6.4.8) to (6.4.11) are written in rather general form and
a reasonable simplification can be made by assuming the particle size
distributions at the three sites to be the same (h = 12 = fe). This assumption
is expected to be valid when rate of release is rapid and the release process is
not strongly dependent on the particle size. Even with this simplification, one
needs two additional pieces of information to solve this set of equations; the
initial size distributions of the fine particles and that of the pores. The solution
procedures adopted by the investigators will be discussed in a later section.
6.5 Correlation between Entrapment and Permeability Reduction
A theoretical correlation has been neither reported between the permeability

(K) and the amount of entrapment (0"2). Nor is there a widely accepted
correlation based on experimental measurements. The inherent difficulties are
similar to the rather well-known ones associated in relating permeability to
porosity. Nevertheless, for the sake of comparing permeabilities at different
stages of entrapment, empirical expressions have been proposed (Khilar and
Fogler, 1983; Groesbeck and Collins, 1982). Two such expressions are used
specifically for the migration offines. These are:
(6.5.1)
!i =
Ko
(1 _ B0"2)2
0"10
(6.5.2)
The parameters a and B are constants for a given porous medium. Ko is the
original permeability of the porous medium. Equation (6.5.1) is proposed by
102 CHAPTER 6
Gruesbeck and Collins (1982) for mechanical bridging leading to plugging

of pore constrictions. Equation (6.5.2) is proposed by Khilar and Fogler
(1983). This equation is derived by considering a cylindrical pore constriction.
The entrapped particles are assumed to deposit uniformly on the cylindrical
surface. The decrease in flow resulting from this deposit causes an increase in
the hydraulic resistance which can be determined by using Hagen-Poiseuille
equation. The consequent decrease in permeability can be also obtained by
means of Darcy's law. The values of a and B are to be determined from the
experimental measurements for a given porous medium.
Sharma and Yortsos (1987a) used the effective medium approximation to
represent the porous medium by an effective conductance, gm. With this
approach, the permeability reduction can be written as
(6.5.3)
Here gmo is the initial effective conductance of the medium. The instantaneous
effective conductance, gm is related to the number of open pores which in tum
is related to the amount of entrapment.
The effective medium conductivity, gm, can be approximately obtained from
the following set of equations:
10 00
G(gR)t(gR) dg R = 0, (6.5.4)
() gm - gR
(6.5.5)
t gR = gR - (1 - (Zj2)) gm
gR ex r; (2 < n< 4) (6.5.6)
Here, Z is the coordination number.
6.6 Solution Procedures, Results and Comparisons with Experimental

Measurements
There are three reported modeling studies concerning the in situ generation and
migration of fines in porous media.(Gruesbeck and Collins, 1982; Khilar and
Fogler, 1983; Sharma and Yortsos, 1987c). In each of these studies simplifying
assumptions are made and the equations are solved to describe the decline in
permeability due to migration of the fine particles in porous media. Gruesbeck
and Collins (1982) addressed the problem of hydrodynamically induced fines
migration. For colloidally induced fines migration, Khilar and Fogler (1983)
used a mass balance approach, while Sharma and Yortsos ( 1987c) have used a
population balance and pore closure approach. We shall discuss these models
separately as each has some unique features.
6.6.1 THE MODEL OF GRUESBECK AND COLLINS FOR

HYDRODYNAMICALLY INDUCED MIGRATION OF FINES
Gruesbeck and Collins (1982) considered two types of depositions in porous

media: plugging type of deposition in small pores and surface type of depostion
in large pores. Thus, they conceptually partitioned the fluid flow path into two
parallel flow pathways: plugging and non-plugging pathways. They further
assumed that the fraction of cross-sectional area made of plugging pathways
is given by J, which is determined by the pore and fine size distibutions. With
J defined, the pore velocity v and pore deposition 0"2 are written as follows:
v = JvP + (1 - j)v np ) (6.6.1)

0"2 = J0"2p + (1 - j)0"2np) (6.6.2)
Here, the subscripts p and np refer to the quantities III plugging

and non-plugging pathways. They considered hydrodynamically induced
release/entrainment only in non-plugging pathways and the rate equations
for deposition can be written for the two pathways as
(6.6.3)
(6.6.4)
Note that Equations (6.6.3) and (6.6.4) are similar to the equations presented
in Sections 5.2 and 5.3. Likewise, the relationship between the permeability
and amount of entrapment are:
(6.6.5)
104 CHAPTER 6
and
(6.6.6)
where a and E are the phenomenological constants. The mass balance equations,
Equation (6.4.3) can be written for plugging and non-plugging pathways as
follows:
(6.6.7)
and
BC BC Ba2np
at + v np Bx = -----at. (6.6.8)
The partition of flow for the two parallel flow pathways can be determined
by the equation given as follows:
vp _ kp(a2p)
(6.6.9)
v kp(a2p) + knp (a2np)·
Gruesbeck and Collins (1982) numerically solved Equations (6.6.7)

and (6.6.8) with specific values of various phenomenological parameters of
model to obtain a2p and a2np. These values were used to calculate kp and
k np , which in turn were used to calculate the pressure rise at constant pore
velocity v. Figure 6.6.1 presents the calculated increase in pressure drop as a
function of pore volumes ofliquid permeating for two different flow velocities.
Interestingly, we observe from this figure that the permeability decline or the
rise in pressure drop is higher for the lower flow velocity, an observation
confirmed by their experimental measurements using the packed bed of sand
grains. The rise in pressure is expected to be a strong function of f. With high
values of f, the rise in pressure drop will be higher than shown in Figure 6.6.1
and may resemble to that observed in water sensitivity of sandstones.
6.6.2 THE MODEL OF KHILAR AND FOGLER FOR COLLOIDALLY

INDUCED MIGRATION OF FINES
Khilar and Fogler (1983) simplified the model equations by dividing Berea
sandstone core into segments and assuming that there are no spatial variations
(l..- 5.-----------------...,
00.
0:: <J 4
0'6.
~<J 3
~ ~ 2
(l)Q
~~ 1
(l.~ O~---~--~~-~~--~~--~~
500
PORE VOLUME
Figure 6.6.1. Numerical simulation of deposition and entrainment experiment
(Groesbeck and Collins, 1982).
in any given segment. The release of fine particles occurs only after the salt
concentration in that segment is decreased below the CSc.
We consider ith segment (see Figure 5.2.1) and simplity the Equation (6.4.3)
to write
(6.6.10)
The rate of colloidally induced release of fines is written (see Table 6.2.1)
(6.6.11)
where
ai = 0, for C sai > CSC,

for C sai < CSC,
and C sai is the salt concentration of the liquid in the ith segment.
Likewise the rate equation for capture is reproduced from Section 6.3 as
(6.6.12)
Let ti be the time required for the salt concentration in ith segment, C sai to
106 CHAPTER 6
reach the CSC. Equation (6.6.11) can be integrated to obtain
(Tli = (TliO, for t < ti,

(6.6.13)
(Tli = (TIO exp[-ai(t - ti)], for t ~ k
Equation (6.6.10) can be solved after substituting for (Tli to obtain Gi , which
can then be substituted in Equation (6.6.12) to obtain (Khilar and Fogler, 1983)
(T2i = 1 _ ai exp[-,6(t - ti)]- ,6exp[-ai(t - ti)].

(6.6.14)
(TIO ai - ,6
The time ti can be determined using the tank-in-series model by the

following equation
CSC
Gsao
( ti )
-exp -
Ti
-2:- j!
j=O
-
i-I 1 (ti)
Ti
j
=0
'
(6.6.15)
where G saO and Ti are the initial salt concentration and the residence time of
the ith segment, respectively.
Knowing the amount of entrapment in the segment, (T2i (t), the permeability
Ki(t) can be determined using Equation (6.5.2). The overall permeability for
a core divided into N, equal segments can be calculated as
(6.6.16)
Khilar and Fogler (1983) carried out calculations using the above set of
equations to determine the rapid and drastic permeability decline in Berea
sandstone. Simulations showed that a segment of length of 5 mm gave the
same results as segments of length lower than 5 mm, and therefore a segment
length of 5 mm was used in all calculations. The idea of treating a segment
of a core as an independent unit for permeability decline calculations emerged
from their experiment with thin cores. Figure 6.6.2 presents the permeability
reduction of thin cores as a function of core length. We observe that, for cores
of lengths greater than 5 mm, the permeability decline is independent of the
core lengths while with cores of length below 5 mm, the permeability decline
decreases with the decrease in core length (Khilar, 1981).
Khilar and Fogler (1983), compared the model predictions with chosen
values of a and,6 for a number of water shock experiments of Berea sandstone
1.0 , - - - - - - - - - - - - - - - - - - ,
BEREA SANDSTONE
Flow Rote = 2.7 It 10- 2cc/sec
0.01
O.OOIL...-----::--'::--::-----:~-_=_==__~:-:::-__:_:=_=_--'----'
0.25 0.50 0.75 1.00 1.25
LENGTH OF THE CORE (Inches)
Figure 6.6.2. Permeability reduction in thin cores (Khilar, 1981).
at different flow rates and temperatures. Figures 6.6.3 and 6.6.4 show the
comparison between the model predictions and the data at two different flow
rates, with different values of 0: and (3.
It was found that both 0: and (3 vary linearly with the flow velocity. While
the variation of (3 with flow velocity can be rationalized by its effect on the
capture resulting from direct interception, the variation of 0: with flow velocity
was shown to be related to the rate of variation of salt concentration with the
flow rate. Both the values of 0: and (3 vary from 10- 4 to 10- 2 S-I in the range
of fluid velocities 10- 6 to 10- 4 ms-I.
Figure 6.6.5 shows the comparison of the model predictions with the data at
three different temperatures of 0, 30, and 60°C. We observe that the temperature
affects both the initiation of permeability reduction and the rate of permeability
reduction. The effect on initiation of the decline is related to the effect of
108 CHAPTER 6
BEREA
I" Oio., j" Length
Q ~ 2.77 x 1O- 4 cc/sec
1.0,-<=>---<"-.;;::------------------,
0.6 o Doto
- - Model
0.6
k/kI
0.4
0.2 o
o 0
1.0 2.0 3.0 4.0 5.0

PORE VOLUME
Figure 6.6.3. Comparison of model with experiment (Khilar and Fogler,

1983).
BEREA
1'·O.a,"· Length
Q ~ 6.22 x 10- 2 cClsec
1.0
0.8 0 Dolo
- - Model
0.6
k/kr
04
0.2
0
0 1.0 2.0 3.0 4.0 5.0
PORE VOLUME
Figure 6.6.4. Comparison of model with experiment (Khilar and Fogler,

1983),
BEREA SANDSTONE
I" Dio., 2" Length
Flow Rate: 3.61 .IO·'cc/sec
1.0,-aJ(~-40::----------------,
O,D,lI Dolo
08
- Model
o·c
0.6
k/kI
0.4
0.2
O.L--L__~-J~~~~~__~~~~~
o 10 2.0 3.0 4.0 5.0
PORE VOLUME
Figure 6.6.5. Effect of temperature on the permeability reduction (Khilar and

Fogler, 1983).
temperature on the CSC. The observed higher rate of reduction in permeability

at higher temperatures indicates that the release coefficient increases with the
temperature.
6.6.3 THE MODEL OF SHARMA AND YORTSOS FOR COLLOIDALLY

INDUCED MIGRATION OF FINES
Sharma and Yortsos (1987c) numerically solved Equations (6.4.6) to (6.4.10)

for a single particle size distribution and when the pore closure occurs resulting
from size exclusion. Solutions were obtained to determine the number of open
pores as a function of axial distance and time. This information was then used
to compute effective medium conductance which was in tum used to compute
the local permeability. The overall permeability was calculated as the harmonic
mean of the local values.
The decline in overall permeability with time was computed for a number
of system variables such as rate of release, rate of capture, particle size, and
the coordination number. For high release rates Sharma and Yortsos found the
existence of a permeability reduction front that divides the formation into two
zones: one of original permeability and the other of reduced permeability. It
is further shown that the permeability reduction is a very sensitive function
of the coordination number of the medium and the particle size. Figures 6.6.6
110 CHAPTER 6
I. 0 ~=:::::::::::::r:~;::::~x~CDC!:J
0.9
0.8
0.1
0.6
"" o.~
INJECTION
0.4 B -1.0
0.3
0.2
0.1
O.OO!o----...L..------'2,....-J
Figure 6.6.6. Overall permeability, effect of coordination number Z (Sharma

and Yortsos, 1987c).
I.o.~----r-----'T""-'
O. A "'1.0
O.
O.
Q
A=o.a
INJECTION
Z-12
e- 1.0
CT- 0.5
A-aS8
Figure 6.6.7. Overall permeability, effect of fine size (Sharma and Yortsos,
1987c) (A is defined as the ratio of average size of pore to that of fines).
and 6.6.7 show the decline of overall permeability with time at different values
of coordination number and particle size respectively for a set of parameters
relating to the release and capture offines. We observe from the Figure 6.6.6
that the lower the coordination number Z, the higher is the rate of decline in
permeability. Likewise, we observe from Figure 6.6.7 that generation oflarger
fine particles results in rapid and drastic decline observed in the case of water
sensitivity of sandstones.
CHAPTER 7
USE OF NETWORK MODELS FOR PREDICTION OF

PERMEABILITY REDUCTION DUE TO FINES
ENTRAPMENT
by
Venkat Ramachandran, Kartic C. Khilar and H Scott Fogler
7.1 Need for network models
The mathematical models for predicting permeability reduction of a porous

medium due to fines entrapment discussed in the previous two chapters fall
under the class of continuum models. In a continuum model, the porous
medium is treated as a continuum within which quantities such as temperature,
fluid species concentration (e.g., ionic strength), and solids concentration are
defined as smooth functions of position. The use of such continuum models
is appropriate when the porous medium under consideration is homogeneous
in terms of the pore structure and the medium surface properties. If the above
conditions are satisfied, continuum models are attractive to use because of
their conceptual simplicity and their ability to provide useful insight into
the process of fines release and capture, and subsequent evolution of the
permeability of the porous medium (as seen in Chapters 5 and 6). Of course,
the assumption of a suitable relationship between the permeability and the
amount of retained fines is necessary. This assumption is typically based on
empirical correlations and is a major limitation of continuum models. When a
porous medium is spatially heterogeneous, however, it will not be possible
to rely on continuum models for accurate prediction of the permeability
reduction because the pore interconnectivity, which the continuum models
generally ignore, becomes crucial. As any particle transport process through a
porous medium involving entrapment is essentially a percolation phenomenon
113
114 CHAPTER 7
(Imdakm and Sahimi, 1987), the permeability of the porous medium is very
sensitive to the connectivity of the pore space. Network models can account
for the pore-level heterogeneity and interconnectivity of the pore space, and
hence are more suitable than continuum models for studying processes that
depend strongly on pore level effects. This chapter discusses the various issues
involved in constructing a network model and using it for predicting the
permeability reduction during fines transport within porous media. Network
models have been used to study several processes that depend strongly on
the pore structure such as dispersion (Sahimi et al., 1983), fluid displacement
(Simon and Kelsey, 1971; 1972), carbonate acidizing (Hoefner and Fogler,
1988), and catalyst deactivation (Sahimi and Tsotsis, 1985).
While all mechanisms of particle capture can generally be incorporated in
a network model, the focus here will be on straining dominated transport of
fines since deposition will not be a major factor under conditions conducive
for the release of fines.
7.2 Network Models
Network models are statistical in nature and consider the porous medium to
be an interconnected network of bonds and nodes. This idealization is based
on observations that the void space in natural porous media is an arrangement
of converging-diverging channels with a distribution of sizes (Dullien and
Dhawan, 1974), as shown in Figure 7.2.1. The bonds in the network represent
the flow channels or pore constrictions in the porous medium while the nodes
represent the pore bodies. The bonds determine the resistance offered to fluid
flow by the network, and hence, the permeability of the porous medium
being simulated. The nodes are the points of fluid mixing where the bonds
intersect. In a network model, interconnectivity refers to the manner in which
the nodes are connected to each other via the bonds. A network used to model
a heterogeneous porous medium will have a distribution of bond sizes and may
have a variable interconnectivity.
7.2.1 NETWORK CONSTRUCTION AND LATTICE ARRANGEMENTS
A realistic representation of the porous medium using a network model depends

on the arrangement of nodes in the network, interconnection of the nodes via
the bonds, and the shape and dimensions of the bonds. The bonds and nodes in
a network can be arranged in a number of ways. Networks can be two- or three-
dimensional. The nodes can be distributed on a regular grid (ordered) or can
NETWORK MODELS FOR FINES ENTRAPMENT 115
Pore Pore Bond Node

throat body
(bond) (node)
Figure 7.2.1. A Schematic diagram showing pore structure in natural porous

media (top) and its network model analog (bottom).
be arranged in a random manner by using a stochastic algorithm (disordered).

As mentioned previously, the interconnectivity of the pore space is a major
structural property of porous media. In a network model, the interconnectivity
of the network is dictated by its coordination number which is the total number
of bonds intersecting at a node (the total number includes bonds that support
flow both into and out of the node). For typical consolidated porous media, the
coordination number varies between 6 and 14 (Sharma and Yortsos, 1987a).
Depending on the lattice arrangement and interconnectivity, several types of
arrangements such as square, triangular, Voronoi, single hexagonal and double
hexagonal are possible (Figure 7.2.2). Of these different arrangements, the
116 CHAPTER 7
Square Triangular Voronoi
Single Hexagonal Double Hexagonal
Figure 7.2.2. Lattice arrangements used in network modeling.
Voronoi lattice is disordered while the rest are ordered. While the Voronoi
lattice most closely resembles real porous media due to its random nature, in
most of the applications the topology of the network used has been regular,
i.e., the coordination number of the network is constant. The use of regular
networks is justified by the work of J erauld et al. (1984) who have compared
the percolation and conduction properties of triangular networks with Voronoi
networks (the two networks have the same coordination number - triangular
has exactly six and the average coordination number of a Voronoi in 2D is also
6). Their work has shown that, as long as the average coordination number of a
regular network is equal to that of a topologically disordered system in 2D, the
transport properties ofthe two are essentially identical. Use of regular networks
is desirable because flow calculations are computationally more efficient in a
regular arrangement.
The bonds in a network, in principle, can have any shape and size. However,
in practice, they are usually assumed to be cylindrical in shape to which
effective radii are assigned. To represent porous media realistically, the bond
size distribution of the network should be similar to the pore throat size
distribution existing in the porous medium being considered. Therefore, the

radii of the bonds are usually selected from an experimentally measured pore
size distribution obtained by, for example, porosimetry. While most of the
network models have used cylindrical tubes for bonds, tubes with sinusoidally
varying cross sections have been used in a few applications (Lin and Slattery
1982; Dias and Payatakes, 1986; Hopkins and Ng, 1986). Lengths to the
bonds can be assigned by various methods: constant length, length directly
proportional or inversely proportional to the radius, and lengths having a size
distribution with a mean and standard deviation similar to that of the radii.
7.2.2 FLUID TRANSPORT THROUGH THE NETWORK
The methods for simulating fluid transport in network models depends largely
on the process being studied. A common approach used in petroleum- related
modeling is due to Fatt (1956) and involves obtaining an equation for the
pressure at each node based on fluid conservation at the nodes. The fluid
conductivity of a bond is governed by its shape, size, and length. For example,
in the case of cylindrical bonds, the conductivity is determined from the
Hagen-Poiseuille equation relating the pressure drop across a cylindrical tube
to the flowrate through the tube, the fluid viscosity, and the tube radius
and length. To determine the flow distribution in the network, solution of
the simultaneous equations for the pressure at all nodes is required. Once
a network is created and bond conductivities assigned, the use of effective
medium theory is an alternative to the numerical solution of the overall
conductivity of the network (Kirkpatrick, 1973). This technique involves the
replacement of the interconnected bonds in a medium, each with an averaged
bond having an effective-medium conductivity. However, since the bonds are
typically heterogeneous, use of this technique may not be suitable for modeling
phenomena that depend strongly on pore level effects.
7.3 Use of network models for studying particle capture
Network models were pioneered by Fatt (1956), who used a square network
of capillary tubes to study the flow of two immiscible fluids in porous media.
Since then, network models have been used in many diverse applications in
which the heterogeneity of the porous medium plays a crucial role. Here we will
discuss the evolution of the network model for studying particle entrapment by
straining and deposition during flow within porous media. Network models, as
will be shown later, have the ability to predict the permeability decline without
118 CHAPTER 7
the use of empirical relations. Donaldson et al. (1977) first used the network
model for studying particle flow through porous media. The network consisted
of a bundle of parallel capillaries with a distribution of diameters. Todd et al.
(1984) used a square network to study formation damage in oil reservoirs. As
the path of the particles within the network was modeled as a random-walk
process (pure diffusion), the influence of flow on the particle movement was
not accounted for. The model predictions were not in quantitative agreement
with experimental data. Houi and Lenormand (1986) proposed a similar
model in which the motion of the particles was biased towards the filter
represented by a square network. Their model did not account for particle
and pore size distributions which are known to strongly affect the evolution
of the permeability of the porous medium. Sharma and Yortsos (1987a;1987b)
used a combination of a network model representation of the porous medium
and an effective medium approximation, EMA to study fines migration and
deposition in porous media. Uniform deposition of the particles within the
bonds was assumed. While an interconnected network of pores was used to
represent the porous medium, continuum population balances, together with
the EMA, were used to obtain the flow field distribution. The model was
rigorous as it took into account pore interconnectivity and pore throat size and
particle size distributions, and predicted experimental data reasonably well.
However, since the EMA approximation is accurate only when the network is
far from the percolation threshold (Kirkpatrick, 1973), the applicability of the
model is limited to the initial stages of permeability reduction and in situations
where particle straining is dominant or where multilayer particle deposition
occurs within pores. Rege and Fogler (1987) developed a model to study
fines entrapment by straining or size exclusion. The model accounted for the
effects of fluid flow on the particle transport path and, unlike previous network
models, considered the simultaneous entry of a number of particles into the
network. With the use of a single parameter that was invariant for different
particle concentrations and for two different kinds of sandstones, the model
was able to match experimental data of Donaldson et al. (1977) quite well. The
model was extended subsequently to study entrapment of solid particles and
emulsions by simultaneous straining and deposition (Rege and Fogler, 1988).
Imdakm and Sahimi have also studied the straining dominated transport of
solid particles using a network model. Their model is similar in most of its
features to that of Rege and Fogler (1987), the only difference being that
Imdakm and Sahimi (1987) used a 2D square network (coordination number
4) while Rege and Fogler used a triangular network (coordination number 6).
Imdakm and Sahimi (1991) later presented a simple-cubic 3D network model

to predict the reduction in permeability due to fines entrapment by straining
and deposition. Burganos et al. (1992) have significantly improved model
predictions of deep bed filtration using a 3D network of sinusoidally shaped
bonds. Models studying particle deposition are further discussed in Section 7.5.
7.4 Application of network models for prediction of permeability

reduction due to straining dominated fines entrapment
In this section, the use of a network model to predict the evolution of the
permeability during fines transport through porous media will be demonstrated.
As the focus here is to show the ability of network models to account for
pore space interconnectivity and to calculate permeability without the use
of empirical correlations, straining is the only mechanism of fines capture
considered. The model to be discussed here is based on the work of Rege and
Fogler (1987) and Imdakm and Sahimi (1987).
7.4.1 NETWORK CONSTRUCTION
The first step in constructing the network involves the selection of a suitable
lattice arrangement for the nodes and bonds. Rege and Fogler (1987) used a
2D triangular network with cylindrical capillary tubes for bonds. This regular
lattice arrangement has a coordination number of 6 which equals the average
coordination number of a disordered Voronoi lattice. Imdakm and Sahimi
(1987) have used a 2D square network which has a coordination number of 4.
The standard practice while constructing a network is to assign no volume to
the nodes representing the pore bodies in the network. Thus, all the porosity
of the system is provided by the bonds (pores). To represent the porous media
realistically the bond size (tube diameter) distribution in the network should
be similar to the pore throat size distribution existing in the porous medium
being considered. The pore throat size distribution can be determined from
a sample of the porous medium using techniques such as photomicrography
and mercury porosimetry. In the case where it is not possible to determine the
pore throat size distribution, the Rayleigh distribution can be used to assign the
bond sizes. The density function for the Rayleigh distribution, !(R), is given
by
(7.4.1)
120 CHAPTER 7
where R is the bond radius, and ,,(-I is a characteristic pore radius. This
distribution mimics qualitatively the pore size distributions determined by
several investigators (lmdakm and Sahimi, 1987). Rege and Fogler (1987)
assigned lengths to the bonds from a distribution having a mean and standard
deviation similar to that of the bond size distribution. This assignment is
justified in modeling an isotropic porous medium and has the advantage of not
introducing another parameter in the model as would be the case if the bond
lengths are chosen to be proportional or inversely proportional to the bond
size. Note that the concept of having a distribution of bond lengths is entirely
mathematical as it is geometrically not possible to have a distribution of bond
lengths in a regular lattice.
7.4.2 CALCULATION OF FLOW DISTRIBUTION IN THE NETWORK
To determine the fluid flow distribution in the network, the pressure at each
node is to be determined. The pressures can be determined by the application
ofthe fluid conservation equation to each node.
The conservation equation written for the node shown in Figure 7.4.1 is
given by
(7.4.2)
where qk is the fluid flow rate in bond k. Similar to Ohm's law for current
through a linear electrical element, the flow rate through the bond can be
expressed as the product of the hydraulic conductivity of the bond, g, and the
pressure drop across the bond, !:l.P . Hence the conservation equation for node
i, j can be rewritten in terms of the pressure drop across the bonds connected
to it:
(7.4.3)
Depending on the sign of !:l.Pk, qk can be either positive (flow into the node)
or negative (flow out of the node). The hydraulic conductivity of a cylindrical
bond is obtained from the Hagen-Poiseuille equation for pressure drop during
laminar flow in a tube:
(7.4.4)
Z
tT1
...,
~
o
~
2::
o
tl
tT1
r-'
en
."
__ 'II '"I II '"' II,"' ,"" ,,"" , "'" , II'" , 'I'" , __ o
i<:I
."
~
en
tT1
1
~
i~ ~
2::
tT1
~
Figure 7.4.1. Regular triangular network and the application of fluid conservation at a node.
>-'
tv
122 CHAPTER 7
where Rk and Lk are the radius and length of bond k respectively, and J.L is
the fluid viscosity. The fluid conservation equations when written for each
node constitute a set of simultaneous linear algebraic equations the solution of
which gives the nodal pressures throughout the network. At the two horizontal
boundaries of the network, standard wrap-around boundary conditions are often
used. The wrap-around boundary conditions allow particles at the bottom of
the network to flow to corresponding nodes at the top. This boundary condition
eliminates edge effects and makes it possible to simulate large porous media
without using proportionately large networks. The pressure at the outlet is set
to zero. At the end of each time step, the pressure at the inlet can be calculated,
thereby allowing the determination of the macroscopic pressure drop across
the network which is (.A.niet - Poutlet), or simply .A.niet, as Poutlet = O.
For an MxN network, there are MxN equations (the pressures at the
MxN nodes) to be solved at each time step. Rege and Fogler (1987) used
the Gaussian elimination technique for inverting the conductivity matrix for
solving the pressure equations. While the use of a very large network may
allow accurate prediction of the permeability reduction, it will be prohibitive
in terms of the computational time and memory requirements. On the other
hand, as will be seen later, use of too small a network can result in significant
scatter in the calculated quantities.
7.4.3 PARTICLE MOVEMENT WITHIN A NETWORK
The particles are considered to be effectively spherical (circular in 2D) and the
radii are determined from a suitable distribution depending on the application.
It is generally assumed that particles move with the average velocity of the
fluid within the bonds. The disturbance of the flow pattern within the pore
due to the presence of the particle is therefore neglected. This assumption is
acceptable in light of the work of Leichtberg et al. (1976) who have shown
that, even for large ratios of particle diameter to bond diameter (as high as
0.95), the deviation of the particle velocity from the average fluid velocity is
negligible.
When a particle arrives at a node during flow within the network, it has a
choice of exit paths, i.e. choice of bonds through which the particles can leave
the node. To take the stochastic nature of the filtration process into account,
Todd et al. (1984) chose the particle path using the random-walk concept
wherein all paths have an equal probability of being chosen. Considering
particle movement within the network as a random walk process is, however,
unrealistic because it does not account for the influence of fluid flow on the
,.;---- 20%
10%
Figure 7.4.2. Application of flow-biased probability for selection of particle

path.
paths taken by the particle. Rege and Fogler (1987) introduced the concept of
flow-biased probability to detennine the particle paths which was also used by
Imdakm and Sahimi (1987). The exit channel is chosen randomly but with a
bias towards the paths with greater flow rates. Based on the flow rate through
each path, a probability is assigned to that path. The greater the flow rate,
the greater the probability of the particle choosing that path. For example, in
Figure 7.4.2, if the network is such that 70 % of the flow leaving the node is
through bond A, 20% is through bond B, and 10% is through bond C, then
based on flow-biased probability the probabilities of the paths A, B, and C
being chosen by a particle are 0.7, 0.2, and 0.1 respectively.
Several particles are injected into the network at each time step. The actual
number of particles injected depends on the particle concentration in the
suspension and the time step. The particle sizes are chosen from the specified
distribution. The particles are placed randomly at the nodes on the entry face
with the fraction of particles injected through a particular node is proportional to
the flow rate through that node. During each time step, particles at the entrance
and within the network advance forward with different particles traveling
different distances based on the flow path selected. When a particle encounters
a node, its path out of the node is chosen based on flow-biased probability. If
the bond selected is smaller in size than the particle, the particle is considered
to be captured and the bond is plugged. In this manner, the straining mechanism
of particle capture is incorporated in the model. Particles arriving at bonds that
124 CHAPTER 7
have been plugged in the same time step are also assumed to be captured. After
all particles have been advanced forward by one time step, the flow distribution
in the network is recalculated to take into account the plugged pores. From the
calculations, the reduction in permeability of the network can be determined
and is given by
-Anlet,t
(7.4.5)
Poutlet,o'
for a constant flow rate injection, where K is the permeability of the network
and the subscript '0' refers to the start of the simulation (t = 0). Results can be
plotted as permeability reduction of the network as a function of pore volumes
injected or the number of particles injected (see Figures 7.4.3 and 7.4.4).
In the simulations, particle injection can be carried out until there is no
significant change in the permeability of the network, or the permeability has
essentially vanished due to pore plugging. Alternately, the simulations can be
stopped after the injection of a fixed number of particles, and the network
can be examined to determine the locations of the particle plugging. Due to
the degree of randomness involved in laying down the network, selecting the
particle path at a node, and generating different size particles at the entry
face, the results from different realizations of the network have to be averaged
to get reliable information from network models. Different realizations are
generated by using different initial seeds for the random number generator
(the physical conditions are of course kept identical during the runs with the
different realizations).
7.4.4 EFFECT OF NETWORK SIZE
Rege and Fogler (1987) performed simulations on a triangular network to study

the effect of network size on model predictions. Figure 7.4.3 shows their model
predictions using a lOx 50 network for different realizations. The particles
were assigned a size (Dpa) of one unit (mono disperse) and the bonds were
assigned a normal size distribution with a mean size (Dpo) of one unit and a
standard deviation (O"po) of 0.33 units. The particle concentration was set at
1 x 10 8 particles per unit volume. The model predicts that the permeability
attains a constant value after a sufficient number of pore volumes of the
suspension have been injected. Since the particles are monodisperse, pores
having sizes larger than the particle size will not be plugged and the model
prediction agrees with what one would intuitively expect. Furthermore, because
there is insignificant scatter in the results from the different realizations, it

appears that a single run is sufficient to obtain a representative prediction for
the case of flow of monodisperse particles.
Figure 7.4.4 compares predictions using networks of different size (from a
single realization). One observes that the scatter is not only large but in one
instance, for the lOx 80 network, the model predicts total plugging after only
a few pore volumes. Rege and Fogler (1987) explain that this anomalous result
is related to the percolation probability of a lOx 80 network. Such behavior of
the network models becomes less probable as the network size is increased.
They have used 40 x 40 and 40 x 60 networks in their simulations, which were
first estimates from scaling theories, to determine representative network sizes
in percolation processes. Imdakm and Sahimi (1987) have used 40x40 and
60 x 60 networks in their simulations.
7.4.5 ABILITY OF NETWORK MODELS TO ACCOUNT FOR

POLYDISPERSE PARTICLES AND PORES
Rege and Fogler (1987) studied the effect of mean size of particles by
considering three cases shown in Figure 7.4.5 : low degree of overlap (A),
moderate degree of overlap (B), and high degree of overlap (C) between particle
and pore size distributions. Figure 7.4.6 shows the permeability reduction as a
function of pore volumes injected predicted by the model for the three different
cases. As expected, the number of pore volumes required for complete plugging
decreases significantly as the degree of overlap between the particle and pore
size distributions increase. The model predictions shown are averages from
simulations using four realizations and 15 runs with each realization.
7.4.6 COMPARISON OF MODEL PREDICTIONS WITH

EXPERIMENTS
Rege and Fogler (1987) used the network model to predict the experimental
data of Donaldson et al., (1977) who studied the flow of particle suspensions
through sandstone cores. Log-normal distributions for particle and pore sizes
reported by Donaldson et a1., (1977) were incorporated in the network. As
the network represents only a small cross section of the actual cores used
by Donaldson et a1., (1977), a scaling parameter was used for quantitative
comparison of the model predictions with the experimental data. The scaling
parameter used is the ratio of the core pore volumes to the network pore
volumes. Rege and Fogler (1987) compared their model predictions with data
126 CHAPTER 7
1.0
0.9 Dpo= I °pO:O·33
0.8
Dpa= I
0.7
0.6
KlKO 0.5
0.4
0.3
0.2
0.1
0.0
0 10 20 30 40 50 60 70 80 90 100
Pore Volumes
Figure 7.4.3. Inherent variability in model predictions due to stochastic nature

of the calculations.
Dpo:: 1 O"po =0.33

1.0
0.9 Dpa:: 1
0.6
0.7
0.6 10 x 10
KlKO 0.5
0.4 10 x 100
0.3
0.2 10xSO
0.1 10 x 80
0.0
0 10 20 30 40 50 60 70 80 90 100
Pore Volumes
Figure 7.4.4. Comparison of model predictions for different network sizes.

Medium Overlap
Low Overlap
DpO· I "PO. O.S Dpo -I "po -O.S
Dp •• 0.2 "P•• 0." 0 .• Dpa-O.S "pa-O.4
0.' Z
N(D) 0.' tTl
N(D) -l
:E
0.2
o
0.2
~
0.1 1.' 2.S ~ .
$:
o.s 2.S o
I.' otTl
Diameters Diameters
t""
A. [/)
B. "I'l
o
i'"
High Overlap "I'l
Z
tTl
[/)
o. Dpo-' "po-a.s
0.&
Dpa_O .• "pa-a.4
~
N(D) ~
"0
0.4
$:
tTl
Z
-l
0.1 1.1 2.S
Diameters
C.
Figure 7.4.5. Polydisperse particle size and pore distributions used in simulations.
N
-.)
-
128 CHAPTER 7
tao
0.80
0.60
K/KO
0.40
0.20
c
_--_--_---"1
0.00 +-----"-_ _........
0.00 2.00 4.00 6.00 B.OO 10.00
Pore Volumes
Figure 7.4.6. Effect of polydispersity in particle size on permeability response.
for two different types of sandstones and different particle concentrations.

The scaling parameter was determined by forcing the model predictions to
fit experimental data for one run and was used unchanged in the rest of the
simulations. Figure 7.4.7 shows the comparison between the model and the
experiments and it can be seen that the fit is very good. The comparisons have
been made for experiments using two different sandstones and two different
concentrations for each sandstone.
We note in these figures the steep increase in pressure drop once a certain
level of plugging occurs. This point is the point at which we have catastrophic
closure of porous media. We note that the network model predicts the
experimentally observed catastrophic closure.
7.5 Accounting for particle deposition in network models
Network models have also been used for studying fines transport though porous
media under conditions when both straining and deposition are significant.
To incorporate particle deposition within bonds in the model, rules have to
be developed for determining whether a flowing particle will be deposited
under the existing conditions and whether a previously deposited particle can
be reentrained as flow continues. Other issues requiring consideration are
the effect of the deposited particles on fluid flow within the bonds and how
deposited particles affect the deposition of flowing particles. It is essential to
lOa T 100 J
Run #1 IBerea Sandstone) Run #2 (Berea Sandstone)
C.5.1El0 ~/m3 80
r C=1.04El1 #/m 3
60
60
KO/K
KO/K z
40 ...,m
"1 :;::
'20 t .-f. ::1 0
~
o•
a 200 400 600 800 1000 1200
o e_e='
0
J;
200
,
400 600 800 1000 1200
s:0
Pore Volumes tl
Pore Volumes m
r
C/J
'TI
0
::<l
200 T 'TI
Run #3 (Cleveland Sandstone)· •
C=4.5El0 #/m} 200 I Run .4 (Cleveland Sandstone) Z
C~9.SE 10 .Im '
m
C/J
150 -i • I 150 m
I• ...,Z
KO/K 100 + • / ~
"0
I•
s:m
50 Z
50 -i ...,
•
..~ I
"""1 J o·
400 1000 1200
0----· 0 200 600 800
0 200 400 600 800 1000 1200 Pore Volumes
Pore Volumes
Figure 7.4.7. Comparisons between model predictions of Rege and Fogler (1987) and experiments of Donaldson et aI.,
(1977). (The run numbers listed in the figures refer to their experiments.)
N
\0
-
130 CHAPTER 7
detennine the increase in resistance to fluid flow as particles are deposited

since the goal of the modeling is to predict penneability reduction due to fines
entrapment. Particle deposition on the pore walls is governed by a balance
between hydrodynamic forces, surface colloidal forces, and gravity forces
acting on particles. In the case of continuum models, it is possible to use
rate expressions for particle deposition and release as in the work of Shanna
and Yortsos (1987a), who used continuum population balances in conjunction
with a network representation of the porous medium. They have used rate
expressions for the deposition and release of Brownian particles with and
without energy barriers. In the case of Monte Carlo simulations of particle
movement in a network, it is obvious that such rate expressions cannot be used
since the motion of individual particles is being monitored. In their network
model simulation of deep bed filtration, Rege and Fogler (1988) have used
an expression for the probability of particle capture that contains a lumped
parameter accounting for the various forces mentioned previously. The effect of
fluid velocity was explicitly accounted for in their model. Imdakm and Sahimi
(1991) studied fines capture by straining and deposition using a 3 D cubic
lattice by incorporating particle trajectory calculations in their simulations. In
trajectory calculations, the motion of an individual particle within a bond is
detennined by solving the equations of motion obtained from a force balance
on the particle (Tien, 1989). From the calculated particle trajectory, one can
detennine whether the particle will be retained within the bond or will flow
through without capture. Reentrainment of deposited particles were also taken
into account. Burganos et al. (1992) have developed a detailed network model
for studying deposition of particles. The main features of their model are the
use of bonds having a sinusoidal shape, 3D particle trajectory analysis within
pores, and numerical detennination of the creeping flow conductivity of the
bond. Sinusoidally shaped bonds having a converging-diverging cross section
are a more realistic representation of the pore space in granular beds.
7.6 Improved network models for prediction of permeability reduction
The network model representation of the porous medium described in this

chapter has one drawback in that direct simulation of all macroscopic properties
of a porous medium such as penneability, porosity, surface area, saturation,
etc. is not possible. For example, in the conventional network model made up
of void spaces connected together with cylindrical tubes, the issues of porosity
and network dimension are not well defined because the tube radii and length
alone govern permeability without a clear correlation to porosity. In the case
of modeling fines deposition, the current models allow fines deposition only
in the pore throats and not in the pore bodies. Including deposition in the
pore bodies is important for three reasons: (1) most of the surface area of
a porous medium is in the pore bodies, (2) fluid flow which significantly
affects deposition is entirely different in the pore throats and pore bodies,
and (3) particle deposition in the pore bodies will have a negligible effect on
the permeability. Thus, to be able to accurately predict both the permeability
decline and the effluent suspension concentration, accounting for deposition in
the pore bodies is essential. The physically representative network, PRN model
developed by Bryant et al. (1993) addresses the above issues and has proven
to be the most significant recent advance in network modeling. In this model, a
three-dimensional randomly packed bed of spheres is mapped onto a network
using a Delaunay tesselation. The result is a completely described, disordered
pore space (formed by the voids between the packed spheres) which accurately
represents the converging-diverging character of pores, the non-circular shape
of pore throats, and the contribution of pores to the porosity in natural porous
media. Properties such as porosity and surface area can be summed across the
network to obtain overall values for the packed bed.
Thompson (1995) has developed an algorithm for generating a PRN model
with a wide range of heterogeneity. By rigorously accounting for pore level
fluid mechanics and taking into account the converging diverging nature of the
pores in the pressure drop calculations, Thompson (1995) has demonstrated
that experimental permeabilities can be accurately reproduced in the model.
More significantly, excellent quantitative fits to residence time distributions in
an experimental packed bed were obtained using the PRN model.
The network model is an indispensable tool for modeling pore scale
phenomena in porous media. The use of the PRN model for simulating the
porous medium, together with the well established rules for particle movement
within the network and the determination of fluid flow distribution discussed
in this chapter, will improve the accuracy of model predictions of permeability
reduction due to fines entrapment in porous media.
CHAPTER 8
METHODS TO PREVENT THE RELEASE OF FINES
A number of techniques to prevent the of migration offine particles in sandstone

oil reservoirs have been developed (Valey, 1969; Reed, 1972; Clementz, 1977;
Sydansk, 1984; Borchard and Brown, 1984). Descriptions of these techniques,
commonly known as clay stabilization techniques, are available in patented
literature. This chapter will focus on the underlying mechanisms of these
techniques and attempt to elucidate their influence on the retardation of the
release of fines.
Because these techniques have been developed to combat the water
sensitivity, they are primarily applicable to prevent the release of migratory clay
particles. Some techniques are, nevertheless, versatile enough to be applicable
to other types of migrating fine particles. As noted in Chapter 1, there are a
number of other adverse practical consequences of migration offines. Possible
extension of these techniques to other cases needs to be explored. In order to
achieve this, a sound understanding of the underlying mechanisms is required.
The first step in the migration phenomenon is the release of fine particles
from the pore surfaces. Therefore, an effective method to prevent pore plugging
is to prevent the release offine particles. All industrial techniques are based on
this concept. As discussed in Chapters 3 and 4, the release of fine particles is
a function of the balance between two forces; force of adherence (attachment)
and force of detachment. While the major forces of atttachment are the London-
van der Waals and possibly the AB forces, the major detachment forces are the
electric double-layer repulsion and the hydrodynamic forces. The release can
be then prevented either by enhancing the force of adherence or by reducing
the force of detachment or a combination of these two methods.
133
134 CHAPTER 8
1.0
0.9
0.8
0.7
0.6 °
0.5
KlK o
0.4
0.3
0.2
0.1
0 200 400 600 800

TEMPERATURE, °c
Figure 8.1.1. Effect of temperature on permeability reduction for Berea

sandstones (Khilar and Fogler, 1987).
8.1 Enhancement of Force of Adherence/Attachment
Methods to strongly attach the fine particles to pore surface can be put into three
categories based on their attachment mechanism: physical, physicochemical
and chemical attachments.
In a physical method of attachment, Berea sandstone cores were fired in an
oven to high temperatures resulting in the clay fine particles being fused to the
quartz grains of the pore surface; consequently, they could not be released in
water shock experiments (Sydansk, 1980). Figure 8.1.1 shows the reduction
in permeability obtained from standard water shock experiments of Berea
sandstone cores fired at different temperatures for 24 hours (Khilar and Fogler,
1987).
We observe from this figure that the permeability of sandstone cores fired
at temperatures higher than 500°C, remains unchanged after a water shock
experiment. Partial reductions in permeability occur however,for cores fired to
a temperature between 200 and 500°C, indicating that not all fine particles get
fused at these temperatures. Considering the fact that fine particles are fused
to the surface, the release of clay particles due to hydrodynamic forces cannot
occur. It is, however, only useful in laboratory scale experiments and can not
be applied at the field scale.
As an example of a physicochemical method, the release of fine particles
METHODS TO PREVENT RELEASE OF FINES 135
1.6
• CONTROL
1.4 • SOA BOTTOMS - TRE ...TED
1.2
1.0
O.B
0.6
w 0.'
~ 0.2
8
15
~, 1.6
~ 1,4
'"~ 1.2
1.0
0.8
0.6
0.'
0.2
oAW~~~ __L-~-L~__L-~~
a 10 20 30 4G 5U &II 10 80 90 100
FLOW R~TE. tclmin/cm 2
Figure 8.1.2. Stabilization of a friable sandstone through adsorption of

petroleum heavy ends from a benzene solution (Clemetz, 1977).
can be reduced by adsorbing petroleum heavy ends (resins and asphahenes )

on the clay-pore surface (Clementz, 1977). This technique has been applied to
minimize the water sensitivity problem in Berea sandstone, as well as to sand
control for formation sands where clays are the primary cementing materials.
Figure 8.1.2 shows sand production as a function offlowrate for untreated and
treated samples of a California sand. A small amount of fines are released for
the treated sample.
While the exact mechanism is not yet known, it is believed that the adsorption
of petroleum heavy ends (resin, asphaltene etc.) on the clay surface makes the
surface oil-wet. This technique appears to be potentially useful in controlling
136 CHAPTER 8
soil erosion by stabilizing the clayey fines in compacted soil mass. However,
research needs to be carried out to learn whether this technique can control the
migration of other kinds of fines.
Most chemical methods essentially use acidic or alkaline solutions to
chemically bind the clay fine particles to the pore surface. These methods
have been used in the fields satisfactorily. For example aqueous solutions of
fluoboric acid, H B F4 have been used successfully to stabilize the clayey fines
(Thomas and Crowe, 1981). Fluoboric acid hydrolyzes to form hydrofluoric
acid, which then reacts with clays and other siliceous fines. In addition to
dissolution, the hydrolysis products of H BF4 extracts aluminum and replaces
it with boron to form a borosilicate. Formation of borosilicate as well as the
possible precipitation of silica species may be the cause of clay stabilization.
Figure 8.1.3 shows the scanning electron micrographs of kaolinite platelets
before and after fluoboric acid treatement. We observe that the kaolinite
platelets appear to have fused to the matrix after the treatment. It is also shown
that this treatment reduces the cation exchange capacity of clays, thereby
reducing their dispersion capabilities.
In another method, Sydansk (1984) used 15 percent by weight of aqueous
potassium hydroxide KOH solution to stabilize clays. Potassium hydroxide
solution was sent through Berea sandstone cores at low flowrates for a period
of 16 hours at 85°C. A kinetically slow chemical reaction is believed to occur
between KOH, fines, and the pore surface. As a result, the fines are stabilized by
forming chemical bonds with the pore surface (Sydansk, 1984). This techniques
has been found to be successful in field tests (Sloat, 1990).
However, for smectitic clay containing sandstones, it has been shown that
if a K+ treatment is followed by flow of a Na+ solution such as in formation
water, damage can result (Mohan, 1996b).
The chemical methods to strongly bind the fine particles to pore surface
have potential applications in other areas of migration of fines. The challenge,
however, lies in developing inexpensive chemicals that can react with the pore
materials and form chemical bonds between the fines and the pore surfaces.
Research in this direction is required particularly to problems associated with
migration of fines in environmental engineering.
8.2 Reduction of Force of Detachment
We have seen in Chapters 3 and 4 that the primary force of detachment in

colloidally induced fines migration is the double-layer force while that in
hydrodynamically induced migration is the shear force. Techniques have been
Figure 8.1.3. SEM photographs of kaolinite platelets (a) before and (b) after
the fluoboric acid treatment (Thomas and Crowe, 1981).
138 CHAPTER 8
developed to minimize the magnitude of these forces during field operations

involving flow through porous media.
Reduction in the double-layer force of repulsion can be achieved by reducing
the zeta potentials of the fines and the pore surface as well as reducing the extent
of overlap of double-layers. It has been shown that the release offine particles
can be prevented in Berea and other sandstones by adding simple monovalent
(NaCl, KCl) and/or divalent salts (CaCh) to the permeating solution to raise
the salt concentration above the critical salt concentration, CSC (Jones, 1964;
Khilar and Fogler, 1984; Lauzon, 1982; Smith et aI., 1964). We have seen
in Chapter 3 that when the salt concentration is above CSC, the double-layer
force repulsion is not strong enough to cause detachment of the particles. In
addition, it has been found that the fine particles are not released in Berea
sandstone when low pH solution (pH <2.5) is sent through a Berea sandstone
core in a water shock experiment (Kia et aI., 1987a). It is shown that at low pH,
the zeta potential of the fine particles is small, resulting in small double-layer
repulsion (Kia et aI., 1987a).
Divalent salt systems are more effective than monovalent cations in pre-
venting the release of clay fine particles. (Jones, 1964; Kia et aI., 1987b;
Lauzon, 1982; Khilar and Sarma, 1990). Here the divalent cations such as
Ca2+ adsorb on the clay and pore surfaces leading to a reduction in zeta po-
tential. It has also been shown that, for mixed salt systems, the ionic strength
should be above a critical total ionic strength, CTIS, which is analgous to the
CSC for monovalent ions, (Khilar and Sarma, 1990). Recently, Zaltoun and
Berton (1992) found that by adding high molecular weight polymers (anionic
and/or non-ionic) such as polyacrylamide to brine solution, one can stabi-
lize the swelling clays such as montmorillonite. The CSC for swelling clays
decreases in presence of this polymer.
Unlike the case of double-layer force, the hydrodynamic force can be re-
duced by simply maintaining a lower flow velocity. In Chapter 4 we showed
that a critical stress or a critical velocity exists below which the fine particles are
not released. Therefore, maintaining a flow velocity below the critical velocity
will ensure the prevention of hydrodynamically induced fines migrations.
All the above techniques are useful to prevent migration of fines in porous
media. These techniques will fail, however, if water from underground sources
such as an aquifer, invades the porous formation and interacts with the migra-
tory fine particles.
8.3 Minimization of Fines Release: Enhancing the Attachment Forces

and Reducing the Detachment Forces
This double-action technique is very effective and has been widely used com-
mercially. The mechanism uses strongly adsorbing polycations on to the clay
surface. The resulting charge reversal causes strong electrostatic attraction be-
tween the pore wall and the particles. Often polycations fonn links between
the pore surface, and the fine particles as well as between the particles on the
surface, thereby strengthening the adherence of the particles to the pore sur-
face. In addition, the adsorption of the polycations reduces the zeta potential of
the fine particles and, therefore, decreases the double-layer force of repulsion.
A number of patented techniques have been developed based on this mech-
anism. Valey (1969) showed that hydrolyzable metal ions can stabilize clay
fine particles. The polyvalent metals such as aluminum and zirconium can un-
dergo hydration followed by hydrolysis to fonn hydrated ions. These hydrated
ions then polymerize to fonn polynuclear ions to valency as high as 24. The
chemistry of the fonnation of polynuclear metal ions is briefly described by
Valey where polyvalent metals such as iron, chromium, lanthanum, hafnium,
and titanium have been used (Valey, 1969). In addition to the strong attraction
between the pore surface and the fine particles, chemical bonds between the
particles and the pore surface are believed to fonn (Valey, 1969). Figure 8.3.1
shows that the penneability of Berea sandstone treated with a hydrolyzable
metal salt solution does not get damaged (i.e. rapid and drastic reduction in
penneability) during the water shock ofthe sandstones.
Based on Valey's findings, Reed (1972) developed hydroxyl-aluminum so-
lution to combat the fines migration problem. It is shown that the disper-
sion and/or swelling of clays can be prevented by injecting a O.IM hydroxy
alminum solution having OH/AI ratio of 2.0 (Reed, 1972). In this case, the
polynuclear metal ions, AI6(OH){26 strongly adsorb to the clay particles.
There have been several investigations both in the laboratory and the field
to evaluate the use of hyrolyzable metal ion solution to stabilize the fonna-
tion clays (Coulter et ai.,1979; Coppel et ai., 1973; Peters and Stout, 1997;
Hesterberg and Reed, 1991). The results reported show effective clay stabi-
lization.
The use of water soluble cationic organic polymers is another approach
developed to prevent the release of clay particles (McLaughlin et ai., 1977;
Young et ai., 1980; Borchard and Brown, 1984; Himes et ai., 1991). These
additives commonly known as COP, are composed of long chain organic
polymers of molecular weights as low as 5000 to well over 1,000,000. In
140 CHAPTERS
lOa r---r-,---,--,---,r---r-r--,--,---,
~outow 0,1.fO'Cl~u' n
no I--+--I--+--i----::~•.:::',.r_'.:.:(O::.:."'Or":.:.'_+- '-+-.,
--- -,-- __._J ___ , -- --- ---~---"1.
. ! ,. ! I ,"-------.
.I.
100 1---+-!-+-+--++---I1f-~'--t--++-+--;
eo f--+H--:-!-". :
! I !_~i~H--+~i__;+-~~
....
_E .'J =: I 0
- 60L--I-r-~' g-#' ~H--+r- -;+-~-1
i 1:5
i ~
-+ __
- . 2 . -
r
I . ~ S ! ~ ! ~
40 I ~--+- ~-+): -~ rT-;: SJ ~
~i:'!~ig
.. 0 I C ; 1;
c, ,_ . A
~ ! ::
10 t-S--!-=;.....;.... "' -i.
• , .. : ~ I
-=-~
~ 1 1.c
'-+-=
I -:::
'--
o~! .I!~ ! ti O!I! 0
o I 1 , I 6 I '0
HUN DREDS Of PORE VOLUMES flU ID
Figure 8.3.1. Typical behavior of water sensitive sandstones, particularly

Berea, when treated with hydrolyzable metal solutions (Valey, 1969).
general, they are hydrophobic and carry positive charge. The details of the
specific molecules used as clay stabilizers are not available in open literature.
Although the exact mechanism of clay stabilization by COP has not been
reported, it is believed to be similar to the charge reversal mechanism of the
hydrolyzable metal salt solutions.
These techniques (hydrolyzable metalloids and COP) have been primarily
applied to clay fines in sandstone formations. Because of the inherent similar-
ities, it appears reasonable to assume that these techniques will work in cases
of other migratory fines such as organic and biocolloidal fines in soil masses,
and in other porous media.
CHAPTER 9
SOIL POLLUTION DUE TO MIGRATION OF FINES
Many toxic chemicals adhere to fine particles; consequently the migration of

fines containing contaminants can spread through our environment. There are
two primary mechanisms by which the pollution can spread and grow:
1. the mobilization of fine particles carrying contaminants sorbed on their
surface and thereby facilitating the transport (McCarthy and Zachara,
1989), and
2. the migration of biocolloidal fines, such as bacteria and viruses, in soil
(McDowel-Boyer et aI., 1986; Gerba et aI., 1975; Harvey et aI., 1989;
Scholl and Harvey 1992).
In general, the soil and the aquifer environment can be subdivided into
four zones: shallow surface soil (top soil), subsurface soil (vadose), aquifer
and aquitard (McCarthy and Zachara, 1989). The top soil usually includes
the upper 0.25 to 1 m layer of the soil environment, while the depth of the
subsurface soil may run up to a few meters. If the water table were to rise,
water fills in the void space in the surface and subsurface soil and comes in
contact with the fines present; as a result from this water intrusion, colloidal
and biocolloidal fin'es are released and migrate with the flow, spreading the
contaminants.
We shall first discuss the facilitated contaminant transport where fines act
as carriers and then shall discuss the migration ofbiocolloidal fines.
9.1 Facilitated Contaminants Transport due to Migration of Fines
Recently, it has been shown in a number of studies, that the transport of

low solubility contaminants in soil is facilitated by migration of colloidal
fines (McCarthy and Zachara, 1989; Eichholz et aI., 1982; Penrose et aI.,
1990; Magaritz et aI., 1990; Torok et aI., 1990; PuIs and Powell 1992;
Ryan and Elimelech, 1996). The fines, are in general, clay minerals, oxides,
carbonates, as well as organic colloids and humic substances. The contaminants
141
142 CHAPTER 9
include metals, radionuclides, and hydrophobic organic contaminants (HOC),

like polychlorinated biphenols and pesticides. These contaminants adsorb or
partition on the surface of the fine particles and are carried as the fines migrate.
The following three processes must occur to facilitate the contaminant
transport (Ryan and Elimelech, 1996):
1. The generation of colloidal fines,
2. The association (sorption) of contaminants with the colloidal fines,
3. The transport of colloidal fines.
Colloidal fines are generated primarily due to their release from the pore
surfaces. In some cases, however, colloidal size precipitates can fonn, resulting
from changes in the geochemical conditions. The generation/release as well as
the migration aspects are presented in Chapters 3 to 7 and consequently will
not be repeated here.
9.1.1 SORPTION OF HYDROPHOBIC CONTAMINANTS ON FINES IN

SOIL MASSES
The sorption of hydrophobic compounds of very low water solubility (ppb

or ppm) by soil particulates has been studied extensively due to its profound
importance in soil and ground water pollution. Voice and Weber (1983) present
a discussion on the theoretical background of this sorption process in the light
of solute-solid and solute-solvent interactions.
In general, the contaminants are present in the aqueous phase at a
concentration much below their solubility. In this very low concentration range,
it has been shown that the sorption isothenn can be approximated as a simple
linear relationship. Figures 9.1.1 and 9.1.2 show some typical isothenns. We
observe from these figures that these isothenns are by and large linear though
some systems show non-linear isothenns.
Based on these observations, a simple proportional relationship has been
proposed (Karickhoff et aI., 1979);
(9.1.1 )
Here X is the concentration of sorbate on the solid, based on dry weight,

(ppb/ ppm); Ce is the equilibrium sorbate concentration in solution (ppb/ppm),
and Kp is the sorption/partition coefficient. The sorption coefficient is found
to depend on (1) organic carbon content of the solid fine particles, (2)
concentration of the fine particles, (3) the size distribution of the fine
SOIL POLLUTION DUE TO MIGRATION OF FINES 143
-50
&
0-
; : ~5 .. Cloy Concentrotion~12mg/ml
z
w
::E~
(5
w
VI 35
;?;
~~
g:~ 25
Z ...
W
~ 20 --Kp=1100
o
u
~ 15
:3
G 10
>-
X
o
:I:
t-
W
:::E
5 ~ • W ~ ~ ~
METHOXYCHLOR IN WATER (ppb)
Figure 9.1.1. Adsorption isotherm for methoxychlor on Hickory hill clay.

(Klrickhoff et al., 1979).
•
•
2 6 7 e 9 1l110, 2 3 4 ~ 6 7 6 ~Xl02
Equilibrium solufion coneemrallon I Ilg I-I)
Figure 9.1.2. Sorption of Aroclor 1254 by three clays. (Weber et al., 1983).
144 CHAPTER 9
particles and (4) the octanollwater distribution coefficient of the contaminants

(Karickhoff et aI., 1979; O'Connor and Connolly, 1980; Voice and Weber, 1983;
Voice et aI., 1983). This sorption process, in general, is reversible and has
negligible hysterisis effects.
The typical values of the partition coefficient, Kp for clay fines range from
103 to 104 . This value can be increased by orders of magnitude when the
clay surface is modified by polyvalent inorganic ions and cationic surfactants
(Srinivasan and Fogler 1990a; 1990b) and can be of order of 106 for suspended
organic colloids and biocolloids. While the linear isotherm model works for
most sorption systems of hydrophobic contaminants on soil, other models such
as Freundlich and BET may better describe the data in some cases. The sorption
of metals and radionuclides on fines can be modeled in analogous manner to
that of organic pollutants.
9.1.2 MODELING OF TRANSPORT OF CONTAMINANTS

FACILITATED BY MIGRATION OF FINE PARTICLES
The fine particles carrying sorbed contaminants migrate over long distances.
This migration is related to the piping phenomenon discussed in Chapter 5.
However, the process is actually more complex due to the presence of the
contaminants and their interactions with the solid and aqueous phases.
Corapcioglu and Jiang (1993) have proposed a mathematical model to
describe the facilitated contaminant transport. Using the general framework
of the release and capture model for migration of fines, they have considered
contaminants present in the phase, sorbed on the fines in suspension and at
the captures sites as well as sorbed on the solid matrix of the soil mass. Mass
balance equations need to be written both for the fine particles as well as for
the contaminants. Since the contaminant species reside in four different sites
(mobile fines, captured fines, liquid, and solid phases), mass balance equations
must be written for each site. The other assumptions and considerations are
similar to those presented in sections 5.2 and 5.3. For the sake of consistency
we shall continue using the notations we have already introduced in this book.
Additional notations required to write the model equations are given below:
C concentration of particles in solution, kgm- 3 ;

Cc = concentration of contaminant species in the aqueous
phase, kgm- 3 ;
XI mass fraction of the contaminant species adhered to fine
particles in suspension;
X2 = mass fraction of contaminant species adhered to the captured
fine particles;
X3 = mass fraction ofthe contaminant species adhered to the solid
matrix, based on dry mass of solid;
Ka = rate constant for sorption on the fines, S-I;
Kd = rate constant for desorption from the fines, s -I ;
Ke = sorption/partition coefficient of the contaminant species on
the solid matrix, kgm- 3 .
Pb = bulk density of solid matrix, kgm- 3 ;
DB Brownian diffusivity of fine particles, m 2 S-I;
Dc longitudinal dispersion coefficient of the contaminant species, m 2 S-I
x longitudinal distance
Tc = rate of capture of particles, kgm- 3 S-I
Tr rate of release of particles, kgm- 3 S-I
Unsteady mass balance for fine particles in suspension is given as
(9.1.2)
Note this equation has one additional term (the first term on rhs) compared to
Equation (6.4.3). This term resulting from Brownian motion can be neglected
if fine particles are of a size greater than 1 /Lm.
Unsteady mass balance for the contaminant on the fine particles in
suspension can be written as
(9.1.3)
In Equation (9.1.3), the only term on lhs represents the rate of accumulation
of contaminant on the fines in suspension. The first two terms on the rhs
146 CHAPTER 9
represent the flux of contaminants due to the Brownian and convective motion
of the fines. The third and fourth terms on the rhs represent the flux of
contaminant resulting from the release and capture offines. In Equation (9.1.3),
the last two terms on the rhs represent the contributions of adsorption and
desorption of contaminants respectively. Note, bl represents the fraction of
total adsorption that takes place on the fines in suspension.
Unsteady mass balance for the contaminant species on the captured fines
can be given as
(9.l.4)
In the Equation (9.1.4), the term on the lhs represents the rate of accumulation
of contaminant species on the captures fines and the first term on the rhs
represents the influx of contaminant species due to capture of fines. The last
two terms represent the rate ofthe adsorption and desorption of the contaminant
species respectively as regard to captured fines. Note, b2 represents fraction of
total adsorption that occurs on the captured fines.
Finally an unsteady mass balance for the contaminant species in the liquid
phase and on the solid matrix together yields
(9.l.5)
In Equation (9.l.5), the first term, on the lhs represents the rate of
accumulation of contaminant species on the solid matrix while the second
term represents that in the liquid phase. The terms on the rhs are similar to
the other equations excluding the first term which represents the flux of the
contaminant species due to hydrodynamic dispersion.
Furthermore, by assuming the equilibrium partition between the liquid phase
and solid matrix, we obtain
(9.l.6)
.... K• • 0.01)8/.
...uo ;!.D - •• K• • 0.001/11
.
• ••• k. • 0.000/.
.....
~
~c
.,• 'D
Z
IJ
~ O, ~
~
Oista nce X (em }
Figure 9.1.3. Total mobile contaminant concentration at different capture

coefficients (Corapcioglu and Jiang, 1993) .
0
2.0
"u
"-
u
V
.,
~
0
1 .~
!l•
·c 1.0
0
c
.·
0
u
___ K, = o.ooe;.
> o.~
::l •••• K, " 0 .006/5
d
._ .... K, = 0.003/ •
0:
- 1'I'ithouL colloid
0.0 -+-',,",-....,..--r--,.....--r-...,...--r--.--.....,.--1
o ~o 40 SO 80 100 120 uo !60 180 200
Pore Vo l ume V/Vo
Figure 9.1.4. Spatial variations of total mobile contaminant concentration at

different capture coefficients (Corapcioglu and Jiang, 1993) .
148 CHAPTER 9
In addition to Equations (9.1.2) to (9.1.6), we need the rate equations

for release and capture of fine particles, which are discussed in Chapter 6.
Furthermore, if the permeability changes due to the migration offine particles,
then the resulting flowrate variation must also to be accounted for and is
discussed in Chapters 5 and 6.
Corapcioglu and Jiang (1993) solved this set of equations numerically for a
system of finite length and where the fine particles, contaminant species and
fluid are fed at a constant rate to the system. They carried out simulations
describing the variations in total mobile concentration of contaminant as a
function of time and axial length for various values of the system parameters
used in their study. The total mobile concentration Get is the sum of the
concentration in the liquid phase, Gc and that on the fines in suspension, ()2X2.
Figures 9.1.3 and 9.1.4 show some of their simulation results. We observe from
these figures that, the total mobile concentration, in general, is higher when
compared with the concentration without migration of fines.
The simulations showed that higher rate of sorption of contaminant species
as well as higher rate of release significantly facilitate the transport of
contaminants. On the other hand, higher rates of desorption and capture of
fine particles, weakly facilitate the transport.
An equilibrium approach may be adopted in appropriate situations where
the sorption of the contaminant species on the fine particles is at equilibrium
(i.e. when mass transfer limitations are negligible). The mass fractions, Xl and
X2 become equal and are related to the concentration in the aqueous phase.
The solution has been obtained for this equilibrium case (Magee et ai., 1991).
The model can be modified to incorporate for the variations in flowrate due to
piping/entrapment and it can also be extended to a case where more than one
polluting species is involved.
9.2 Migration of Biocolloidal Contaminants
Biocolloidal contaminants are the micro-organisms such as bacteria, fungi

or viruses that are of colloidal size. Knowledge of the migration of these
biocolloidal fines is important in assessing the spreading of certain diseases as
well as of soil and groundwater pollution. This is also required for the design
of bioremediation techniques.
9.2.1 CHARACTERISTICS OF BIOCOLLIDAL CONTAMINANTS
In addition to the colloidal characteristics of size and charge, bio-characteristics

relevant to their migration are the metabolic growth (birth) and decay (death).
Bacteria growth rate has been shown to be important in determining how far
the bacteria migrate, and therefore it is a factor in migration of biocolloidal
fines (Harvey and Garabedian, 1991).
The growth rate of the bacteria is given by a commonly used expression
know as Monod equation (Fogler, 1992)
Cx S
Tg=Jhm Ks +S' (9.2.1)
where Jhm = a maximum specific growth reaction rate, S-I; CX = bacteria/cell

concentration, kgm- 3 ; S = substrate concentration, kgm- 3 ; Ks = parameter
analogous to Michaelis constant, kg m- 3 .
The decay or death of bacteria is given as
(9.2.2)
where C t = concentration of substance toxic to the bacteria, kgm- 3 ; kd =

natural death rate constant, s -I; C t = substrate concentration, kg m - 3 ; k t =
death rate constant due to toxic substance, m 3 kg- I S-I.
In addition to substrate utilization for producing new cells, a part of the
substrate is used to maintain the bacteria population, and the rate is written as
proportional to the concentration of the bacteria, C x (Fogler, 1992).
9.2.2 MODELING THE MIGRATION OF BIOCOLLOIDAL FINES
The mathematical modeling to describe the migration of biocolloida1

fines is essentially based on the similar framework as that of colloidal
fines discussed earlier (McDowel-Boyer et aI., 1986; Bales et aI., 1991;
Harvey and Garabedian, 1991; Lindqvist et aI., 1994; Tan et aI., 1994;
Lappan and Fogler, 1996). In addition, some modifications to account for
a number of mechanisms specific to biocolloids have been incorporated. An
unsteady mass or number balance for the active bacteria/cell in suspension
can be written with the major assumptions and considerations as outlined in
150 CHAPTER 9
section 5.2, 6.1, and 9.1 as
(9.2.3)
In the Equation (9.2.3), the terms are similar to the terms in the balance
equation for fines in suspension. The last two terms on rhs represent the rate of
generation and decay of bacteria as discussed in the preceding section. It can
be mentioned here that Harvey and Garabedian (1991) added two additional
terms representing the reversible adsorption and desorption of bacteria. They,
however, did not include the release term in the balance equation and included
an irreversible adsorption term in place of the capture term. The convective flux
resulting from sedimentation and other sources have been generally neglected,
and therefore Vt can be taken as equal to the flow velocity, v. Furthermore,
the growth and decay terms have also been neglected for low-residence-time
experiments in laboratory. These terms, however, may be non-negligible in
large scale fields. Equation (9.2.3) with the above simplifications has been
solved by several investigators under a variety of conditions (Harvey and
Garabedian, 1991; Lindqvist et al., 1994; Lappan and Fogler, 1996). We shall
discuss the modeling study of Lappan and Fogler (1994; 1996) who considered
both release and capture ofbiocolloidal fines.
Other investigators have neglected the release term. The rate of release or
entrainment of bacteria is given as
Tr = 0: (a2x - a2XO) , a2X > a2XO,

(9.2.4)
= 0, a2X < a2XO·
Likewise the capture/deposition of bacteria is given as
Tc = j3Cx· (9.2.5)
Lappan and Fogler solved these Equations (9.2.2) to (9.2.5) along with
the mass balance equation for deposited/capture bacteria with appropriate
boundary and initial conditions to obtain Cx as a function time and axial
length. The following parameters values are assumed in their calculations:
a2XO = 5.9 x 10 9 dm- 3 ; 0: = 1.3 X 10- 2 and j3 = 8.9 X 10- 3 min-I.
Further they assumed the dispersion coefficient to be proportional to the
flow velocity with proportionality constant of 2.3 mm. The inlet bacteria/cell
10.0...----------------,
7.5
5.0
2.5
100 200 300 4(KI 500 600 700

Time (min)
Figure 9.2.1. Inoculum model effluent planktonic cell concentration (Lappan

ana Fogler, 1996).
concentration was maintained at 8.6 x lOll dm- 3 for one hour. Figure 9.2.1
shows the comparison between their data and the model predictions. We
observe that model predictions agree very well with the data, implying that the
migration ofbiocolloidal fines can be represented by their model.
Some investigators also have solved the model equations without the release
term, and their model predictions were compared with data (Harvey and
Garabedian, 1991; Tan et aI., 1994; Lindqvist et aI., 1994). Their experimental
systems were typically sand packs oflength varying from a few cm to a few m,
consisting of sand particles of size of the order of 100 pm. The flow velocity
was of the order 10- 1 mms- 1 •
The bacteria suspended in distilled water or in NaCI solution were injected
either in the form of pulse or for a short interval of time. The only variable
measured was the concentration of bacteria at the outlet. The flowrate and
pressure drop were not measured in these experiments. Figures 9.2.2 and
9.2.3 show the typical comparison between the measurements and the model
calculations. We observe from these figures that the agreement is reasonably
good.
The major experimental findings reported can be summarized as follows:
1. Bacteria could move through the sandpack readily. The sandpack retained
the bacteria depending on its retention capacity.
2. The breakthrough was generally retarded due to the adsorption and
straining of bacteria.
152 CHAPTER 9
1.2
'0
1.0
~
~
=
.9
'la
0.8
b
rJ 0.6
•
u
U
t::
0 •
~
0.4
• •
•
.~
"0 0.2
~
0.0
••
0.0 0.5 1.0 1.5 2.0 2.5
Time (hours)
Figure 9.2.2. Breakthrough plot for a cell concentration of 107 ml- l and a
flow velocity of 0.20 mm.s- l (Tan et aI., 1994).
g
'"'b
1.2
1.0
...
•
=
0
-.::I
0.8
• •
~
5 0.6
§
U 0.4
.,
Q
.~
'la
0.2
~
0.0
0.0 0.5 1.0 1.5 2.0 2.5
Time (hours)
Figure 9.2.3. Breakthrough plot for a cell concentration of 108 ml- l and a
flow velocity of 0.20 mm.s- l (Tan et at, 1994).
3. The transport of bacteria was affected by the flow velocity, bacteria

concentration and type, and the ionic concentration of the suspending
liquid.
The modeling of migration of bacteria in porous media is a topic of current
research. Aspects such as bacteria-pore surface interactions, permeability
change due to retention of bacteria, and the effects of birth and death of
bacteria on migration need to be studied.
REFERENCES
Adler, P. M., Porous Media: Geometry and Transports, Butterworth-

Heinemann, Boston, Ch. 2, 1992.
Aitchinson, G.D., and C.C. Wood, Some interactions of compaction,
permeability and post-construction deflocculation affecting probability of
piping failure in small earthen dams, 6th Inti. Conf. on Soil Mechanics and
Foundation Engr., Montreal, Canada, Vol. 2, 442-446, 1965.
Anderson, W.G., Wettability literature survey - Part I: Rock-Oil-Brine
interactions and the effects of core handling on wettability, J. Pet. Technol.,
October, 1125-1144, 1986a.
Anderson, W.G., Wettability literature survey - Part 2: Wettability
measurements,J. Pet. Technol., November, 1246-1262, 1986b.
Arulanadan K, P. Loganathan, and R.B. Krone, Pore and eroding fluid
influences on surface erosion of soil, J. Geotech. Engin. Div., ASCE, 101 (1),
51--65, 1975.
Arulanandan K, E. Gillogley, and R. Tulley, Development of quantitative
methods to predict critical shear and rate of erosion of natural undisturbed
cohesive soils, Technical Report GL-80-5, USCOE Waterways, Exp. Stn.,
Vickshurg, Miss. 1980.
Arulanandan K, and B.B. Perry, Erosion in relation to filter design criteria in
earth dams, J. Geotech. Engin. Div., ASCE, 109(5),682--698, 1983.
Bales, R.C., S.R. Hinkle, T.W. Kroeger, K Stocking, and C.P. Gerba,
Bacteriophage adsorption during transport through porous media: Chemical
perturbations and reversibility, Environ. Sci. Technol., 25, 2088-2095,
1991.
Barenblatt, G.I., V.M. Entov, and V.M. Ryzhik, Theory ofFluid Flows through
Natural Rocks, Kluwer Academic Publishers, Boston, Chapter 4, 1990.
Barouch, E., T.H. Wright, and E. Matijevic, Kinetics of particle detachment 1.
General considerations, J. Colloid Interface Sci., 118(2), 473-481, 1987.
Bear, J., and V. Bachmat, Introduction to Modelling of Transport in Porous
Media, Kluwer Academic Publishers, Boston, Chapter 4, 1990.
155
156 REFERENCES
Bhuniya S.K., Flow of Suspension through Packed Beads: Studies on the

Influences of the Particle Concentration, M. Tech. Thesis, lIT Bombay,
1996.
Biswas, T.D., and S.K. Mukherjee, Text Book of Soil Science, Tata McGraw
Hill, New Delhi, Chapter 2, 1994a.
Biswas, T.D., and S.K. Mukherjee, Text Book of Soil Science, Tata McGraw
Hill, New Delhi, Chapter 4, 1994b.
Borchardt, J.K., and D.L., Brown, Clay stabilizers improve injection rates, Oil
Gas J., Sept., 150-156, 1984.
Bryant, S.L., D.W. Mellor, and c.A. Cade, Physically representative network
models of transport in porous media, AIChE J., 39(3), 387-396, 1993.
Burganos, V.N., C.A. Paraskeva, and A.C. Payatakes, Three dimensional
trajectory analysis and network simulation of deep bed filtration, J. Colloid
Interface SeL, 148, 167-181, 1992.
Cerda, C.M., Mobilization of kaolinite fines in porous media, Colloids
Surfaces, 27, 219-241, 1987.
Cerda, C.M., Mobilization of quartz fines in porous media, Clays Clay
Minerals, 36,491--497,1988.
Chedda, P., D. Grasso, and C.J. van ass, Impact of ozone on stability of
montmorillonite suspensions, J. Colloid Interface Sci., 153(1), 226-236,
1992.
Christenson, H.K., Non-DLVO forces between surfaces - solvation, hydration
and capillary effects, J. Dispersion Sci. Techno!., 9(2) 171-206, 1988.
Cleaver, J.W., and B. Yates, Mechanism of detachment of colloidal particles
from a flat substrate in a turbulent flow, J. Colloid Interface SeL, 44(3),
464--474, 1973.
Clementz, D.M., Clay stabilizattion of sandstones through adsorption of
petroleum heavy ends, J. Pet. Technol., 21, 1061-1066, 1977.
Collins, R.E., Flow of Fluids through Porous Materials, Reinhold Publishing
Corporation, New York, Chapter 1, 1960.
Collins,R.E., Flow of Fluids through Porous Materials, Reinhold Publishing
Corporation, New York, Chapter 1, 1990.
Coppe1, c.P., H.Y. Jennings, and M.G. Reed, Field results from wells treated
with hydroxy-aluminum, J. Pet. Technol., Sept., 1108-1112, 1973.
Corapciog1u, M.Y., and S. Jiang, Colloid-facilitated groundwater contaminant
transport, Water Resour. Res., 29(7), 2215-2226, 1993.
Coulter, A.W., C.T., Copeland, and W.H. Harrisberger, A Laboratory study of
clay stabilizers, Soc. Pet. Engin. J., Oct., 267-270, 1979.
REFERENCES 157
Craig, F.F., The Reservoir Engineering Aspects of Water-Flooding, Monograph

Series, SPE, Richardson TX, 3, 1971.
Dahneke, B., Kinetic theory of escape of particles from surface, J. Colloid
Interface Sci., 50(1), 89-107, 1975a.
Dahneke, B., Resuspension of particles, J. Colloid Interface Sci., 50(1), 194-
196, 1975b.
Das, S.K., R.S. Schechter, and M.M. Sharma, The role of surface roughness
and contact deformation on the hydrodynamic detachment of particles from
surface, J. Colloid Interface Sci., 164,63-77, 1994.
Derjaguin B.V., V.M. Muller, and VP. Toporov, Effect of contact deformation
on the adhesion of particle, J. Colloid Interface Sci., 53, 314-326,1975.
Dias, M.M., and A. C. Payatakes, Network models for two-phase flow itt porous
media: Part 1. Immiscible micro-displacement of non-wetting fluids, J.
Fluid Mech., 164,305-336,1986.
Donaldson, E.C., B.A. Baker, and H.B. Carrol, Particle transport in
sandstones, SPE 6905, 52nd Annual Technical Conference and Exhibition
of SPE, 1977.
Dullien, F.A.L., and G.K. Dhawan, Characterization of pore structure by a
combination of quantitative photomicrography and mercury porosimetry,
J. Colloid Interface Sci., 47(2),337-349, 1974.
Dullien, F.A.L., Porous Media, Fluid Transport and Pore Structure, Academic
Press, New York, Chapter 3, 1979a.
Dullien, F.A.L., Porous Media, Fluid Transport and Pore Structure, Academic
Press, New York, Chapter 2, 1979b.
Eichholz, G.G., B.G. Wahlig, G.F. Powell, and T.F. Craft, Subsurface migration
of radioactive waste materials by particulate transport, Nuclear Technol.,
58,511-520, 1982.
Flair, G.M., J.e. Geyer, and D.A. Okun, Water and Wastewater Engineering,
Vol. 2, Water Purification and Wastewater Treatment and Disposal, John
Wiley and Sons, New York, Chapter 27,1968.
Fair, M.C., and J.L. Anderson, Electrophoresis of non-uniformly charged
ellipsoidal particles, J. Colloid Interface Sci., 127(2),388-400, 1989.
Fatt, I., The network model of porous media, Pet. Trans. AIME, 207,141-188,
1956.
Feke, D.L., N.D. Prabhu, J.A. Mann Jr., and J.A. Mann III, A formulation
of the short range repulsion between spherical colloidal particles, J. Phys.
Chem., 88, 5735-5739, 1984.
Fogler, H.S., Elements of Chemical Reaction Engineering, 2nd Ed., Prentice
158 REFERENCES
Hall, Englewood Cliffs, New Jersey, Chapter 12, 1992.

Fowkes, F.M.,Role of acid-base interfacial bonding in adhesion, J. Adhesion
Sci. Technol., 1, 7-27, 1987.
Gerba, C.P., C. Wallis, and J.L. Melnick, Fate of wastewater bacteria and
viruses in soil, J. Irrig. Drain. Div. ASCE, 101, 157-174, 1975.
Goldenberg, L.c., M. Magaritz, and S. Mandel, Experimental investigation on
irreversible changes in hydraulic conductivity in seawater-fresh water in
coastal aquifers, Water Resour. Res., 19(1), 77---!d5, 1983.
Goldman, S., K. Jackson and T.A. Brusztynsky, Erosion and Sediment Control
Handbook, McGraw Hill, New York, Chapter 1, 1986.
Gray, D.H., and R. Rex, Fonnation damage in standstones caused by dispersive
clays, Proc. 14th Nat!. Con! on Clays and Clay Minerals, Berkeley, 355-
366, 1966.
Gray, D.H., Khilar, K. C., and Fogler, H.S., Piping and plugging as a result
of clay dispersion I swelling in porous media, Interim Report Submitted to
NSF Grant No. CEE-8313765, Washington, D.C., 1985.
Greenkorn, R.A., Flow Phenomena in Porous Media: Fundamentals and
Applications in Petroleum, Water and Food Productions, Marcel Dekker
Inc., New York, Chapter 3,1983.
Grim, R.E., Clay Mineralogy, McGraw Hill, New York, Chapter 6, 1968.
Gruesbeck, C., and R.E. Collins, Entrainment and deposition of fine particles
in porous media, Soc. Pet. Engin. J., Dec., 847---!d56, 1982.
Hardcastle, J.H., and J.K. Mitchel, Electrolyte concentration penneabi1ity
relationships in sodium-illite silt mixtures, Clays Clay Minerals, 22, 143-
154, 1974.
Harvey, R.W., George, L.H., Smith, R.L., and LeBlanc, D.R., Transport of
micro spheres and indigenous bacteria through a sandy aquifer: Results of
natural- and forced tracer experiments, Environ. Sci. Technol., 23,51-56,
1989.
Harvey, R.W., and S.P. Garabedian, Use of colloid filtration theory in modeling
movement of bacteria through a contaminated sandy aquifer, Environ. Sci.
Technol., 25, 178-185, 1991.
Helffrich, F., Ion Exchange, McGraw Hill, New York, Chapter 4, 1962.
Herrero, C., P. Pradanos, J.1. Calvo, F. Tejerina and A. Hernandez, Flux dec1in;;
in protein microfiltration: influence of operative parameters, J. Colloid
Interface Sci., 187,344-351,1997.
Herzig J.P., D.M. Leclerc, and P. Legoff, Flow of suspension through porous
media - Applications to deep bed filtration, Ind. Engin. Chem., 62(5), 8-35,
REFERENCES 159
1970.
Hesterberg, D., and M.G. Reed, Volumetric treatment efficiencies of some
commercial stabilizers, SPE Production Engin., Feb., 57---62, 1991.
Hewitt, C.H., Analytical techniques for recognizing water sensitive reservoir
rocks, J Pet. Technol., Trans AIME, 15(8), 813-817, 1963.
Hiemenz, P.e., Principles of Colloid and Surface Chemistry, Second Edition,
Revised and Expanded, Marcel Dekker Inc, New York, Chapter-13, 1986.
Himes, RE., E.F. Vinson, and D.E. Simon, Clay stabilization in low
permeability formations, SPE Production Engin., Aug., 252-258, 1991.
Hjeldness E.I., and B. Y.K. Lavania, Cracking leakage and erosion of earth dam
materials, J Geotech. Engin. Div., ASCE, 106, 117-135,1980.
Ho, N.F.H., and W.I. Higuchi, Preferential aggregation and coalescence in
heterodisperesed systems, J Pharm. Sci., 57, 436--443, 1968.
Hoefner, M.L., and H.S. Fogler, Pore evolution and channel formation during
flow and reaction in porous media, AIChE J, 34(1), 45-54, 1988.
Hogg, R., T.W. Healy, and D.W. Fuerstenau, Mutual coagulation of colloidal
dispersions, Trans. Faraday Soc., 62, 1633-1651, 1966.
Hopkins, M.R., and K.M. Ng, Liquid-liquid relative permeability, network
models and experiments, Chem. Engin. Commun., 46, 253-279,1986.
Houi, D. and R Lenormand, Particle accumulation at the surface of a filter,
Filtration Separation, 238-241, 1986.
Hubbe, M.A., Theory of detachment of colloidal particles from flat surfaces
exposed to flow, Colloids Surfaces, 12, 151-178, 1984.
Hunter, RJ., Zeta Potential in Colloid Science, Academic Press, New York,
Chapter 4, 1981.
Hunter, R.J., Foundations of Colloid Science, Vol. 1, Oxford Univ. Press,
Chapter 7, 1989a.
Hunter, R.J., Foundations of Colloid Science, Vol. 1, Oxford Univ. Press,
Chapter 4, 1989b.
Imdakm, A.O., and M. Sahimi, Transport of large particles in flow through
porous media, Phys. Rev. A, 36(11), 5304-5309,1987.
Imdakm, A.O., and M. Sahimi, Computer simulation of particle transport
processes in flow through porous media, Chem. Engin. Sci., 46(8), 1977-
1993, 1991.
Ioannidis, M.A., and I. Chatzis, Network modeling of pore structure and
transport properties of porous media, J Colloid Interface Sci., 48(5),951-
972, 1993.
Israelachvili, J.N., Intermolecular and Surface Forces with Applications to
160 REFERENCES
Colloidal and Biological Systems, Academic Press, New York, Chapter 12,
1992a.
Israelachvili, J.N., Intermolecular and Surface Forces with Applications to
Colloidal and Biological Systems, Academic Press, New York, Chapter 11,
1992b.
Jerauld, G.R., lC. Hatfield, L.E. Scriven, and H.T. Davis, Percolation and
conduction on voronoi and triangular networks: A case study in topological
disorder,.!. Phys. C. Solid State Physics, 17,1519-1529,1984.
Jones, Jr., F.O., Influence of chemical composition of water on clay blocking
of permeability, J. Pet. Technol., April, 441--446,1964.
Kallay, N, and E. Matijevic, Particle adhesion and removal in model systems
Part-IV: Kinetics of detachment of hematite particles from steel,'!' Colloid
Interface Sci., 83(1),289-300, 1981.
Kallay, N., B. Biskup, M. Tomic, and E. Matijevic, Particle adhesion and
removal in model systems Part- X: The effect of electrolytes on particle
detachment,'!' Colloid Interface Sci., 114(2),357-362,1986.
Kaplan, D.l., P.M. Bertsch, D.C. Andriano, and W.P. Miller, Soil-borne mobile
colloids as influenced by water flow and organic carbon, Environ. Sci.
Technol., 27, 1193-1200, 1993.
Kar, G., S. Chandar, and T.S. Mika, The potential energy of interaction between
dissimilar electrical double layers, J. Colloid Interface Sci., 44(2), 347-355,
1973.
Karickhoff, S.W., D.S. Brown, and T.A. Scott, Sorption of hydrophobic
pollutants on natural sediments, Water Res., 13,241-248, 1979.
Khilar, K.C., The Water Sensitivity of Berea Sandstones, Ph.D. thesis, The
University of Michigan, Ann Arbor, 1981.
Khilar K.C., and H.S. Fogler, Permeability reduction in water sensitivity of
sandstones, in Surface Phenomena in Enhanced Oi/Recovery, Ed. D. Shah),
Plenum, 721-739,1981.
Khilar, K.C., and H.S. Fogler, Water sensitivity of sandstones, Soc. Pet. Engin .
.!., 23(1), 55--64, 1983.
Khilar K.C., H.S. Fogler, and J.S. Ahluwalia, Sandstone water sensitivity :
Existence of a critical rate of salinity decrease for particle capture, Chem.
Engin. Sci., 38, 789-800, 1983.
Khilar, K.C., Effect of streaming potential on permeability of sandstones, Ind.
Engin. Chem. Fundam., 22, 264--266, 1983.
Khilar, K.C., and H.S. Fogler, The existence ofa critical salt concentration for
particle release,.!. Colloid Interface Sci., 100(1), 214--224, 1984.
REFERENCES 161
Khilar K.C., H.S. Fogler, and D.H. Gray, A model for piping and plugging in
earthen structures, J Geotech. Engin. Div. ASCE, 111(7),833-846, 1985.
Khilar, K.C., and H.S. Fogler, Colloidally induced fines migration in porous
media, Rev. Chem. Engin., 4(1 & 2), 41 - 108, 1987.
Khilar K.C., and D.S.H. Sita Ram Sarma, Colloidally and hydrodynamically
induced fines migration in porous media, Chapter-19, in Encyclopedia of
Fluid Mechanics, Vol. 10, Surface and Ground Water Flow Phenomena,
Ed. N.P. Cheremisinoff, 623-663,1990.
Khilar, K.C., R.N. Vaidya, and H.S. Fogler, Colloidally induced fines release
in porous media, J Petro. Engin. Sci., 4, pp. 213-221,1990.
Kia, S.P., H.S. Fogler, and M.G. Reed, Effects of pH on colloidally induced
fines migration, J Colloid Interface Sci., 118(1), 158-168, 1987a.
Kia, S.P., H.S. Fogler, M.G. Reed, and R.N. Vaidya, Effect of salt composition
on clay release in Berea sandstone, SPE Production Engin., Nov., 277-283,
1987b.
Kirkpatrick, S., Percolation and conduction, Rev. Mod. Phys., 45(4),574-588,
1973.
Kolakowski, J.E., and E. Matijevic, Particle adhesion and removal in model
systems, Part 1 - Monodispersed chromium hydroxide on glass, J Chem.
Soc. Faraday Trans., I 75(1), 65-78,1979.
Kuo, R.J., and E. Matijevic, Particle adhesion and removal in model systems:
Part 2 - Monodispersed chromium hydroxide on steel, J. Chem. Soc.
Faraday Trans., I 75(8), 2014 - 2026, 1979.
Kuo, R.J., and E. Matijevic, Particle adhesion and removal in model systems:
Part III - Monodispersed ferric oxide on steel, J. Colloid Interface Sci.,
78(2), 407-421, 1980.
Lappan, R.E., and H.S. Fogler, Leuconostoc mesentroides growth kinetics
with applications to bacterial proifle modification, Biotech. Bioengin., 430,
865-873, 1994.
Lappan, R.E., and H.S. Fogler, Reduction of porous media permeability from
insitu leuconostoc mesentroides growth and dextran production, Biotech.
Bioengin., 50, 6-15, 1996.
Lauzon, R.V., Chemical balance stops formation damage, Oil Gas J, Sept.,
124-126, 1982.
Leichtberg, S., R. Pfeffer, and S. Weinbaum, Stokes flow past finite coaxial
clusters of spheres in a circular cylinder, Int. J Multiphase Flow, 3, 147-
169, 1976.
Lin, C.Y., and lC. Slattery, Three dimensional, randomized network model
162 REFERENCES
for two-phase flow through porous media, AIChE J., 28(2), 311-324,1982.
Leone, J.A., and M.E. Scott, Characterization and control offormation damage
during water flooding of a high clay content reservoir, SPE Reservoir
Engin., Nov., 1279-1285,1988.
Lever A., and R.A. Dowe, Water sensitivity and migration of fines in the
Hopeman sandstone,J. Pet. Geology, 7(1), 97-108,1984.
Lin, C., and M.H. Cohen, Quantitative methods for microgeometric modeling,
J. Appl. Phys., 53(6), 4152-4165,1982.
Lindqvist, R., J.S. Cho, and G.G. Enfield, Kinetic model for cell density
dependent bacterial transport in porous media, Water Resour. Res., 30(12),
3291-3299, 1994.
Lymberopoulos, D.P. and A.e. Payatakes, Derivation of topological,
geometrical, and correlational properties of porous media from pore-chart
analysis of serial section data, J. Colloid Interface Sci., 150(1), 61-81,
1992.
Magaritz, M., A.J. Amiel, D. Ronen, and M.C. Wells, Distribution of metals
in polluted aquifer: A comparison of aquifer suspended material to fine
sediment of adjacent enviomment, J. Contaminant Hydrology, 5, 333-347,
1990.
Magee, B.R., L.W. Lion, and A.T. Lemley, Transport of dissolved organic
macromolecules and their effect on the transport of phenanthrene in porous
media, Environ. Sci. Technol., 25,323-331, 1991.
McCabe, W.L., J.e. Smith, and P. Harriott, Unit Operations of Chemical
Engineering, 5th Edition, McGraw Hill, New York, Chapter 30, 1993.
McCarty, P.L., Environmental Biotechnology: Bioengineering Issues Related
to Insitu Remediation of Contaminated Soils and Groundwater, Plenum,
New York, 143-162,1987.
McCarthy, J.P', and J.M. Zachara, Subsurface transport of contaminants,
Environ. Sci. Technol., 23(5), 496-502, 1989.
McDowel-Boyer L.M., J.R. Hunt, and N. Sitar, Particle transport through
porous media, Water Resour. Res., 22(13), 1901-1921, 1986.
McLaughlin Sr., H. C., E.A. Elphingstone, R.E. Remington II, and S. Contes,
Clay stabilzing agent can correct formation damage, World Oil, May, 58-
60,1977.
Mohan K.K., Water Sensitivity of Porous Media Containing Swelling Clays,
Ph.D. thesis, The University of Michigan, Chapter V, 1996a.
Mohan K.K., Water Sensitivity of Porous Media Containing Swelling Clays,
Ph.D. thesis, The University of Michigan, Chapter III, 1996b.
REFERENCES 163
Mohan K.K., and H.S. Fogler, Colloidally induced smectitic fines migration:
Existence of microquakes, AIChE J., 43,565-576, 1997.
Mohanty, J. and B.W. Ninhan, Dispersion Forces, Academic Press, New York,
Chapter 2, 1976.
Monaghan, P.H., R.E. Salatheil, and B.E. Morgan, Laboratory studies of
formation damage in sands containing clays, J. Pet. Techno!. Trans, AIME,
11(2),209-215,1959.
Moore, J.E., How to combat swelling clays, Petrol. Engineer, 32, 78-101,
1960.
Morrision, A., Hazarads waste landfills: Can clay liners prevent migration of
toxic leachate?, Civil Engin. ASCE, 60-63, July 1981.
Muecke, T.W., Formation fines and factors controlling their movement in
porous media, J. Pet. Technol., Feb., 144-150, 1979.
Mungan, N., Permeability reduction through changes in pH and salanity, J.
Pet. Techno!., Trans. AIME, 17, 1449-1453, 1965.
O'Connor, D.J. and J.P. Connolly, The effecct of cocncentration of adsorbing
solids on the partition coefficient, Water Res., 14, 1517-1523,1980.
O'Neill, M.E., A Sphere in contact with a plane wall in a slow linear shear
flow, Chern. Engin. Sci., 23, 1293-1297, 1968.
Pandya, v.B., S. Bhuniya and K.C. Khilar, Existence of a critical particle
concentration in plugging ofa packed bed, AIChE J., 44, 978-981, 1998.
Parker, G.G., and E.A. Jenne, Structural failure of western US highways caused
by piping, 46th Annual Meeting of Hwy. Res. Board, Washington, D.C.,
1967.
Payatakes, A.C., H.Y. Park, and J. Petrie, A visual study of particle
depositon and reentrainment during depth filtration of hydro sols with a
polyelectrolyte, Chern. Engin. Sci., 36, 1319-1335, 1981.
Penrose, W.R, W.L. Polzer, E.H. Essington, D.M. Nelson, and K.A. Orlandini,
Mobility of plutonium and americum through a shallow aquifer in a
semiarid region, Environ. Sci. Technol., 24(2), 228-234, 1990.
Peters, F. W. and C.M. Stout, Clay stabilization during fracturing treatments
with hydrolyzable zirconium salts, J. Pet. Technol., AIME, Feb., 187-194,
1977.
PuIs, RW., and RM. Powell, Transport of inorganic colloids through natural
aquifer materials: Implications for contaminant transport, Environ. Sci.
Technol., 26(3), 614--621,1992.
Quirk, J.P., and R.K. Schofield, The effect of electrolyte concentration on soil
permeability, J. Soil Sci., 20(1), 163-178, 1966.
164 REFERENCES
Reed, M.G., Stabilization of fonnation clays with hydroxy-aluminum

solutions, J Pet. Technol., July, 860-864, 1972.
Rege, S.D., and H.S. Fogler, Network model for straining dominated particle
entrapment in porous media, Chem. Engin. Sci., 42(7), 1553-1564,1987.
Rege, S.D., and H.S. Fogler, A network model for deep bed filtration of solid
particles and emulsion drops, AIChE J, 34(11), 1761-1772, 1988.
Rowel, D.L., D. Payne, and H. Ahmed, The effect of the concentration and
movement of solutions on the swelling and disperson, J Soil. Sci., 20(1),
176-188, 1969.
Ruckenstein, E., and D.C. Prieve, Adsorption and desorption of particles and
their choromatographic separation, AIChE J, 22(2), 276-283, 1976a.
Ruckenstein,E., and D.C. Prieve, On reversible adsorption of hydro sols and
repeptization, AIChE J, 22(6), 1145-1147, 1976b.
Ryan, J.N., and P.M. Gschwend, Effects of ionic stength and flow rate on
colloid release: Relating kinetics to intersurface potential energy, J Colloid
Interface Sci., 164,21-34,1994.
Ryan, J.N., and M. Elimelech, Colloid mobilization and transport in
groundwater, Colloids Surfaces, 107, 1-56, 1996.
Sahimi, M., H.T. Davis, and L.E. Scriven, Dispersion in disordered porous
media, Chem. Engg. Commun., 23,329-341,1983.
Sahimi, M., and T.T. Tsotsis, A percolation model of catalyst deactivation by
site coverage and pore blockage, J Catal., 96, 552-567, 1985.
Sahimi, M., Flow and Transport in Porous Media and Fractured Rocks: From
Classical Methods to Modern Approaches, VCH, New York, Chapter 5,
1995a.
Sahimi, M., Flow and Transport in Porous Media and Fractured Rocks: From
Classical Methods to Modern Approaches, VCH, New York, Chapter 6,
1995b.
Sanchez, R.L., A.I. Strutynsky, and M.A. Silver, Evaluation of erosion potential
of embankment core materials using the laboratory biaxial erosion test
procedure, Technical Report, GL-83-4, USCOE Waterways Exp. Stn.,
Vickshurg, Miss, 1983.
Schechter, R.A, Oil Well Stimulation, Prentice Hall, Englewood Cliffs, New
Jersey, Chapter 1, 1992.
Scheidegger, A., The Physics of Flow through Porous Materials, University
of Toronto Press, Toronto, Chapter 1, 1974.
Scholl, M.A., and R.W. Harvey, Laboratory investigations on the role of
sediment surface and ground water chemistry in transport of bacteria
REFERENCES 165
through a contaminated sandy aquifer, Environ. Sci. Technol., 26, 141 ~

1417,1992.
Schweich, D., and M. Sardin, Transient ion exchange and solubilization of
limestone in an oil field sandstone : Experimental and theoretical wave
front analysis, AIChE J., 31(11), 1882-1890, 1985.
Sharma, M.M., and T.F. Yen, Interfacial electrochemistry of oxide surfaces
in oil-bearing sands and sandstones, J. Colloid Interface Sci., 98, 39-54,
1984.
Sharma, M.M., IF. Kuo, and T.F. Yen, Further investigation of the surface
charge properties of oxide surfaces in oil-bearing sands and sandstones, J.
Colloid Interface Sci., 115(1), 9-16, 1987.
Sharma, M.M., and Y.C. Yortsos, Transport ofparticulate suspensions in porous
media: Model formulation, AIChE J., 33(10), 1636-1643, 1987a.
Sharma, M.M., and Y.C. Yortsos, A network model for deep bed filtration
processes, AIChE J., 33(10), 1644-1653, 1987b.
Sharma, M.M., and Y.e. Yortsos, Fine migration in porous media, AIChE J.,
33(10),1654-1662,1987c.
Sharma, M.M., H. Chamoun, D.S.H.R Sarma, and RS. Schechter, Factors
controlling the hydrodynamic detachement of particles from surfaces, J.
Colloid Interface Sci., 149(1), 121-134, 1992.
Shaw, D.J., Electrophoresis, Academic Press, New York, Chapter 3, 1969.
Sherard, J.L., RS. Decker, and N.L. Ryker, Piping of earth dams of dispersive
clays, Proc. ASCE Speciality Con! on Earth and Earth Supported
Structures, 1(1),589-626,1972.
Sherard, J.L., L.P. Dunnigan and J.R Talbot, Filters for silts and clays, J.
Geotech. Engin. Div., ASCE, 110(6),701-719,1984.
Simon, R, and F.J. Kelsey, The use of capillary tube networks in reservoir
performance studies - I, Soc. Pet. Engin. J., June, 99-112, 1971.
Simon, R., and F.J. Kelsey, The use of capillary tube networks in reservoir
performance studies - II, Soc. Pet. Engin. J., Aug., 345-351,1972.
Slattery, J.e., Momentum, Energy and Mass Transfer in Continua, McGraw
Hill, New York, Chapter 4, 1972.
Sloat, B.F., Field test results with alkaline potassium solutions to stabilize clays
permanentlay, SPE Reservoir Engin., May, 143-146,1990.
Smith, C. F., J.P. Pavalich, and RL. Solvinsky, Potassium, calcium treatments
inhibit clay swelling, Oil Gas J., Nov. 8~1, 1964.
Solomentsev, Y.E., Y. Pawar, and lL. Anderson, Electrophoretic mobility of
nonuniformly charged spherical particles with polarization of the double
166 REFERENCES
layers, J. Colloid Interface Sci., 158, 1-9, 1993.

Srinivasan K.R., and H.S. Fogler, Use ofinorgano-organo-clays in the removal
of priority pollutants from industrial wastewaters: Structural aspects, Clays
Clay Minerals, 38(3), 277-286, 1990a.
Srinivasan K.R., and H.S. Fogler, Use ofinorgano-organo-clays in the removal
of priority pollutants from industrial wastewaters: Adsorption ofbenzo(a)
pyrene and chlorophenols from aqueous solutions, Clays Clay Minerals,
38(3),287-293, 1990b.
Svarvsky, L., Solid-liquid Separation Proccess and Technology, Handbook of
Powder Technology, Vol. 5., Chapter 4, Elsevier, Amsterdam, 1985.
Swanton, S.W., Modeling colloid transport in ground water: The prediction of
colloid stability and retention behaviour, Adv. Colloid Interface Sci., 54,
129-208, 1995.
Swartzen-Allen, S.L., and E. Matijevic, Colloid and surface properties of
clay suspensions: Stability of montmorillonite and Kaolinite, J. Colloid
Interface Sci., 56(1), 157-167, 1976.
Sydansk, R.D., Discussion of the effects of temperature and confining on
single-phase flow in consolidated rocks, J. Pet. Techno/., 32, 1329-1330,
1980.
Sydansk, R.D., Stabilizing clays with potassium hyroxide, J. Pet. Technol., 36,
1366-1374, 1984.
Tan, Y., J. T. Gannon, P. Baveye, and M. Alexander, Transport of bacteria in
an aquifer sand: Experiments and model simulations, Water Resour. Res.,
30(12),3243-3253,1994.
Thomas, R.L., and C.W. Crowe, Matrix treatment employs new acid system
for stimulation and control of fines migration in sandstone formations, J.
Pet. Technol., 33, 1491-1500, 1981.
Thomson, G., N. Kallay, and E. Matijevic, Particle adhesion and removal in
model systems - IX: Detachment of rod-like ,B-FeOH particles from steel,
Chem. Engin. Sci., 39(7/8), 1271-1276, 1984.
Thompson, K.E., Interfacial Reactions for the Modification ofFlow in Porous
Media, Ph.D. thesis, University of Michigan, 1995.
Tien, C., Granular Filtration of Aerosols and Hydrosols, Butterworth
Publishers, Chapter 5, 1989.
Todd, A.c., lE. Somerville, and G. Scott, The application of depth offormation
damage measurements in predicting water injectivity decline, SPE 12498,
Denver, CO,1984.
Torok, l, L.P. Buckley, and B.L. Woods, The separation of radionuclide
REFERENCES 167
migration by solution and particle transport 1ll soil, J. Contaminant

Hydrology, 6, 185--203, 1990.
Tsakiroglou, e.D., and A.C. Payatakes, A new simulator of mercury
porisimetry for the characterization of porous materials, J. Colloid Inteiface
Sci., 137(2),315--339,1990.
Tsakiroglou, C.D. and A.C. Payatakes, Effects of pore-size correlations on
mercury porosimetry curves, J. Colloid Inteiface Sci., 146(2), 479-494,
1991.
Vaidya, R.N., and H.S. Fogler, Formation damage due to colloidally induced
fines migration, Colloids Surfaces, 50, 215--229, 1990.
Vaidya, R.N., Fines Migration and Formalation Damage, Ph.D. thesis, The
University of Michigan, Ann Arber, 1991.
Valey, e.D., How hydrolyzable metal ions react with clays to control formation
water sensitivity, J. Pet. Technol., Sept., 1111-1118, 1969.
van alphen, H., Introduction to Clay Colloid Chemistry, John Wiley and Sons,
New York, Appendix-II, 1977.
van ass, C.J., M.K. Chaudhary, and R.J. Goed, Interfacialliftshitz- van der
Waals and polar interactions in macroscopic systems, Chem. Rev., 88, 927-
941, 1988.
van ass, C.J., Interfacial Forces in Aqueous Media, Marcel Dekker, Inc., New
York, Chapter 1, 1994.
Visser, J. in Surface and Colloid Science, E. Matijevic Ed., Chapter 8, Wiley,
New York, 1976.
Voice, T.e., and W.J. Weber Jr., Sorption of hydrophobic compounds by
sediments, soils and suspended solids - I, Theory and background, Water
Res., 17(10), 1433-1441,1983.
Voice, T.C., C.P. Rice, and W.J. Weber Jr., Eftect of solid concentration on
sorptive partitioning of hydrophobic pollutants in aquatic systems, Environ.
Sci. Tech., 17(9), 513-517, 1983.
Vrbanac, M.D., and J.e. Berg, The use of wetting measurements in the
assessment of acid-base interactions at solid-liquid interfaces, Acid-Base
Interactions at Solid-Liquid Interfaces, Ed. by Mittal K.L and H.R.
Anderson, Jr. VSP, 67-78, 1991.
Weber Jr., W.J., T.J. Voice, M. Pirbazari, G.B. Hunt, and D. Ulanoft, Sorption
of hydrophobic compounds by sediments, soils and suspended solids - II
Sorbent evaluation studies, Water Res., 17, P 1443-1452, 1983.
Wiese, G.R., and T.W. Healy, Effect of particle size on colloid stability, Trans.
Faraday Soc., 66, 490-499,1970.
168 REFERENCES
Williams, D.J.A., and K.P. Williams, Electrophoresis and zeta potential of

kaolinite, J Colloid Interface Sci., 65(1), 79-'(,7, 1978.
Xu, Z. and R Yoon, The role of hydrophobic interactions in coagulation, J
Colloid Interface Sci., 132(2),532-541, 1989.
Yanuka, M, F.A.L. Dullien, and D.E. Elricgh, Serial sectioning and
digitization of porous media for two- and three- dimensional analysis and
reconstruction, J Micro Sci., 135, 159-168, 1984.
Yao, K., M.T. Habibian, and C.R. O'Melia, Water and waste-water filtration:
Concepts and applications, Environ. Sci. Technol., 5(11),1105-1172,1971.
Young, B.M., H.C. McLaughlin, and J.K. Borchardt, Clay stabilization agents
- their effectiveness in high temperature steam,J Pet. Techno!., Dec. 2121-
2131,1980.
Zahka, J., and L. Mir, Ultrafication of Latex emulations, Chern. Engin. Prog.,
73(12),53-55, 1977.
Zaltoun A. and N. Berton, Stabilization of montmorillonite clay in porous
media by high molecualr weight polymers, SPE Production Engin., May,
160-166, 1993.
INDEX
A coordination number, 11, 13, 74, lO9,

115
AB interaction, 36 critical flow velocity, 67, 69, 93
acidization, 2 critical particle concentration, 75, 83,
adhesive force, 65, 66 87
attapulgite, 22 critical rate of salinity decrease, 79, 88
critical salt concentration, 46, 93
B critical shear stress, 67, 69
critical total ionic strength, 51
bacteria, 5 crystalline regime, 54
Berea sandstone, 2, 20, 23, lO6
Bethe lattice, 11
Biocolloidal contaminants, 148 D
biocolloidal fines, 5, 26
Boltzman constant, 32 Darcy's law, 14,82
bonds, 114 deep bed filtration, 7
Born repulsive potential, 36 direct interception, 95
Brownian fines, 55 DLVO theory, 44
double layer repulsion, 29, 40
c drilling, 2
cake filtration, 8 E
capture coefficient, 82, 95, 98
cation exchange capacity, 24 effective medium approximation, 118
cationic organic polymers, 140 electro-kinetic interactions, 48
charge regulation models, 39 energy barrier, 46, 55
Chemical characterization, 17, 26 entrapment, 7, 73, 74, 103
chemical methods, 136 erosion rate coefficient, 70, 71
clay particles, 2, 20
clay stabilization, 133
Colloidal characterization, 16, 23 F
colloidally induced release, 29
Constant charge, 32 facilitated contaminant transport, 144
Constant potential, 32 fine particles, 1
contact deformation, 66 fines, 19
contamination of soils, 5 Fokker-Plank equation, 55
convective jamming, 95 force of detachment, 136
169
170 INDEX
G migration of fine particles, 1,3,5

migratory fines, 19,25
ground water pollution, 5
Monod equation, 149
montmorillonite, 20, 26
H moving permeability front, 93
multiparticle bridging, 95
Hagen-Poiseuille, 120
Hamaker constant, 33, 35
hydrodynamic detachment, 63 N
hydrodynamic forces, 64
hydrodynamically induced release, 29 Network models, 114, 117, 128
hydrolyzable metal ions, 139 nodes, 114
hydrophobic interaction, 37 non-Brownian particles, 55
hydroxyl-aluminum, 140
o
I
osmotic regime, 54
illite, 20,21,26
interaction force boundary layer, 56 p
interconnectivity,74, 115
ion exchange, 52 percentage of calcium, 50
permeability, 14,83, 101, 103
K permeability reduction, 89, 102, 106,
109
kaolinite, 20, 21, 26,52 physical method of attachment, 134
physically representative network, 131
L piping, 3, 4, 73, 75, 76, 79
plugging, 3, 7, 73, 76, 78, 79, 103
lattice arrangement, 115 Poisson-Boltzman equation, 31
leaching of polycations, 52 Pollution, 5
Lifshitz theory, 34
pore, 9
lift force, 37, 64
pore chamber size distribution, 10
London-van der Waals, 29, 33, 40
pore chambers, 10, 11
pore closure, 98
M pore constriction size distribution, 10,
11
mass balance for the fine particles in
pore constrictions, 10, 11
the suspension, 81
mica, 20 pore size distribution, 13
micro-organisms, 5, 148 pore structure, 74
microquake, 54 porosity, 14,80, 83, 84
migration of biocolloidal fines, 141, porous media, 1, 9, 74
149 primary minimum, 40, 41, 54
INDEX 171
R T
Rate expressions for release offine par- total interaction energy, 39, 41
ticles,96 total mobile concentration of contami-
rate of capture, 82 nant, 148
rate of erosion, 70 trajectory calculations, 130
release and capture mechanisms, 91 tunnel erosion, 3, 4
release coefficient, 56, 59, 61, 109
retarded interaction, 34
u
s Unconsolidated porous media, 10
Sand bed filters, 87

sand control, 135
v
saponite, 22 viruses, 5
secondary minimum, 41, 54 void fraction, 10
sequential reduction in permeabi1ity,
93
size exclusion, 95 w
Smoluchowski's equation, 55
soil erosion, 84, 136 washout, 7
soil microflora, 19 water flooding, 2
sorption coefficient, 142 Water sensitivity, 2
sorption of hydrophobic compounds, water sensitivity, 2, 3, 88
142 water shock, 2
specific characteristics of cation, 48 wettability, 15
Stevens sandstone, 46 wrap-around boundary conditions, 122
streaming potential, 16
surface charge, 17,23,24
surface erosion, 4
z
surface potentials, 39 zeta potential, 16,24,25,48
surface roughness, 15
swelling clays, 54
Theory and Applications of Transport in Porous Media
Series Editor:
Jacob Bear, Technion -Israel Institute o/Technology, Haifa, Israel
1. H.1. Ene and D. Polissevski: Thermal Flow in Porous Media. 1987

ISBN 90-277-2225-0
2. J. Bear and A. Verruijt: Modeling Groundwater Flow and Pollution. With Computer
Programs for Sample Cases. 1987 ISBN 1-55608-014-X; Pb 1-55608-015-8
3. G.I. Barenblatt, V.M. Entov and V.M. Ryzhik: Theory o/Fluid Flows ThroughNatural
Rocks. 1990 ISBN 0-7923-0167-6
4. J. Bear and Y. Bachmat: Introduction to Modeling o/Transport Phenomena in Porous
Media. 1990 ISBN 0-7923-0557-4; Pb (1991) 0-7923-1106-X
5. J. Bear and J-M. Buchlin (eds.): Modelling and Applications o/Transport Phenomena
in Porous Media. 1991 ISBN 0-7923-1443-3
6. Ne-Zheng Sun: Inverse Problems in Groundwater Modeling. 1994
ISBN 0-7923-2987-2
7. A. Verruijt: Computational Geomechanics. 1995 ISBN 0-7923-3407-8
8. V.N. Nikolaevskiy: Geomechanics and Fluidodynamics. With Applications to Reser-
voir Engineering. 1996 ISBN 0-7923-3793-X
9. V.1. Selyakov and V.V. Kadet: Percolation Models for Transport in Porous Media.
With Applications to Reservoir Engineering. 1996 ISBN 0-7923-4322-0
10. J.H. Cushman: The Physics 0/ Fluids in Hierarchical Porous Media: Angstroms to
Miles. 1997 ISBN 0-7923-4742-0
11. J.M. Cralet and M. EI Hatri (eds.): Recent Advances in Problems 0/ Flow and
Transport in Porous Media. 1998 ISBN 0-7923-4938-5
12. K.C. Khilar and H.S. Fogler: Migration o/Fines in Porous Media. 1998
ISBN 0-7923-5284-X
Kluwer Academic Publishers - Dordrecht / Boston / London

Migration of Fines in Porous Media

Uploaded by

Copyright:

Available Formats

Migration of Fines in Porous Media

Uploaded by

Document Information

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

Migration of Fines in Porous Media

Uploaded by

Copyright:

Available Formats

MIGRATION OF FINES IN POROUS MEDIA

Theory and Applications of Transport in Porous Media

Springer-Science+Business Media, B.Y.

ISBN 978-90-481-5115-8 ISBN 978-94-015-9074-7 (eBook)

Printed on acid-free paper

All Rights Reserved

1 PRACTICAL CONSEQUENCES OF RELEASE AND

2 CHARACTERIZATION OF PORE SPACE AND FINES 9

3 COLLOIDALLY INDUCED RELEASE OF FINES IN

3.1.4 AB (Acid-Base) Interaction 36

4 HYDRODYNAMICALLY INDUCED RELEASE OF FINES IN

5 ENTRAPMENT OR PIPING OF FINES DURING MIGRATION 73

6 MATHEMATICAL MODELS FOR PERMEABILITY

6.5 Correlation between Entrapment and Permeability Reduction . 101

7 USE OF NETWORK MODELS FOR PREDICTION OF PERME-

8 METHODS TO PREVENT THE RELEASE OF FINES 133

9 SOIL POLLUTION DUE TO MIGRATION OF FINES 141

9.1.1 Sorption of Hydrophobic Contaminants on Fines in

The mechanism of release is probably the most important basic process in

at lIT Bombay and at the University of Michigan have indirectly contributed

PRACTICAL CONSEQUENCES OF RELEASE AND

1.1 Migration of Fine Particles in Porous Media: An Introduction

1.2 Practical Consequences in Petroleum Engineering

A widely recognized consequence of migration of fines in sandstone oil

3°10 Noel Solution Fresh wo~er

Figure 1.2.1. Permeability reduction in water sensitivity of sandstones (Khilar

operations such as solution mining, and hydraulic fracturing etc. It may be

1.3 Importance in Geotechnical Engineering

Figure 1.3.1. A schematic illustration of piping and tunnel erosion across a

presence of a leak. In these cases, it begins by surface erosion of soil at the

1.4 Effects in Environmental Engineering

Contam nant movement

(b) Three-phase system

Figure 1.4.1. A schematic diagram of movement of contaminants (McCarthy

Figure 1.5.1. Flux vs. concentration plot (McCabe et at, 1993).

over a mile. Furthermore, the migrating contaminants were shown to be present

1.5 Importance in Chemical Engineering

Unconsolidated porous media are widely used in many chemical engineering

et. al., 1997).

CHARACTERIZATION OF PORE SPACE AND FINES

2.1 Pore Structure

A porous material/medium contains voids (free of solids) dispersed in a solid

Figure 2.1.1. Two types of packing of spherical particles and corresponding

2.1.1 PORE CONSTRICTION AND CHAMBER SIZE DISTRIBUTIONS

Figure 2.1.2. Electron micrographs of some porous materials (Dullien, 1979a).

Figure 2.1.3. Optical micrograph of a soil sample (Biswas and Mukhetjee,

TABLE 2.1.1. Pore constriction and chamber sizes of some sandstones

Sandstone Pore constriction Pore chamber Reference

Berea 0.5 to 5.0 5 to 50 Dullien and Dhawan (1974)

Boise 39-74 Ioannidis and Chatzis (1993)

STl 33-120 33-250 Lymberpolous and

distribution in rocks (Sahimi, 1995b).

2.1.2 PORE CONNECTIVITY

The connectivity of a pore network is generally characterized by a coordination

2.1.3 POROSITY AND PERMEABILITY

2.2 Pore Surface

characterization, will be discussed: surface, colloidal, and chemical.

2.2.1 SURFACE CHARACTERIZATION OF PORE SURFACE

50 UNBAKED BEREA SANDSTONE

Figure 2.2.1. Variation of zeta potential of crushed Berea sandstone with pH

-IO'--_~_~_.J'~ ..........~'___'_ _~~_