-
.Applied Clay Science 15 1999 447462
Influence of pore water chemistry on the swellingbehavior of
compacted clays
Waddah S. Abdullah a,1, Khalid A. Alshibli b,),Mohammed S.
Al-Zoubi c,2
a Department of Ciil Engineering, Jordan Uniersity of Science
and Technology, PO Box 3030,Irbid, Jordan
b Department of Ciil and Enironmental Engineering, Uniersity of
Alabama in Huntsille,Huntsille, AL 35899, USA
c Department of Ciil Engineering, Uniersity of Illinois at
Urbana-Champaign, Urbana, IL61801, USA
Received 16 June 1998; received in revised form 7 May 1999;
accepted 11 May 1999
Abstract
The influences of the exchange complex and pH of the solution
used for cation saturation onAtterberg limits, compaction, and
swelling potential of a compacted clay were investigated. Thestudy
involved transforming the exchange complex from a heterogeneous to
a homogeneous oneso that a frame of reference can be set for the
clay behavior under such an ideal condition. Theemployed method for
altering the exchange complex successfully yielded homo-ionic clay.
Theintroduction of different species of cations gave rise to
different particles associations. Whenintroduced to the tested
clay, potassium cations bond its particles with a rather strong
bond . K-linkage , causing a drastic decrease in the specific area
of the clay about one-fourth of its
.untreated specific area , a decrease in the CEC, as well as a
drastic decrease in the swell potential.For example, the swell
pressure decreased from 1.87 kgrcm2 for the untreated samples to
0.4
2 .kgrcm for the K-treated samples under the same conditions .
Also, the swell potential vs. timerelationships can be modeled
accurately using a rectangular hyperbola. q 1999 Elsevier
ScienceB.V. All rights reserved.
Keywords: expansive clays; swelling; cation exchange;
compaction; Atterberg limits
) Corresponding author. Tel.: q1-256-544-3051; Fax:
q1-256-890-6724; E-mail:[email protected]
1 E-mail: [email protected] Tel.: q1-217-333-7311; Fax:
q1-217-333-9464; E-mail: [email protected]
0169-1317r99r$ - see front matter q 1999 Elsevier Science B.V.
All rights reserved. .PII: S0169-1317 99 00034-4
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462448
1. Introduction
It is well known that the exchange complex has an important role
incontrolling properties and engineering behavior of clays. The
magnitude of theinfluence varies according to the properties under
consideration, the type of clayinvolved, and other environmental
conditions. Atterberg limits are fundamentalproperties that are
extensively used as a measure for soil classification and
asparameters for correlation to predict the soils engineering
behavior such asswelling and compressibility.
Swell phenomenon is known to cause serious damage to low rise
buildings,dams, and highways. Ultimate heave and rate of heave are
the two factors thatcause structures built on expansive soils to
experience damages due to differen-tial settlements or expansion.
In clayey soils, the physico-chemical swell is themajor and the
important part of the soils heave. The claywaterelectrolytesystem
is the main factor affecting the physico-chemical swell. The clay
mineralparticles together with the adsorbed exchangeable cations
constitute the so-called
.exchange complex Marshall, 1977; Hillel, 1980 . An exchange
reaction maytake place which results in altering the fractions of
the exchangeable cationswithin a heterogeneous or a homogeneous
exchange complex. A typical ex-change reaction is given according
to the following equation:
In this paper, a physico-chemical study was conducted using an
expansiveclay known as Azraq Green clay. The investigation involves
altering the
q 2q 2qexchange complex from a heterogeneous untreated having Na
, Mg , Ca ,q . and K exchangeable cations to a homogeneous exchange
complex having
.mainly one species of cations present on the clay . This is
considered as afundamental step for establishing a frame of
reference, for future studies, forassessing real conditions of
heterogeneous exchange complex. The alteration
.process was conducted three times, to transform the exchange
complex to i . .K-dominated ii Na-dominated or iii Ca-dominated
cation exchange complex,
so that the swell behavior under such an ideal condition can be
established firstand then used as a basis for establishing soil
behavior under real conditions ofheterogeneous exchange
complex.
The influence of clay structure on the swell behavior was also
investigated.The structure of compacted clays can be made to fit
certain research needs
ranging from flocculated to oriented structure Lambe, 1953; Seed
and Chan,.1959 . Because of that, compacted clays can offer an
excellent opportunity for
studying the influence of clay structure on its behavior. In
this investigation, the
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
449
Fig.
1.X
-ray
diffr
acto
gram
soft
hete
sted
soil.
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462450
.structure has been manipulated via the use of i the molding
moisture content, . .ii exchangeable cations, and iii the pH value
of the solution used for cationsaturation.
2. Materials and methods
2.1. Clay compositional properties
The clay used in this investigation is natural clay known as
Azraq green clay,obtained from Azraq basins, Jordan. This clay
spreads over large area in manyparts of Jordan.
The clay minerals present in the Azraq green clay are: mixed
layerillitersmectite, kaolinite, palygorskite, montmorillonite, and
discrete illite.Kaolinite presence was evident as can be seen from
the X-ray diffractograms of
Fig. 2. Transmission electron microscope photo of
palygorskite.
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
451
. .the tested soil Fig. 1 . The strongest peaks of the kaolinite
d spacing in A are7.12 and 3.58. These peaks disappeared and became
amorphous to X-rays afterheating to 5508C for 1 h. As for the mixed
layer IrS, since the diffractionpatterns of the ethylene glycol
solvated and the air-dried are different then weshould be having a
certain mineralrsmectite. Since the EG has a peak at
16.8 A and the air-dried has a peak at 14.1 A, then the mixed
layer must be an .IrS Moore and Reynolds, 1997 . The peak at 5.01
also an indication of the
.presence of IrS. The peaks d spacing in A at 10.56, 4.46, 4.26
and 3.18 are .indicative of the presence of palygorskite Moore and
Reynolds, 1997 . Illite is
unaffected by ethylene glycol solvation and heating to 5508C.
The strongest .peaks d spacing in A are 10.1, 5.01, and 3.34. The
peaks at 16.8 A and 5.65 A
are indicative of the presence of montmorillonite.Fig. 2 is a
Transmission Electron Microscope of a clay specimen showing a
concentration of palygorskite laths with other clay minerals
particles this photo.does not reflect actual percentages of the
present clay minerals . The approxi-
mate percentages of the clay minerals present in the soil,
palygorskite, illite,illitersmectite, kaolinite, and
montmorillonite, are about 16%, 39%, 9%, and10%, respectively. The
specific surface area of the tested clay was determined
from different locations in Azraq basin, using the
2-ethoxyethanol Ethylene ..Glycol Monoethel Ether EGME surface area
method. Cation Exchange Ca-
. .pacity CEC was determined according to Polemio and Rhoades
1977 for anumber of clay specimens obtained from different
locations in the Azraq basin.The CEC vs. specific surface
relationship is shown in Fig. 3.
Fig. 3. Cation exchange capacity vs. surface area of Azraq green
clay.
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462452
Table 1Influences of exchange complex and pH on Atterberg
limitsType pH value LL PL PI SLUntreated clay 6.2 107 52 55
22Na-clay 2 58 37 21 21
7 57 33 24 2212 58 40 18 22
Ca-clay 2 82 45 37 217 86 51 35 21
12 87 55 32 21
K-clay 2 56 45 11 237 63 48 15 23
12 55 45 10 24
2.2. Alteration of the exchange complex
In order to make a specific species of cations dominant in the
exchange .complex, the clay was treated with concentrated solutions
1yN concentration
of sodium, calcium, and potassium chlorides. In each case, the
solution was . .buffered for three pH values; acidic pHs2 , neutral
pHs7 , and basic
.pHs12 . The clay was washed three times for each case. After
three washes,the clay became nearly homo-ionic more details about
the cation saturation
.procedure and results are referred to Abdullah et al., 1997
.
2.3. Specimen preparation and testing procedures
.Liquid limit, plastic limit, and shrinkage limit Atterberg
Limits for theuntreated and treated samples were measured following
the standard proceduresD-4318, D-4318, and D-427 of the American
Society for Testing and Materials .ASTM , respectively. Table 1
shows this highly plastic clay has a Liquid Limit . . .LL of 107, a
Plastic Limit PL of 52, and Plasticity Index PI of 55. Theeffects
of exchangeable cation type and pH-value on Atterberg Limits
wereinvestigated.
Table 2States of compaction considered in the investigation
3 . .Compaction state Molding moisture content % Dry density
grcmState No. 1 25.3 1.25State No. 2 31.5 1.25State No. 3 40.5
1.25State No. 4 31.5 1.30
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
453
.Fig. 4. Moisture content vs. dry density relations for the
tested clays standard proctor test .
The percent free swell and swelling pressure experiments were
performed oncompacted specimens using Oedometer Apparatus with a
diameter of 76 mm
.and a height of 20 mm following ASTM Test Methods for
One-Dimensional .SwellrSettlement Potential of Cohesive Soils
Methods D 4546 . The specimens
were prepared for four dry densities vs. molding moisture
content compactionstates as defined in Table 2 and shown in Fig. 4.
To prepare a specimen for apercent swell test, dry clay was mixed
with deionized water to give the requiredwater content, placed in
the consolidation cell, and compacted to give therequired density.
A surcharge of 0.07 kgrcm2 was applied, then the specimenwas
saturated with deionized water and the values of swelling with time
wererecorded. The measurements continue until the swelling
increment reach negligi-ble values. Also, all swelling pressure
tests were performed using the samepreparation procedure and
apparatus. Swelling pressure is defined as the valueof pressure
required to keep the sample at zero swelling after saturating it
withdeionized water.
3. Results and discussion
3.1. Moisturedensity relationships
Different species of cations present in the exchange complex
influences theinteraction between clay particles. The kind of
interaction is generated through
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462454
.the repulsive forces due to the diffuse double layer DDL and
the attractive vander Waals forces, leading to various types of
particles associations. Accord-
.ingly, the moisture-dry density vr relationship assumes
different locationsdin the vr domain. This behavior is depicted
through the results of thed
.standard Proctor tests ASTM, D 698-78 on the untreated and
treated samples .Fig. 4 . The K-clay and the Ca-clay produced
rather close vr relationships,dhaving higher maximum dry densities
and smaller Optimum Moisture Contentsas compared with that of the
Na-clay and the untreated clay. This type of
.behavior is attributed for the Ca-clay , to the dominating
effect of the van derWaals forces since the electric DDL is
depressed drastically due to the presence
of a divalent Ca cations in the exchange complex Lambe, 1953;
Olson, 1963;.Van Olphen, 1977 . The net attractive force leads to a
face-to-face aggregation
of clay particles. Consequently, particles become bigger due to
their stacking.K-clay behaves in a similar fashion as Ca-clay, due
to face-to-face aggregationof particles. However, the mechanism
responsible for the aggregation differsfrom that of the Ca-clay. As
consequence of the isomorphous substitution a highelectric
polarization exists near the clay particle surface especially
within theStern layer. Such a high electric polarization coupled
with a low hydration
.energy of the potassium cations 80 kcalrg ion, Lebedev, 1958
causes thehydrated K cations to shed all its water of hydration. As
a result, K ions with an
. .unhydrated radius of 1.33 A enters the hexagonal holes of the
siloxane silica .sheets having a radius of 1.32 A and provides a
strong linkage between clay
.particles Grim, 1968; Greenland and Mott, 1978 . The K-linkage
provided bythe Potassium ions causes the clay particles to
aggregate and form domains .Mesri and Olson, 1970 . The K-linkage
is stronger than the secondary valancevan der Waals attractive
forces that are responsible for forming domains for thecase of
Ca-clay. Thus, the domains formed by the K-linkage are more
stable
.than that formed by the van der Waals forces Ca-clay . The size
of eachdomain is much larger than the discrete clay particles. The
K-clay can be easilydistinguished from the untreated clay by its
coarse texture. Fig. 5 is a TEMpicture of a soil specimen treated
with Kq which shows large domains formedby K-linkage. The larger
the particles the less the colloidal activity, and the lessthe
plasticity index of the K-clay. The plasticity index was reduced
drasticallyfrom 55 at pHs6.2 for the untreated samples to 15 at
pHs7 for the K-clay.
A percolation test with distilled de-ionized water was performed
on theK-treated and the Na-treated samples. The percolation test
was followed by a
hydrometer test with 15 min mechanical agitation and with adding
dispersing.agent to determine whether changes to the soil particles
association has taken
place and became permanent. The result of the hydrometer test is
shown in Fig.6 which shows that the K-treated sample has become
coarser indicating that theK-linkage in the K-treated samples is
permanent. On the other hand, theNa-treated clay produces thick
DDL; however, high concentration of Na reducesthe thickness of the
DDL. Percolation with distilled and de-ionized water
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
455
Fig. 5. TEM of the K-clay soil showing lathes of polygorskite
and large domains formed by theK-linkage.
reduces Na concentration; and with subsequent agitation and
adding dispersingagent in the hydrometer test bring the Naq-treated
soil to a state similar to thenatural soil as far as grain size is
concerned.
For the Na-clay, the thick DDL leads to an important osmotic
repulsive forcethat overwhelms the van der Waals attractive forces
and causes a net repulsiveforce. This situation is conducive to a
large volume occupied by water instead ofsolids resulting in low
dry density and higher moisture content compared withthe other
treated samples.
3.2. Atterberg limits
When the clay was treated with potassium, the specific surface
area wasreduced from about 465 m2rg for the untreated samples to
about 135 m2rg forK-clay. The CEC was also reduced from 33 meqr100
g to 10.3 meqr100 g .Fig. 3 . The reason for these changes is in
the nature of the potassium ion and
.the surface nature of the clay minerals silica sheet . As
mentioned earlier, theK-linkage causes face-to-face association of
clay particles and the major part of
. .the K ions bonding the layer silicates stays in an
unreplaceable fixed state,
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462456
Fig. 6. Grain size distribution of the natural, Na-treated, and
the K-treated soils.
hence causing the actual CEC to decrease. As a direct
consequence, clayparticles are bigger and fewer, causing the
specific surface area to decrease as
.well Fig. 3 .A change in the electric potential of the clay
particles, brought about by
changes in the exchange complex, influences the state of
attraction or repulsion .of the clay particles, and thus influences
its properties. Atterberg limits Table 1
were greatly affected by changes in the exchange complex. The
potassiumtreated samples were the most affected by these changes,
since the K-linkagetransformed it into a coarser type, with less
activity due to the reduction in themeasured CEC.
3.3. Clay structure and swelling behaior
.For a certain clay, the structure is influenced by: i the
exchange complex, . . .ii molding moisture content, iii the pH
value, and iv method of compaction.The exchange complex is mainly
influenced by valence, concentration, size, andhydration
characteristics of the cations present in the exchange complex.
TheDDL is influenced by the exchange complex. There are a number of
theoriesused to predict the DDL thickness as well as the
distribution of ions adjacent tocharged surfaces in colloids. The
GauyChapman theory is the most widelyused one. Predictions provided
by most DDL theories have major limitation dueto the assumption
that ions are point charges with no interactions between them,and
the charge deficiency is uniformly distributed on a planner surface
extend-ing large distance in the plane. Though quantitatively the
DDL theories may not
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
457
yield accurate predictions depending on the real condition
compared to the.assumptions . However, they can be used
qualitatively to interpret clay behavior
.Lambe, 1953; Seed and Chan, 1959 . The type of clay structure
formed due tothe effect of the DDL and the molding moisture content
was discussed earlier.The pH of the saturation cation solution does
not influence particles surfacescharge. Such charge is mainly
caused by isomorphous substitution. The pH
value mainly affects the hydroxyl group clay minerals at
particle edges Green-.land and Mott, 1978 . The charges on
particles edges arises from the associa-
q.tion of the hydroxyls with hydrogen ion H below Point of Zero
Charge . qPZC giving rise to a positive charge, or loss of H above
the PZC giving riseto a negative charge. Therefore a low pH value
promotes a positive edge to
.negative surface interaction often if other mentioned factors
are favorableleading to a flocculated structure. Oriented structure
often occurs under condi-
.tions of high pH values i.e., pH)PZC .Clay structure is not an
easy term to quantify nor a state that can be uniquely
described. For instance, a flocculated structure can occur at
low molding . moisture content dry of optimum high in water
deficiency Seed and Chan,
.1959 . Dry of optimum molding moisture content, low pH
condition, high in .cation concentration having low percentage of
mono-valent cations are condi-
tions conducive to a flocculated structure. Wet of optimum
molding moisturecontent, high pH, low concentration of cations, and
high percentage of mono-va-lent cations are conditions for a
dispersed or oriented structure.
Swell potential and clay structure are intimately related. The
more the .flocculated structure for a specific clay the more the
swell potential. Thus, the
factors that give rise to a flocculated structure are conducive
to a high swellingpotential and vice versa. Increase in pH value
contributes to a dispersed
structure an oriented structure assembles due to the negative
charge on clayparticles surfaces and the rise of negative edges
charges, thus contributing to
.lower swelling potential swell pressure as well as percent free
swell as shownin Figs. 7 and 8. Potassium dominated exchange
complex produced the lowestvalues of swell as compared to untreated
and other treated samples. TheK-linkage provided by the potassium
cations contributed to strong face-to-faceassociations of clay
particles preventing water molecules to penetrate through,thus
causing drastic decrease in swell potential. Another aspect of the
potassiumdominated exchange complex is its insensitivity to pH
changes. The K-linkagecontributed, in fact, to decrease in specific
area of the clay causing low surficial
.activity giving rise to larger size particles , and hence edge
charges becameinsignificant.
3.4. Modeling swell potentialtime behaior
.Swell potential swell pressure or percent free swell is
time-dependentprocess. Swell pressure or free swell constitute a
single point on the swell
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462458
Fig. 7. Influence of pH and compaction state on swell
pressure.
potentialtime relationship. Knowing the full picture of the
swelling behaviorrepresents an indispensable need for accurate
analyses of swelling problems.Although, such relationships might be
very difficult to incorporate in anyanalytical solution, they are
easy to implement in finite element analysisprovided that such
relationship can be modeled by a certain
mathematicalformulation.
Moreover, measured swell potential in the laboratory requires
saturation ofthe tested clay specimen. In the field, clay
saturation is not necessary to takeplace. The amount of swell
potential reached in the field is dependent on the
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
459
Fig. 8. Influence of pH and compaction state on percent free
swell.
amount of available moisture and time. Thus, swell potential
prediction, basedon laboratory tests, represents an upper bound.
The closeness of the predicted tothe actual is dependent on the
amount of water absorbed by the clay which is, inturn, a time
dependent process. The presence of a swell-time relationship
modelprovides a powerful tool for accurate prediction of soils
swell potential.
Fig. 9 shows typical results of percent free swell and swell
pressure vs. squareroot of time relationships. These experimental
results, very clearly, indicate that
.the behavior of the swelling potential vs. time or square of
time can be
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462460
Fig. 9. Measured and predicted swell potentials relations.
modeled accurately by a rectangular hyperbola. The mathematical
form of therectangular hyperbola is given by:
tSs 2 .aqb t
where S is the percent free swell or swelling pressure, t is
time; and a and b areconstants of the hyperbola to be determined
from experimental results. Theconstants a and b are the slope and
intercept of the straight line fit of 1rS vs.
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999 447462
461
1r6t as given below and shown in Fig. 9. Let hs1rS and js1r6t:
.substituting h and j in Eq. 2 and rearranging yields:
hsajqb 3 .
4. Summary and conclusions
An experimental study was conducted on influence of the exchange
complexon Atterberg limits, compaction, and swell potential of an
expansive clay. Thestudy involves transforming the heterogeneous
exchange complex to a homoge-neous one. Swell potential tests were
performed on compacted specimens.Compaction states were chosen so
that clay structure influence on swellpotential can be included.
CEC and specific surface area were determined so asto assist in the
interpretation of the test results. The following conclusions
arebased on data, analyses, and discussion presented in this
paper.
.1 The K-dominated exchange complex caused fundamental changes
to theinvestigated clay properties and behavior. These changes are
caused by theK-linkage which bonds the silica sheets of the clay
mineral particles in aface-to-face association. Consequently, the
average size of the particles becamebigger, transforming it to a
rather coarse grained clay.
. .2 The K-linkage caused drastic decrease in: a the plasticity
index from 55 . . .pHs6.2 for the untreated clay to 15 pHs7 for the
K-clay; b the surface
2 2 .area from 442 m rg for the untreated clay to 135 m rg for
the K-clay; c theCEC from 31 meqr100 g for the untreated clay to
11.3 meqr100 g for the
. 2K-clay; d the swell pressure from 2.47 kgrcm for the
untreated clay to 0.492 . .kgrcm for the K-clay under same
conditions ; and e the percent free swell
from 16.9% for the untreated clay to 8.35% for the K-clay under
same.conditions .
.3 Altering pH value influences orientation of clay particles.
The higher thepH value the more the oriented clay particles.
.4 Swell potential is highly influenced by the exchange complex
and to alesser degree by the pH value especially for the
K-clay.
. .5 The swell potential swell pressure and percent free swell
vs. timerelationships can be accurately modeled by a rectangular
hyperbola.
5. Nomenclature
CEC Cation exchange capacityDDL Diffuse double layerKq, Naq,
Ca2q Potassium, sodium, and calcium cations, respectivelyLL Liquid
limit
-
( )W.S. Abdullah et al.rApplied Clay Science 15 1999
447462462
.PZC Point of Zero Charge the pH at which the net charge is
zeroPI Plasticity indexPL Plastic limitS Swelling pressure or
percent free swellSL Shrinkage limitt Timer Clay dry densitydv
Moisture content
References
Abdullah, W.S., Al-Zoubi, M.S., Alshibli, K.A., 1997. On the
physicochemical aspects ofcompacted clay compressibility. Can.
Geotech. J. 34, 551559.
.Greenland, D.J., Mott, C.J.B., 1978. In: Greenland, Hayes Eds.
, The Chemistry of SoilConstituents, Chap. 4. Wiley, New York.
Grim, R.E., 1968. Clay Mineralogy. McGraw-Hill Book, New
York.Hillel, D., 1980. Fundamentals of Soil Physics. Academic
press, New York.Lambe, T.W., 1953. The structure of inorganic
soils. Proc. of the American Society of Civil
Engineers, Vol. 79, Separate No. 315.Lebedev, V.I., 1958. Basic
factors determining adsorption of ions in soils. Soviet Soil Sci.
J., No.
6, pp. 606611.Marshall, C.E., 1977. The Physical Chemistry and
Mineralogy of Soils. Wiley, New York.
.Mesri, G., Olson, R.E., 1970. Shear strength of
montmorillonite. Geotechnique 20 3 , 261270.Moore, D.M., Reynolds,
R.C., Jr., 1997. X-ray Diffraction and the Identification and
Analysis of
Clay Minerals, 2nd edn., Oxford Univ. Press, Oxford, New
York.Olson, R.E., 1963. Effective stress theory of soil compaction.
J. Soil Mech. Found. Div., Am. Soc.
.Civ. Eng. 89 SM. 2 , 2745.Polemio, M., Rhoades, J.D. 1977,
1977. Determining cation exchange capacity: a new procedure
for calcareous and gypsiferous soils. Soil Sci. Soc. Am. J. 41,
524528.Seed, H.B., Chan, C.K., 1959. Structure and strength
characteristics of compacted clays. J. Soil
.Mech. Found. Div., Am. Soc. Civ. Eng. 85 SM. 5 , 87128.Van
Olphen, H., 1977. An Introduction to Clay Colloid Chemistry.
Wiley-Interscience, New
York.