Applied Clay Science 108 (2015) 6169

Contents lists available at ScienceDirect

Applied Clay Science

journal homepage: www.elsevier.com/locate/clay

Research paper

Stabilization treatment of a dispersive clayey soil using granulated blast

furnace slag and basic oxygen furnace slag
A.R. Goodarzi , M. Salimi
a

Faculty of Engineering, Hamedan Branch, Islamic Azad University, Hamedan, Iran

a r t i c l e

i n f o

Article history:
Received 18 October 2014
Received in revised form 19 February 2015
Accepted 20 February 2015
Available online 6 March 2015
Keywords:
Dispersive clay
Furnace slags
Flocculation
Pozzolanic reactions
Soil stabilization

a b s t r a c t
Dispersive clays may pose considerable distress if not adequately taken care of. On the other hand, the treatment
of problematic soils with waste materials has been recently proved to be a useful option from economic and environmental view point. Hence, in this research, the potential use and effectiveness of dispersive soil stabilization
using two types of industrial by-product, including granulated blast furnace slag (GBFS) and basic oxygen furnace
slag (BOFS) were investigated. The slags were separately added (ranging from 2.5 to 30%) to a laboratory dispersed sample and a set of experiments were performed to study the physicochemical, mechanical and microstructural changes of the stabilized soil. The results indicate that the soil dispersion can be eliminated upon
adding 10% BOFS. This is attributed to the exchange of interlayer sodium ions on the clay surfaces by multivalent
cations from the agent. Besides, an increase in the ion concentration of soil-additive mixtures induces a more depression of the diffuse double layer that decreases the soil dispersivity potential. With increasing the curing time,
an improvement in the strength of composite samples is observed. The formation of cementitious compounds
due to the pozzolanic reactions is responsible for such a treatment, as conrmed by the XRD analyses and SEM
micrographs. It appears that the GBFS has a lower activity as compared to BOFS, therefore causes less inuence
on the soil engineering parameters and a higher percentage (2025%) of GBFS is required to govern soil dispersion. Overall, utilization of the studied slags particularly of BOFS is very effective to overcome the problems associated with dispersive soils.
2015 Elsevier B.V. All rights reserved.

1. Introduction
Dispersive clayey soils are one of the most common problematic
soils that have a wide geographical distribution and exist in several
countries such as Australia, Greece, Iran, South Africa, and the United
States (Gutierrez et al., 2003; Ouhadi and Goodarzi, 2006; Fernando,
2010; Abbasi and Nazi, 2013). The soil dispersibility happens when
the attractive forces (Van der Waals attraction) among the clay particles
become less than the interparticle force of repulsion, resulting in weakening of the soil structure, so that the particles repel each other and go
into suspension when the soil mass is even in contact with still water
(Bauder et al., 2008; Zorluer et al., 2010). The soil dispersion is a complicated physicochemical phenomenon which is affected by the clay mineralogy and pore size distribution. It also depends on the chemistry of
the pore and eroding uids, since this can change the diffuse double
layer (DDL) geometry and electrical charge of the clay particles that
may affect the soil particle arrangement (Yildiz et al., 1999). The previous studies indicated that the presence of active clay minerals, such
as montmorillonite, low electrolyte concentration, and exchangeable
sodium ions contribute to dispersive behavior in soils, whereas ordinary
Corresponding author. Tel.: +98 9163620733; fax: +98 8134494127.
E-mail address: amir_r_goodarzi@yahoo.co.uk (A.R. Goodarzi).

http://dx.doi.org/10.1016/j.clay.2015.02.024
0169-1317/ 2015 Elsevier B.V. All rights reserved.

erosion-resistant clays have a preponderance of Ca2+, Mg2+, and K+

cations (Abend and Lagaly, 2000; Baik et al., 2007). Dispersive clays
are generally very stiff in dry state; however, they can be pulverized
and become structurally unstable when subjected to water. This may
pose considerable distress and severe damages to geotechnical structures that come into contact with dispersive soils or constructed out
of them (Gutierrez et al., 2003). For example, the erosion due to soil
dispersion can cause the formation of piping processes which have
contributed to the failure of 37% of earth dams in different parts of the
world. Such a condition may also negatively affect the self-sealing capacity of the landll clay liners and lead to decreasing the safety performance of geoenvironmental projects (Baik and Lee, 2010).
A simple way to prevent the problems encountered in dispersive
soils is to replace them with suitable materials (Ouhadi and Goodarzi,
2006). This method may have some limitations within an economic
range; while, in many cases chemical modication procedures have
been adopted to overcome deciencies in the performance of problematic soils (Manso et al., 2013; Saride et al., 2013; Aldaood et al.,
2014). Chemical stabilization is usually done by using additives such
as lime, gypsum, and cement. However, in recent years, some disadvantages and the increased cost associated with the use of the traditional stabilizers are leading researchers to replace them with new
alternative soil modiers like industrial waste materials, either for

62

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

economic consideration, resource and environmental conservation or

enhancement of soil engineering characteristics (Cokca et al., 2009;
Abdi, 2011; Harichane et al., 2012). Granulated blast furnace slag
(GBFS) and basic oxygen furnace slag (BOFS) are two types of byproducts resulting from the manufacture of iron that are produced in
large amounts throughout the world (Zhang et al., 2011). As a result
of growing environmental awareness, they are highly regarded as a
recycled material that can reduce impacts on the environment due
to their resource-conservation and energy-saving effects. Utilization
of GBFS and BOFS as supplementary cementitious materials in concrete
technology is most well-known (Zhang et al., 2011; Kim and Lee, 2012).
BOFS is also now used in different civil engineering elds in asphaltic
mixes, road-base layers, and the context of embankment construction
projects (Poh et al., 2006). Al-Rawas (2002) reported that the formation
of aggregations resulted in the reduction of the swelling characteristics
of expansive Al-Khod (in northern Oman) soil treated with GBFS. The
other studies conrmed the positive effects of GBFS on the volume
change behavior, strength and durability of clays (Cokca et al., 2009;
Obuzor et al., 2011); however, based on the literature review, its application for dispersive soil treatment has not been accurately investigated. In
addition, there is a lack of detailed studies on different aspects of BOFS for
modication of clayey soils. Thus, the present research was conducted to
address the potential use and effectiveness of GBFS and BOFS to enhance
the stability of a dispersive soil. For this propose a set of macro and micro
level tests including Atterberg limits, cation exchange capacity (CEC),
double hydrometer, sedimentation, unconned compression strength,
pH value, EC measurement, XRD and SEM were performed to investigate
the mechanisms of soil and additives interactions.
2. Materials and methods
2.1. Materials and sample preparation
The problems associated with dispersive soils have generally occurred in clays that contain predominantly expansive lattice type
minerals such as montmorillonite (Abbasi and Nazi, 2013). Besides,
smectite clay (rich in montmorillonite) is utilized in many geotechnical
and geo-environmental projects (Goodarzi and Akbari, 2014). Thus, in
this research, smectite was used to prepare the potentially dispersive
soil. The mineral identication with X-ray analysis and the chemical
tests show that the used smectite contains high amount of montmorillonite and its main exchangeable cation is sodium. Therefore, it can be
considered as a Na-dominant-montmorillonite which has a high tendency to dispesivity. An articial dispersive soil sample was prepared
by adding different concentrations of NaNO3 solutions to the smectite
at laboratory condition by an earlier described procedure of Ouhadi
and Goodarzi (2006). For this purpose, the following manner was
used: I) the presence of exchangeable sodium ions generally contributes
toward dispersive behavior in soils. Therefore, a liter of soil-electrolyte
mixture with a ratio of 1:50 was prepared by mixing smectite and a
specic concentration of NaNO3. Because the soil dispersion is usually
occurred in low ionic concentration, the range of NaNO3 was varied
within 0.001 to 0.25 Normal. II) The soil-electrolyte mixtures were
kept for 96 h and shake for 24 h to achieve equilibrium. III) The
dispersivity of each prepared mixture was quantied employing the
double hydrometer test ASTM D-4221. For this, the standard hydrometer tests were rst performed to obtain the particle size distribution of
samples. The non-standard tests were then run on those samples with
no chemical dispersant (sodium hexa-metaphosphate) and no mechanical agitation of the solution. The dispersion of mixed samples was identied as the ratio of 5 m particles from the parallel test to the same
sized particles in the standard hydrometer test. If the ratio is greater
than 50%, the soil is classied as dispersive; if it is between 30% and
50%, the soil is moderately dispersive; and if it is less than 30%, the soil
is considered non-dispersive. Therefore, this ratio was considered to
obtain the required concentration of NaNO3 to cause the maximum

dispersivity in the soil samples. Finally, the prepared dispersive smectite

sample with the highest percentage of dispersion was selected in the
rest of the research. The engineering properties of laboratory dispersed
soil sample were then measured according to ASTM methods (ASTM,
2006). Its geoenvironmental characteristics were also determined
using the procedures described in the EPA manual (EPA, 1983). The engineering and geoenvironmental characteristics of the soil sample are
given in Table 1.
GBFS and BOFS produced in Iran Ferroalloys Industries Co. were selected as admixtures to improve the soil engineering properties. Their
chemical compositions were determined using the XRF analysis and
are given in Table 2. As it can be seen, the used by-products have a signicant amount of multivalent cations that can easily exchange the
cations having lower valance on the clay surfaces. The additives were
separately added in amounts ranging from 2.5, 5, 10, 15, 20, and 30%
in dry mass to the dispersed soil sample. Prior to mixing, all the constituents were oven dried and passed through a No. 200 (0.075 mm) sieve.
The mixtures were blended with the needed amount of water for each
test. They were then placed in air-tight plastic bags and cured in a
warm humid chamber at temperature 22 1 C and with a relative humidity of 85%. At the end of each curing period (i.e. 1, 3, 7, and 28 days),
the experiments were performed to investigate the effects of GBFS and
BOFS on the dispersivity and engineering properties of the soil sample.
2.2. Macro and micro level tests
To determine dispersibility of soil different chemical and physical
measurement techniques are commonly used such as pinhole test, double hydrometer test and crumb test. These tests may yield different
results for the same soil samples; hence, more than one test should
be used to identify dispersibility of natural soils. However, as the pinhole and crumb tests are qualitative tests, they may not be useful to
monitor the quantitative impact of different percentages of additives
on the soil dispersion. Besides, the aim here was not to dene the
exact dispersibility of unknown soil samples but to assess how the
chemical stabilizers affect the behavior of soil dispersion when compared with untreated sample. Additionally, since the condition of soil
suspension for performing double hydrometer test is almost similar
to that employed in the sedimentation, EC, pH, XRD, and SEM tests
used in this study, it would be possible to compare the physicochemical
interactions and microstructural changes of treated specimens in
conjunction with the dispersivity potential variations which are calculated by the double hydrometer test. Therefore, in this study, the
dispersivity potential of samples was evaluated by the application of
double hydrometer test. To evaluate the impact of GBFS and BOFS on
the dispersivity behavior, a series of double hydrometer tests were performed on the articially prepared dispersive samples. For this purpose,
several similar suspensions of the soil-electrolyte mixtures in a ratio of

Table 1
Engineering and geo-environmental properties of dispersed smectite sample.
Characteristics

Quantity measured

Mineral composition in decreasing amount

Soil-pH
EC, mS/cm
SSA, m2/g
CEC, cmol/kg
Clay fraction, %
Specic gravity, GS
Liquid limit (LL), %
Plasticity index (PI), %
Soil classication
Maximum dry density, gr/cm3
Optimum moisture content, %
Unconned compression strengtha, MPa

Montmorillonite, quartz, calcite

10.01
2.70
463
85.4
82
2.82
393.7
350.5
CH
1.21
45.5
0.54

a
The soil sample was mixed with the optimum moisture content and compacted at the
maximum dry density.

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

Chemical composition

SiO2
Al2O3
Fe2O3
CaO
MgO
Na2O
K2O
P2O5
SO3
TiO2
Loss of ignition

Percentage in weight (%)

GBFS

BOFS

32.1
16
6.4
27.1
8.8
0.6
0.8
b0.1
3.4
2.9
1.85

15.6
8.2
20.9
48
3.4
0.2
b0.1
0.4
0.3
1.3
1.65

1:50 with a 0.005 normal NaNO3, which had maximum soil dispersion,
were prepared and different percentages of additive were added to
them. After equilibrium conditions were attained, the double hydrometer experiments were conducted on those samples to determine the required amount of percentage additives to control the soil dispersivity. In
addition, to determine the extent of the replacement of sodium ions of
dispersed smectite sample by multivalent cations of the additives, each
dispersive sample with a certain amount of additives was centrifuged
for 45 min at 5000 rpm. Following that, the concentration of Na+ ions
was measured using an Atomic Absorption Spectrophotometer modeled
GBC XplorAA-Dual. To monitor the chemical reactions between the clay
lamellae system and the agent, the soil pH and electrical conductivity
(EC) experiments were also conducted using the method of Eades and
Grim (1966). For this purpose, the pH and EC of the soil-additive slurries
were recorded both immediately (2 h) and after 1, 3, 7, and 28 days of
curing at constant laboratory temperature of 22 1 C.
To take into consideration the effect of chemical admixtures on the
Atterberg limits of treated soil, several samples, which had maximum
dispersivity, were mixed with water contents at the liquid limit and
different amounts of each additive. Following the attainment of equilibrium, the liquid limit and plastic limit tests were conducted in accordance with ASTM D-4318. In many stabilization studies (Al-Mukhtar
et al., 2012; Stoltz et al., 2012; Aldaood et al., 2014), the unconned
compression strength (UCS) value has been used as a primary control
check to evaluate the effectiveness of soil stabilizers. Thus, in this research, the UCS tests were conducted on the samples containing different amounts of the additives. For this propose, the soil/additive mixture
required for the maximum dry density was weighed and a hand mixing
technique was employed to enhance the homogeneity of the sample.
The water needed as optimum moisture content was also added to
the sample. The wet homogenized mixture was then placed inside
cylindrical steel mold with 35 mm in diameter and 70 mm in length.
To ensure uniformed compaction, the samples were subjected to a static
compression force using a hydraulic jack to achieve the desired dry density. After the extrusion of samples from the molds, they were cured as
the mentioned method for the preparation of samples and the UCS tests
were performed following ASTM D-2166, up to failure or 20% strain
under a constant strain rate of 1.2 mm/min.
To monitor the changes of soil structure due to chemical treatment,
the X-ray diffraction response of the laboratory prepared dispersive
smectite samples containing different percentages of GBFS and BOFS
was studied. After the curing period and equilibration, the suspensions
of natural and treated samples poured in closed laboratory asks. Each
volumetric ask was then placed on an ultrasound to adequately disperse the soil before XRD testing. Finally, 4 drops of the solution were positioned on a glass slide by using a micro pipette. The coatings on glass
slides for all samples had similar thicknesses. The glass slides were air
dried and were scanned in the 2 range of 2 to 60 for their XRD spectra.
A Bruker D4 apparatus with Cu-K radiation was used to obtain the
X-ray diffraction patterns of soil samples. Moreover, to further evaluate

the interaction between the additives and the clay particles, the soil samples were subjected to image analysis using scanning electron microscope (SEM). Images of samples were magnied 3000 times by means
of an SEM modeled VEGA3-TESCAN. Air dried pieces of the soil samples
collected from posttest UCS specimens were used for SEM analyses.
3. Results and discussion
3.1. Variation of dispersion potential of smectite sample
In an attempt to investigate the effects of GBFS and BOFS on the soil
dispersivity, an articial dispersive smectite sample was rst prepared
at the laboratory condition. Note that the reason to use articial dispersive samples is to ensure that the mineralogy compositions, type of
exchangeable cations and pore uid chemistry of samples in all of the
tests are similar because these parameters are very effective in soil dispersion and in the case of using natural dispersive samples they may be
different, which can affect the reactions between soil and stabilizer.
Therefore, to avoid confusion in interpreting the results due to the
differences in the samples characteristics and also to allow a better comparison of the test results, especially in the microstructure experiments,
the dispersive samples with similar mineralogy and exchangeable
cations were prepared at the laboratory controlled condition. Moreover,
the problems associated with dispersive soils have generally occurred
in clays that contain predominantly expansive lattice type minerals
such as montmorillonite. Thus, to consider the maximum level of soil
dispersibility, we used articial dispersive samples prepared with Nadominant-montmorillonite clay which have a high tendency for soil
dispersion. Based on the presented results in Fig. 1, the interaction of sodium nitrate with smectite inuences the soil dispersivity by the salt
concentration. Dispersivity is observed to increase up to a NaNO3 concentration of 0.005 N, although it becomes less patent as the concentration increases. This conrms the occurrence of soil dispersion increasing
in the presence of low concentration of sodium ions also reporting in the
previous studies (Abend and Lagaly, 2000; Goodarzi and Akbari, 2014).
In fact, at a small amount of electrolyte containing sodium ions, the edge
to face attraction becomes less than the face to face repulsion, the clay
particles become more hydrated and consequently an increase in the
dispersivity potential occurs. The further addition of salt makes a decrease in the thickness of diffuse double layer (DDL) and reduces the
distance between the clay particles due to increasing the osmotic potential (Ouhadi and Goodarzi, 2006; Yong et al., 2009). Based on the results
in Fig. 1, a suspension of smectite-electrolyte mixture in a ratio of 1:50
with a 0.005 N of NaNO3, which has the maximum percentage of dispersion, was used to prepare the potentially dispersive soil sample. In the
Figs. 1 and 2-a, the dispersivity potential of the soil samples is the
ratio of particles smaller than 5 m from the hydrometer test without
120
Smectite + NaNO 3

100
Dispersivity potential, (%)

Table 2
Chemical compositions of used by-product materials.

63

80
60
40
20
0
0.001

0.0025

0.005
0.01
0.025
0.05
Electrolyte concentration, (N)

0.1

0.25

Fig. 1. Variation of dispersivity potential in smectite sample with different concentrations

of NaNO3.

64

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

75

100

Dispersed soil + GBFS

Dispersed soil + BOFS

60

40

Dispersed soil + BOFS

60

Sodium adsorption, (%)

Dispersivity potential, (%)

Dispersed soil + GBFS

80

45

30

15

20

(b)

(a)
0

0
0

10

15

20

25

30

35

10

Percent of additive, (%)

15

20

25

30

35

Percent of additive, (%)

Fig. 2. Effect of BOFS and GBFS addition on the dispersivity potential and sodium ions adsorption of laboratory dispersed soil sample.

Fig. 2-b, the untreated dispersed sample has 62% adsorbed sodium
ions. Following the application of additives, the amount of adsorbed sodium decreased as the additive content increased. It means that the sodium ions in the clay double layer are replaced by the multivalent
cations (e.g. Ca2+) of the additives. It shows that the addition of 30%
BOFS decreases the adsorbed sodium of the untreated specimen from
62% to almost 5%. The quantity of adsorbed sodium is also decreased
from 62% to 17% with the addition of GBFS until it made up 30% of the
dispersed sample. The reduction in adsorbed sodium indicates that ion
exchange (i.e. the tendency of cations having a higher valance to replace
the exchangeable sodium ions) has an important role to control the
dispersivity behavior of the treated soil. This reaction causes a decrease
in the thickness of DDL, hence leading to the change of dispersed structure to occulated structure as the additive content increases.
At constant laboratory conditions, the variations in sedimentation
behavior of clayey soils with similar mineralogy can be ascribed to the
changes in the clay particles arrangement. Therefore, to evaluate the inuence of additives on dispersivity and micro-structural changes of
samples, a set of sedimentation tests were performed on the specimens.
For this purpose, the suspensions of natural and treated soil in a ratio
of 1:10 (soil:water) were prepared in the 1000 ml laboratory ask
and shaken for 24 h on a horizontal shaker for equilibrium. Following
that, the soil suspensions were poured into a glass cylinder 457 mm in
height and 63.5 mm in diameter, which is generally used for the
hydrometer test. After this, the top of the soil suspension level in each
cylinder was monitored at different intervals of time. The sedimentation

chemical dispersant and mechanical agitation to the same sized particles in the standard hydrometer test expressed as a percentage.
3.2. Effects of GBFS and BOFS addition on the soil dispersivity
Fig. 2-a plots percentage of dispersion variation in the dispersed
smectite after the addition of agents with a short time of curing. It can
be seen that the soil-stabilizers interaction causes a signicant decrease
in the dispersivity as the additive content increases. By adding 10%
BOFS, the dispersion percent dropped below 30% and then treated;
however, a higher percentage (2025%) of GBFS was required to govern
the soil dispersivity. With application BOFS (or GBFS) and due to solubility or partial solubility of them, the concentration of multivalent
cations such as Ca2+ (see Table 2) in the pore uid can be increased. It
is well known that the cations having higher valance can easily substitute the cations having lower valance (Al-Mukhtar et al., 2012; Stoltz
et al., 2012; Saride et al., 2013). Therefore, a base exchange may be occurred with the multivalent cations of additive, replacing the Na+ ions
on the clay surface. This phenomenon, which is a part of short-term
reactions between soil and stabilizer, can result in the formation of
occulated structure (Ouhadi et al., 2014) and reduce the dispersivity
potential as indicated in Fig. 2-a.
The nding about the replacement of the soil exchangeable sodium
ions by multivalent cations from the additive was investigated by
monitoring the variation in the Na+ ions adsorption and sedimentation
performance of the soil-addtive mixtures. According to the results in
Dispersed soil + 2.5%GBFS
Dispersed soil + 10% GBFS
Dispersed soil + 20% GBFS

Dispersed soil + 2.5% BOFS

Dispersed soil + 10% BOFS
Dispersed soil + 20% BOFS

Dispersed soil + 5% GBFS

Dispersed soil + 15% GBFS
Dispersed soil + 30% GBFS

40
30%

20%

15%

10%

5%

Percent o f pa rticle sedimenta tio n (%)

40

Percent o f pa rticle sedimenta tio n (%)

Dispersed soil + 5% BOFS

Dispersed soil + 15% BOFS
Dispersed soil + 30% BOFS

2.5%

35

30
25
20
15

Dispersed samples + GBFS

10
5
0

30%

20%

15%

10%

5%

2.5%

35
30
25
20
15
Dispersed samples + BOFS
10
5
0

0.1

10

100

Log time (minute)

1000

10000

0.1

10

100

Log time (minute)

Fig. 3. Dispersivity behavior of treated dispersed soil with different percentages of GBFS and BOFS.

1000

10000

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

65

400

percentage of mixed samples was identied as the ratio of initial height

of soil water slurry in cylinder minus its nal height to the initial height
of soil water slurry. From Fig. 3, it is observed that the sedimentation
percentage of soil continuously increases as the additive content increases. This behavior can be ascribed to the depression of DDL, allowing
for more aggregations to occur. In fact, by increasing the agents and due
to the replacement of exchangeable sodium ions by multivalent cations
of the additives (see Fig. 2-b), the clay particles get closer to each other
and coagulation of soil particles forms, providing a gradual increase in
the particles size and consequently the soil settle continuously increases
with increasing the additive content. The results also show that addition
of BOFS induces a higher sedimentation as compared to GBFS, which
explains that GBFS has less inuence on soil coagulation. This result is
correlated with a lower reduction in soil dispersion when the sample
is treated by GBFS. Such behavior can be attributed to the lower activity
of GBFS, as will be discussed further on.

Dispersed soil + GBFS

Plasticity index, (%)

350

Dispersed soil + BOFS

300
250

200
150
100
50

10

15

20

25

Fig. 5. Effect of different quantities of the additive types on the soil-PI for short curing time
of 1 day.

3.3. EC and pH variations of the soil-additive mixtures

To more clarify the role of additives for controlling dispersivity behavior, variations in EC values of samples, with and without additives,
were measured and presented in Fig. 4-a. It is evident that the electrical
conductivity from 2.7 mS/cm is increased to 9.75 mS/cm by adding
30% BOFS and the EC of GBFS-soil samples are increased from 2.7 to
5.95 mS/cm with the addition of 30% additive. Such an increase in the
EC level is due to increasing the ions concentration resulting from solubility or partial solubility of the additives. This provides appropriate
condition for increasing the osmotic potential (Poh et al., 2006; Yong
et al., 2009; Abdi, 2011), decreasing the repulsive forces, thus producing
larger particles which have a tendency to settle more rapidly as showed
in Fig. 3. As a result of this process, the treated soil tends to exhibit nondispersive behavior. It means that in addition to the ion exchange, an
increase in the EC level of soil-additive mixtures can allow for more occulation to occur (Al-Mukhtar et al., 2012; Goodarzi and Akbari, 2014)
that tends to reduce the dispersive characteristic of soil. The comparison
of results in Figs. 2-a and 4-a indicates that a chemically stabilized dispersive soil can be identied as non-dispersive if its EC level is higher
than 5 mS/cm. It should be pointed out that the presence of BOFS causes
a higher increase in the EC as compared to GBFS. This is due to lower
solubility of GBFS and consequently to control the soil dispersion a
higher percentage of this additive is required, as indicated in Fig. 2-a.
According to Eades and Grim's criteria (Eades and Grim, 1966), the
strongly alkaline pH (12) condition and appropriate amount of calcium
ions are generally needed to activate the pozzolanic reactions that can
produce cementing products [e.g. calcium-silicate-hydrate (CSH) and
calcium-aluminate-hydrate (CAH)] in stabilized clays, resulting in the
soil parameters improvement (Al-Rawas, 2002; Chen and Lin, 2009;
Saride et al., 2013). Therefore, to investigate the feasibility of pozzolanic

activity occurrence in the soil samples, the immediate effect (after 2 h) of

different quantities of the additive types on the soil-pH values are measured and presented in the Fig. 4-b. It can be seen that at presence of
BOFS the pH increases from 10 to about 12 with 10% BOFS, and reaches
about 13 with the further addition of additives. However, the soil pH increased from 10 to about 12 in the GBFS content up to 30%. This signies
that the pH condition to initiate the long-term reactions (i.e. pozzolanic
activity) is more provided in the case of specimens containing BOFS
that can enhance the efciency of stabilizer (Lemaire et al., 2013). This
result is consistent with the EC tests and conrms that BOFS has a higher
activity as compared to GBFS.
3.4. Effects of GBFS and BOFS on the Atterberg limits of dispersed soil sample
The Atterberg limits have been repeatedly shown to be useful indicators of clayey soils behavior. These limits are usually affected by the pore
uid characteristics like ions concentration and valence of cations
(Goodarzi and Akbari, 2014). Based on this fact and with respect
to the results of EC tests, one may expect various changes in the soil
Atterberg limits due to the addition of BOFS and GBFS. Fig. 5 presents
the changes in Plasticity Index (PI = LL PL) values of the treated
dispersed smectite after a short curing time of 1 day. As it can be seen,
the addition of BOFS is more effective than the GBFS application. For example, the PI of the non-treated dispersed smectite clay (PI = 350%)
declines to about 71% with the addition of 30% BOFS, and to about
42% in specimens containing the same amount of GBFS. This behavior
is possible due to the low dissolved ions in the soil-GBFS mixes samples, proving by the EC tests. In fact, at presence of the BOFS, ions

10

13.5

(a)

Dispersed soil + GBFS

(b)

Dispersed soil + BOFS

12.5

pH

EC, (mS/cm)

11.5

Dispersed soil + GBFS

10.5

Dispersed soil + BOFS

9.5

10

15

20

Percent of additive, (%)

30

Percent of additive, (%)

25

30

35

10

15

20

Percent of additive, (%)

Fig. 4. EC and pH variation of laboratory dispersed smectite sample after treatment with GBFS and BOFS.

25

30

35

66

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

2.5

15

10

20

30

2.5

15

10

30

20

Fig. 6. Effect of curing time and stabilizer content on the compressive strength of both GBFS and BOFS treated dispersed smectite samples.

reactions and bind the clay particles together, and consequently improve
the soil geo-mechanical parameters (Poh et al., 2006; Al-Mukhtar et al.,
2012; Manso et al., 2013). The results indicate that the effect of pozzolanic activity on the soil strength is highlighted at least after 3 days of curing. In the other word, the induced occulation at short-time of curing,
which causes a signicant reduction in dispersivity and soil PI, has no
considerable impact on the geo-mechanical properties and the strength
of treated soil is mainly dictated by the pozzolanic process. It is also interesting to note that the strength of samples with 30% BOFS increases by
nearly 10 times after curing time of 28 days. However, limited changes
occur in the samples tested with 2.5% and 5% of BOFS, regardless of the
curing periods. It means that the low amount of additives is mainly consumed in the short-term reactions and the pozzolanic reactions will be
promoted by the high percentage of BOFS (N5%) or GBFS (N15%).
As illustrated in Fig. 6, the addition of GBFS had little effect on the soil
strength as compared to BOFS, suggesting that the BOFS has a higher activity to produce pozzolanic compounds that generate the long-term
strength gain. To support this nding, the EC variations in treated samples at different curing periods were also measured. Fig. 7 indicates that
the addition of BOFS and GBFS caused an increase in the EC values; however, curing had a lower effect of the EC value of soil-GBFS mixtures. It is
well known that the ion exchange occurs in a short time of curing, and

concentration in the soil-electrolyte signicantly increases that can reduce repulsive forces between the clay particles, causing the soil skeleton to shrink, thus promoting occulation and decreasing in the water
retention capacity (Aldaood et al., 2014). On the other hand, based on
the presented results in Fig. 2-b, a more replacement of sodium ions in
the dispersed sample by multivalent cations of BOFS and consequently
a higher reduction in thickness of DDL can also be attributed to a noticeable decrease in the plasticity of soil-BOFS mixtures.
3.5. Effects of GBFS and BOFS on the mechanical capacity of dispersed soil
sample
To investigate the effect of the by-product materials on the geomechanical behavior of dispersive soil, the strength values of composite
samples cured for 1, 3, 7, and 28 days were measured using the UCS tests.
As it can be observed in Fig. 6, the strength values increase with increasing time of curing. Based on the literature review (Cokca et al., 2009;
Obuzor et al., 2011; Stoltz et al., 2012), the observed increase in the mechanical capacity can be explained due to developing long-term reactions (i.e. pozzolanic activity) and forming of cementing compounds,
as will be evaluated by XRD and SEM tests. The new produced materials
can occupy an increasing amount of void space due to the short-term

10

10
2.5% BOFS

10% BOFS
20% BOFS

2.5% GBFS

5% GBFS

15% BOFS

10% GBFS

15% GBFS

20% GBFS

30% GBFS

30% BOFS

(a)
6

EC (mS/cm)

EC (mS/cm)

5% BOFS

(b)
6

2
1

10

100
Log time (hour)

1000

10000

10

100
Log time (hour)

Fig. 7. Effect of curing time on the EC values of both GBFS and BOFS treated dispersed smectite samples.

1000

10000

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

after this, the pozzolanic reactions initiate, which can slowly consume
the additive to form CSH and CAH gels. Therefore, based on the EC
tests, the development of soil properties due to this second level reaction can signicantly take place in the BOFS-treated samples, which is
consistent with the results of UCS tests.

1700
1500

1300

Peak intensity, cps

67

1100

3.6. Effects of additives on the micro-structural performance of dispersed

soil sample

900
700
500

The XRD and SEM were used to precisely investigate the results from
a micro-structural view point. Generally, it is expected that the characteristics of XRD diagrams of clays particularly the intensity of major
basal spacing peaks will be inuenced by the nature of pore uid and
type of chemical stabilizers (Chen and Lin, 2009; Lemaire et al., 2013;
Aldaood et al., 2014; Goodarzi and Akbari, 2014). Therefore, variations
of the peak intensity for the major basal spacing of treated and natural
dispersed smectite samples (i.e. d001 = 12.4 ) are determined from
the results of the XRD tests and reported in Fig. 8. As described in the

Dispersed soil + GBFS, after 1 day of curing

300

Dispersed soil + BOFS, after 1 day of curing

100
1

10

100

Additive content, (%)

Fig. 8. Effect of additives on the peak intensity of the major basal spacing of dispersed soil
(i.e. d001 = 12.4 ).

(c)

(b)

(a)

250

(d)

GBFS-treated soil, after 1 day of curing

GBFS-treated soil, after 3 days of curing
GBFS-treated soil, after 7 days of curing
GBFS-treated soil, after 28 days of curing

200

CSH peak intensity, cps

CSH peak intensity, cps

250

150

100

50

(e)

BOFS-treated soil, after 1 day of curing

BOFS-treated soil, after 3 days of curing
BOFS-treated soil, after 7 days of curing
BOFS-treated soil, after 28 days of curing

200

150

100

50

10

15

20

25

Additive content, (%)

30

35

10

15

20

25

30

35

Additive content, (%)

Fig. 9. XRD analysis; (a) laboratory dispersed smectite, (b) treated dispersed soil with 30% GBFS after 28 days of curing, (c) treated dispersed soil with 30% BOFS after 28 days of curing,
(d) CSH peak intensity variation of dispersed soil treated with GBFS, and (e) CSH peak intensity variation of dispersed soil treated with BOFS.

68

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

materials and methods section, the samples were prepared on the glass
slides and then were air-dried for performing the XRD tests. It is obvious
that at a short run time of 1 day, the addition of BOFS or GBFS leads to a
decrease in the Na-montmorillonite mineral's peak intensity. This is because the clay particles with increasing the concentration of additive exhibit a much-occulated structure, as was previously explained. In such
a case, those occulated particles show lesser peak intensities owing to
a decrease in the reection of incident ray as compared to the dispersed
structure (Goodarzi and Akbari, 2014). Fig. 8 illustrates a relative decrease in the montmorillonite mineral's peak intensity with increasing
the additives contents, thereby reducing the mineral's capacity for
dispersion as indicated in Fig. 2-a. The results show that the peak intensity for the major basal spacing of dispersed sample decreases from
1680 counts/s in its initial condition to 1410 and 650 counts/s in the
presence of 5% and 30% BOSF, respectively. A similar pattern is also
observed in the case of GBFS. However, at the same amount of slag, a
higher reduction in the peak intensity values of the soil-BOFS mixtures
occurs in comparison with that of the samples having GBFS, which is
consistent with the better performance of BOFS on the reduction of
dispersivity potential.
With increasing the curing time, the XRD patterns show a more decrease in the major peak intensity of montmorillonite and illustrate the
formation of new peaks characteristic particularly of calcium silicate
hydrates (CSHs), consistent with other researchers (Song et al., 2000;
Yi et al., 2014) who suggested the main hydration products of GBFS
were CSH. An example of the X-ray graphs for natural and treated smectite with 30% additives and 28 days of during is depicted in Fig. 9. It
can be seen that the addition of additives causes an increase in some
basal spacing and a decrease in the major XRD peak intensity of Namontmorillonite (i.e. d001 = 12.4 ). As mentioned above, the decrease
in the major peak intensity after addition of additives can be attributed
to the formation of occulated structures due to short-term reactions.
This can also be explained by dissolving the clay fractions due to the

pozzolanic reactions (Ouhadi et al., 2014). Besides, the increase in the

basal spacing of 13.7 and 15.04 can be ascribed to the substitution of
multivalent ions of agents with monovalent exchangeable Na+ ions of
dispersed soil sample (Ouhadi and Goodarzi, 2006).
The XRD results in Fig. 9 and the variation of CSH peak intensity (the
increase in the intensity of peaks at 2 about 29) in Fig. 9(d) and (e) indicate that the intensity of CSH peak (i.e. the formation of cementing
materials) starts to form after the curing time up to 3 days and with
the addition of at least 5% BOFS or 15% GBFS. Such a result is considerably
consistent with the UCS tests which describe only with a high percentage
of additives a great improvement in the soil strength is occurred. In addition, the results presented in Fig. 9 indicate that the GBFS-treated samples show less rate of the appearance of the new peak with increasing
in additive content than those of soil-BOFS mixtures. It means that a
more pozzolanic activity occurs in the latter samples. The observations
can be explained as follows: The higher presence of the soluble ions
(e.g. OH, Ca2+ and Fe2+) in BOFS produces a higher increase in the
soil pH and EC value (Fig. 4). This condition causes a further dissolution
of the clay fractions in the soil-BOFS mixtures, which results in a higher
reduction in the montmorillonite mineral's peak intensity (Fig. 8). Consequently, the more dissolved Si and Al ions in the BOFS-treated samples
and the interaction of them with the calcium ions leads to a higher process of pozzolanic reactions and strength gain in this case (Fig. 6).
To further evaluate the interaction between the clay particles
and clarify the formation of cementitious compounds, the natural and
treated soil samples were subjected to image analysis using scanning
electron microscopy (SEM). Fig. 10 shows the SEM micrographs of
dispersed smectite (Fig. 10-a), dispersed smectite with 30% GBFS and
28 days of curing (Fig. 10-b) and dispersed smectite with 30% BOFS
and 28 days of curing (Fig. 10-c) taken at the same magnication. As it
can be seen, the natural sample has dispersed structure and displays
typical morphology of the montmorillonite clay, consisting thin wavy
sheets. On the other hand, at the presence of BOFS and GBFS, the soil

(a)

(b)

Flocculated structure

(c)

Flocculated structure
Cementation products

Cementation products

Fig. 10. SEM micrographs; (a) laboratory dispersed smectite sample, (b) treated dispersed soil with 30% GBFS after 28 days of curing, and (c) treated dispersed soil with 30% BOFS after
28 days of curing.

A.R. Goodarzi, M. Salimi / Applied Clay Science 108 (2015) 6169

fabric transforms from a particle based form to a more integrated

composition. This textural event causes a signicant decrease in the
dispersivity behavior as indicated in Figs. 2 and 3. With addition
of GBFS, the micrograph (Fig. 10-b) provides visual evidences of occulated structure and shows the formation of patches of cementing
products. This phenomenon is more prominently observed in case of
BOFS treatment (Fig. 10-c) that illustrates its greater effect on the
micro-structural performance of dispersed smectite samples as also
conrmed by the macro level tests.
4. Conclusion
- Dispersive clays can be stabilized satisfactorily using granulated
blast furnace slag (GBFS) and basic oxygen furnace slag (BOFS).
Such an application can overcome the soil dispersion, cause a decrease in the plasticity index, and improve the compressive strength,
particularly in the case of BOFS treatment.
- At a short curing age, the role of BOFS and GBFS for controlling
dispersivity can be mainly ascribed to the ion exchange phenomenon. This can also be explained by increasing the EC value as the additive content increases, causing a more occulated structure and
producing larger particles that have a lower tendency to erosion. It
appears that a chemically stabilized dispersive soil may be identied
as non-dispersive if its EC level is higher than 5 mS/cm.
- The induced occulation owing to the short-time reactions has
no considerable impact on the soil mechanical capacity. However,
an increase in the strength of composite samples is observed with
the progression of the reactions. This is due to the formation of
cementing minerals as conrmed by the XRD and EC tests. It was
found that only with a high percentage of BOFS (N 5%) or GBFS
(N15%), these new materials are formed sufciently large in quantity
to develop soil properties.
- The SEM micrographs also support the formation of new cementing
compounds and reveal a reorganization of oriented structure to occulated structure in the treated soil samples, which are consistent
with the observed diminution in the XRD peak intensity of the clay
mineral.
- From a practical point of view, an optimum BOFS content of 10% with
curing time of 7 days is recommended to overcome the deciencies in dispersive soils performance; however, a higher percentage
(2025%) of GBFS is necessary to govern the dispersivity potential.
Acknowledgments
The authors wish to acknowledge the nancial support was provided by the deputy of Research of Islamic Azad University, Hamedan
Branch.
References
Abbasi, N., Nazi, M.H., 2013. Assessment and modication of sherard chemical method
for evaluation of dispersion potential of soils. Geotech. Geol. Eng. 31, 337346.
Abdi, M.R., 2011. Effects of basic oxygen steel (BOS) slag on strength, durability and plasticity of kaolinite. Int. J. Civ. Eng. 9, 8189.
Abend, S., Lagaly, G., 2000. Solgel transitions of sodium montmorillonite dispersions.
Appl. Clay Sci. 16, 201227.

69

Aldaood, A., Bouasker, M., Al-Mukhtar, M., 2014. Geotechnical properties of lime-treated
gypseous soils. Appl. Clay Sci. 8889, 3948.
Al-Mukhtar, M., Khattab, S., Alcover, J.F., 2012. Microstructure and geotechnical properties
of lime-treated expansive clayey soil. Eng. Geol. 139, 1727.
Al-Rawas, A.A., 2002. Microfabric and mineralogical studies on the stabilization of an
expansive soil using cement by-pass dust and some types of slags. Can. Geotech. J.
39, 11501167.
ASTM, 2006. Annual book of ASTM standards. vol. 04.08. American Society for Testing
and Materials, Philadelphia.
Baik, M.H., Lee, S.Y., 2010. Colloidal stability of bentonite clay considering surface charge
properties as a function of pH and ionic strength. Ind. Eng. Chem. 16, 837841.
Baik, M.H., Cho, W.J., Hahn, P.S., 2007. Erosion of bentonite particles at the interface of a
compacted bentonite and a fractured granite. Eng. Geol. 91, 229239.
Bauder, J.W., Hershberger, K.R., Browning, L.B., 2008. Soil solution and exchange complex
response to repeated wetting-drying with modestly salinesodic water. Irrig. Sci. 26,
121130.
Chen, L., Lin, D.F., 2009. Stabilization treatment of soft subgrade soil by sewage sludge ash
and cement. J. Hazard. Mater. 162, 321327.
Cokca, E., Yazici, V., Ozaydin, V., 2009. Stabilization of expansive clays using granulated
blast furnace slag (GBFS) and GBFS-cement. Geotech. Geol. Eng. 27, 489499.
Eades, J.L., Grim, R.E., 1966. A quick test to determine lime requirements of lime stabilization. Highw. Res. Rec. 139, 6172.
EPA, 1983. Process design manual: land application of municipal sludge. Res. Lab. EPA625/1-83-016.
Fernando, J., 2010. Effect of water quality on the dispersive characteristics of soils found in
the Morwell Area, Victoria, Australia. Geotech. Geol. Eng. 28, 835850.
Goodarzi, A.R., Akbari, H.R., 2014. Assessing the anion type effect on the hydro-mechanical
properties of smectite from macro and micro-structure aspects. Geomech. Eng. 7,
183200.
Gutierrez, F., Desirm, G., Gutierrez, M., 2003. Causes of the catastrophic failure of an earth
dam built on gypsiferous alluvium and dispersive clays (Altorricon, Huesca Province,
NE Spain). Environ. Geol. 43, 842851.
Harichane, K., Ghrici, M., Kenai, S., 2012. Effect of the combination of lime and natural
pozzolana on the compaction and strength of soft clayey soils: a preliminary study.
Environ. Earth Sci. 66, 21972205.
Kim, H.K., Lee, H.K., 2012. Effects of high volumes of y ash, blast furnace slag, and bottom
ash on ow characteristics, density, and compressive strength of high-strength
mortar. J. Mater. Civ. Eng. 25, 662665.
Lemaire, K., Deneele, D., Bonnet, S., Legret, M., 2013. Effects of lime and cement treatment
on the physicochemical, microstructural and mechanical characteristics of a plastic
silt. Eng. Geol. 166, 255261.
Manso, J.M., Ortega-L pez, V., Polanco, J.A., Setin, J., 2013. The use of ladle furnace slag in
soil stabilization. Constr. Build. Mater. 40, 126134.
Obuzor, G.N., Kinuthia, J.M., Robinson, R.B., 2011. Enhancing the durability of ooded lowcapacity soils by utilizing lime-activated ground granulated blast furnace slag
(GGBS). Eng. Geol. 123, 179186.
Ouhadi, V.R., Goodarzi, A.R., 2006. Assessment of the stability of a dispersive soil treated
by alum. Eng. Geol. 85, 91101.
Ouhadi, V.R., Yong, R.N., Amiri, M., Ouhadi, M.H., 2014. Pozzolanic consolidation of stabilized soft clays. Appl. Clay Sci. 95, 111118.
Poh, H.Y., Ghataora, G.S., Ghazireh, N., 2006. Soil stabilization using basic oxygen steel slag
nes. J. Mater. Civ. Eng. 18, 229240.
Saride, S., Puppala, A.J., Chikyala, S.R., 2013. Swell-shrink and strength behaviors of lime
and cement stabilized expansive organic clays. Appl. Clay Sci. 85, 3945.
Song, S., Sohn, D., Jennings, H.M., Mason, T.O., 2000. Hydration of alkali-activated ground
granulated blast furnace slag. J. Mater. Sci. 35, 249257.
Stoltz, G., Cuisinier, O., Masrouri, F., 2012. Multi scale analysis of the swelling and shrinkage of a lime-treated expansive clayey soil. Appl. Clay Sci. 61, 4451.
Yi, Y., Liska, M., Al-Tabbaa, A., 2014. Properties of two model soils stabilised with different
blends and contents of GGBS, MgO, lime and PC. J. Mater. Civ. Eng. 26, 267274.
Yildiz, N., Sarikaya, Y., Calimli, A., 1999. The effect of the electrolyte concentration and pH
on the rheological properties of the original and the Na2CO3-activated ktahya bentonite. Appl. Clay Sci. 14, 319327.
Yong, R.N., Ouhadi, V.R., Goodarzi, A.R., 2009. Effect of Cu2+ ions and buffering capacity
on smectite microstructure and performance. Geotech. Geoenviron. Eng. 135,
19811985.
Zhang, T., Yu, Q., Wei, J., Li, J., Zhang, P., 2011. Preparation of high performance blended
cements and reclamation of iron concentrate from basic oxygen furnace steel slag.
Resour. Conserv. Recycl. 56, 4855.
Zorluer, I., Icaga, Y., Yurtcu, S., Tosun, H., 2010. Application of a fuzzy rule-based method
for the determination of clay dispersibility. Geoderma 160, 189196.