Proceedings of the Institution of
Civil Engineers
Construction Materials 163
August 2010 Issue CM3
Pages 157–165
doi: 10.1680/coma.2010.163.3.157
Paper 900024
Received 01/04/2009
Accepted 18/06/2009 Rahmat Mohamad John Mungai Kinuthia
Keywords: Nidzam Senior Lecturer,
roads & highways/recycling & reuse Visiting researcher, Division of Civil
of materials/strength and testing of Division of Civil Engineering, University
materials Engineering, University of Glamorgan,
of Glamorgan, Pontypridd, UK
Pontypridd, UK
Sustainable soil stabilisation with blastfurnace slag – a review
R. M. Nidzam MSc, PhD and J. M. Kinuthia
This paper reviews a relatively recent approach towards
sustainable soil stabilisation, by utilising ground granu-
lated blastfurnace slag, an industrial by-product from
steel manufacture. It describes in detail the reactions and
reaction mechanisms that occur when soil modification
and subsequent stabilisation processes take place. The
effects of incorporating slag in the soil stabilisation
process on the engineering properties of a soil are also
reviewed and discussed. The review demonstrates the
great potential of slag in soil stabilisation, in particular of
the sulfate-bearing soils whose stabilisation using the
traditional stabilisers of lime and/or Portland cement
leads to excessive swelling.
1. INTRODUCTION
Soil stabilisation may be defined as the alteration of the
properties of an existing soil to meet specified engineering
requirements. The main properties that may require to be altered
by stabilisation are
(a) strength – to increase the strength and thus stability and
bearing capacity
(b) volume stability – to control the swell–shrink characteristics
caused by moisture changes
(c) durability – to increase the resistance to erosion, by
weathering or traffic usage
(d) permeability – to reduce permeability and hence the passage
of water through the stabilised soil (Sherwood, 1993). There
may also be secondary requirements, such as short-term
material changes including plasticity, bulking and water
reduction.
Soil stabilisation is widely used in road construction to improve
sub-bases and sub-grades, for rail, road and airport construc-
tion, for embankments, as soil exchange in unstable slopes, as
backfill for bridge abutments and retaining walls, as landfill and
canal linings, for improvement of soil beneath foundation slabs
and for lime piles.
Traditionally, the use of lime in soil stabilisation precedes the
beginning of clearly recorded history. Lime has been success-
fully employed in many countries, particularly in warm
countries because it needs a relatively high temperature to react
with the clay. In the UK, lime stabilisation was first used in the
construction of the A38 in Worchester in 1951 (Brook-Bradley,
Construction Materials 163 Issue CM3
MSc, PhD, MCIHT, FHEA
1952), where part of the sub-base was stabilised with hydrated
lime. Other subsequent examples of lime stabilisation on a large
scale in the UK include the construction of the Oxford–
Birmingham part of the M40 motorway in the late 1980s. Here,
problems of excessive swelling of lime stabilisation of sulfate-
bearing clay soils were exposed. Research work following this
experience has shown that the use of ground granulated
blastfurnace slag (ggbs) can eliminate the expansive tendencies
of lime stabilisation of the sulfate-bearing clay soils (Wild et al.,
1996, 1998).
Higgins and Kennedy (1999), in their report on the utilisation of
ggbs in road construction, summarise research in the UK and
practical experiences relating to the use of ggbs as an industrial
by-product that is readily available in the UK, and its
combination with lime for soil stabilisation. Laboratory research
at the University of Glamorgan and subsequent field trials on the
A421 Tingewick bypass in Buckinghamshire (Higgins, 1998)
have both confirmed that the combination of ggbs and lime are
practical and effective options for soil stabilisation, and provide
technical, economic and environmental benefits. In particular,
the incorporation of ggbs is very effective in combating the
expansion associated with the presence of sulfates or sulfides in
lime and/or cement-stabilised soil.
This review paper encapsulates the entire development of
research involving ggbs in soil stabilisation in one general
document, including the reactions and reaction mechanisms
that are involved. The technology is of interest to the modern
civil engineer who would like to modernise and move on from
the traditional or classical soil stabilisation practices using
traditional stabilisers of lime and/or Portland cement (PC). The
fact that both performance and care for the environment are
addressed by the utilisation of ggbs makes this contribution
more significant.
1.1. Manufacture, properties and utilisation of ggbs
Blastfurnace slag is produced from iron blast furnaces as a by-
product of the iron-making industry. It results from the fusion of
a limestone flux with ash from coke and the siliceous and
aluminous residue remaining after the reduction and separation
of the iron from the ore. Iron blastfurnace slag consists
essentially of silicates and alumino-silicates of lime and other
bases (Lee, 1974), and has potential cementitious reactivity.
Molten blastfurnace slag has a temperature of 1300–1600 ˚C and
is chilled very rapidly to prevent crystallisation. The granulated
Sustainable soil stabilisation with blastfurnace slag – a review Nidzam N Kinuthia 157
 material produced is known as granulated blastfurnace slag
(GBS). During the manufacture of ground slag, slag granules are
ground into a fine powder (ggbs) similar in fineness to PC
(specific surface of ggbs is typically within the range of 320–
380 m2/kg) and with specific gravity in the range of 2?85–2?94
(PC is approximately 3?15). As opposed to granulation, the
molten slag may also be made to undergo a pelletisation process,
although the pellets are not as reactive as the granulated slag. In
both cases a latently hydraulic material that can be activated
using a variety of compounds such as lime, alkalis and PC is
formed. The activated material can then hydrate to produce
hydration products (Gupta and Seehra, 1989). The latent
hydraulic properties of blasfurnace slag were discovered in
Germany in 1862 (Bijen, 1996). The latent hydraulicity means
that, once activated, the slag reacts with water to give
cementitious products. In the UK, the potential of blastfurnace
slag was first realised in Scotland in 1914 when GBS was inter-
ground with PC.
ggbs can be incorporated in cementitious materials to modify
and improve certain properties and also in order to conserve
non-renewable natural resources by utilising industrial by-
products (Nixon and Gaze, 1981). The possibility of recycling or
processing materials for use as partial replacement of cement in
concrete, or to stabilise soils, has great economic benefits in all
areas of the construction industry (Wild and Tasong, 1999).
The use of ggbs is well established in many cement applications
where it provides enhanced durability, high resistance to
chloride penetration, resistance to sulfate attack and protection
against alkali silica reaction (ASR). Its use in soil stabilisation is,
however, still a novel process in the UK although it has been
used in parts of Europe, South Africa and India. The well-
established sulfate-resisting properties imparted to cements by
blending them with ggbs suggests that by blending lime with
ggbs, the latter may impart similar sulfate-resisting properties to
lime-stabilised clay (Wild et al., 1996). In both PC hydration and
hydration of mixtures of sulate-containing clay mixes and PC,
the phases present are very similar: ettringite, calcium–silicate–
hydrate, calcium–aluminate–hydrate and calcium–aluminate–
silicate–hydrate (C–S–H, C–A–H and C–A–S–H) gels. As PC
hydration products are practically the same as those of PC-ggbs
blends, it is likely also that those of lime–ggbs–clay are very
similar or at least comparable to those of PC–ggbs blends
(Kinuthia, 1997; Regourd, 1980; Smolczyk, 1980). In order to
understand how the addition of ggbs alters soil properties, the
properties and hydration of ggbs should first be discussed in
some detail.
1.2. Activation and hydration of ggbs
ggbs used on its own shows only minimal hydration. Caijun and
Day (1993) studied the hydration of Canadian ggbs and found
that when ggbs is in contact with water, a rich layer of Si–Al–O
forms on the ggbs particle surfaces. This layer could absorb H+
ions resulting in an increase in OH2 ions and in the pH of the
solution, but this was insufficient to break the Si–O and Al–O
bonds to allow formation of C–S–H, C–A–H or C–A–S–H
hydration compounds. They also found that only a small
amount of C–S–H was formed after 150 days of moist curing.
For this reason, utilisation of ggbs is based on its activation to
break this stagnation.
158 Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review
The most common method used to activate ggbs is by way of
chemical activation. In chemical activation, an activator and/or
an alkaline medium are required. Several activators have been
suggested in the literature. Calcium hydroxide, calcium sulfate,
standard PC, sodium hydroxide, sodium carbonate, sodium
sulfate and sodium silicate are among the most commonly cited
activators (Gjorv, 1989; Gupta and Seehra, 1989). It is difficult
to confidently explain exactly how the activator works on the
ggbs. Calcium sulfate, for example, is not only a successful
activator but also plays an important role as a reactant (Daimon,
1980; Taylor, 1990). Ouf (2001), in his work on the stabilisation
of clay sub-grade soils using ggbs, reported that a reactant
participates significantly in the reaction process while an
activator creates an appropriate environment for the reaction
process without necessarily playing a significant role in the
reaction. Wu et al. (1990), in work on early activation and
properties of slag cement, suggested that sodium hydroxide,
sodium sulfate and potassium aluminium sulfate used as
activators for ggbs help in breaking the Si–O and Al–O bonds.
The most commonly used activator for ggbs is, however, PC. The
reactions of PC with ggbs and water are also complex. Water
hydration of PC produces mainly calcium hydroxide (Ca(OH)2)
and C–S–H gel. In the hydration of a blended PC–ggbs system,
although minor amounts of alkalis are released, ggbs is mainly
activated by the hydration product Ca(OH)2 (Bijen, 1996;
Hakkinen, 1993). Therefore, lime in the form of Ca(OH)2 may be
added either as an additive or expected from that released from
PC hydration. Due to its high alumina and silica content, PC–
ggbs hydration produces slightly different hydrates from those
formed by the hydration of standard PC. As the main reaction
products of ggbs hydration are C–S–H gel, C–A–H gel and small
amounts of calcium hydroxide (Higgins et al., 1998), the PC–
ggbs system is not significantly different from either the PC or
activated ggbs systems.
In the hydration of alkali-activated ggbs, however, CH is absent
and the hydration products are mainly C–S–H and hydrotalcite-
type phases containing gypsum (Tasong et al., 1999; Wild et al.,
1998). Song et al. (2000) reported the formation of hydrotalcite-
type phases in higher pH pastes along with C–S–H. They also
observed that the pH of the mixing solution may affect the
nature of C–S–H and its Ca–Si ratio. In addition to C–S–H and
C–A–H, ettringite is also a principal early hydration product in
hydrated PC–ggbs blends. The formation of Ca(OH)2 during PC
hydration in hydrated PC–ggbs system produces an alkaline
environment suitable for the dissolution of Al2O3 and SiO2 from
the ggbs and/or any other sources in the reacting system. In the
presence of Ca(OH)2, CaSO4 reacts with Al2O3 to form ettringite
(C3A.3CaSO4.32H2O). Because of the consumption of both
Ca(OH)2 and CaSO4 in the activation of ggbs, the availability of
these components is subsequently suppressed and hence affects
the continued formation of ettringite in a ggbs-containing
system.
Tasong et al. (1999) Wild et al. (1998) and Wild Tasong (1999)
all used lime as an activator in studies into the influence of ggbs
on the sulfate-resistance of lime-stabilised kaolinite. They
observed that the optimum lime–ggbs ratio was 1 : 5 to activate
the ggbs and to prevent attack caused by excess sulfate solution.
They also observed that the lime-activated ggbs hydration
reaction was quicker than the pozzolanic reaction of lime with
Nidzam N Kinuthia
clay. The researchers argued that, due to its high alumina and
silica content, the main reaction products of ggbs activated by
lime are a complex C–A–S–H gel intermixed with hydrotalcite-
type phases containing magnesium.
Douglas et al. (1991) observed that 3% by weight of hydrated
lime can retard the setting time of alkali-activated ggbs
concrete. This would appear to agree with Wild et al. (1998) who
established that only a small proportion of lime is needed for
optimal activation of ggbs (e.g. 1 : 5 lime–ggbs) depending on
the system. Macphee et al. (1989) found that the hydraulic
reactions of ggbs are slower than the hydration of PC and have a
‘pore-blocking’ effect which leads to a greater ultimate strength
and lower permeability. The pore-blocking effect is partly due to
increased gel formation in the system. The foil-like gel
morphology common with ggbs hydration, as opposed to the
fibrillar gel morphology common with PC systems (Taylor,
1990), also enhances the impermeability of the ggbs-containing
systems. Talling (1989) using X-ray diffraction, studied lime–
alkali-activated ggbs. He identified the presence of C4AH13 as
well as the formation of C–S–H. Thus, the wide variety of
hydration products, the change in morphology, together with
the reduced Ca(OH)2 and other improved binding and absorptive
effects, all combine to enhance the resistance of ggbs systems to
sulfate attack.
The effectiveness of ggbs hydration depends on many other
factors. These include the chemical composition of the ggbs,
alkali concentration of the reacting system, glass content,
fineness of ggbs, and temperature, especially during the early
stages of the hydration process (Kinuthia, 1997). The initial
reactions during ggbs hydration produce a porous crystallised
hydrated layer of coatings of alumino-silicate products on the
surfaces of the ggbs particles within a short time of exposure to
water. These surface layers are impermeable to water, inhibiting
further hydration reactions (Daimon, 1980; Richardson et al.,
1994). The hydration products of ggbs have also been found to
be relatively more crystalline than the hydration products of PC,
and so add density to the blended cement paste (Smolczyk,
1980; Taylor, 1990).
2. CLAY–LIME–GGBS REACTIONS
The introduction of ggbs into the lime–clay hydration will
undoubtedly modify the typical lime clay reaction products. The
reactions in a ternary clay–lime–ggbs mixture are therefore
different and more complex relative to the clay–lime reactions
occurring in classical lime stabilisation of soils. There are three
competing reactions rather than one, and their kinetics are
governed by the same factors observed in the relatively more
researched PC–ggbs blended hydration. These factors include
water content, curing environment, the properties of the C–S–H
gel (such as its amount, porosity and permeability) the fineness
and composition of all reactants involved, the efficiency of
mixing, the temperature and the curing period, among other
factors.
In the ternary clay–lime–ggbs mix, the lime remaining after the
initial consumption of lime in the clay–lime reaction provides
the required alkaline environment for the ggbs activation and
hydration. Thus, the strength increase observed in hydrated PC–
ggbs blends relative to the strength when the PC is used on its
own is also likely to take place in clay–lime–ggbs mixes since
Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review
the ggbs, in a similar manner, will also introduce extra and more
freely available alumina and silica. The enhanced formation of
the strength-contributing silicate and aluminate hydrates, in
addition to the products inherent from the hydration of the
clay–lime system, results in enhanced strength-development.
For this reason, the partial replacement of lime by ggbs implies
that using a blended lime–ggbs system instead of lime only
would lead to a reduction in the total binder content needed to
achieve a given strength requirement. However, the degree of
lime replacement should not exceed a certain proportion in
order to maintain the critical minimum lime content that is
sufficient to fully activate the ggbs as well as facilitating the
clay–lime pozzolanic reactions.
In order to effectively unravel the reactions involved in the
complex ternary clay–lime–ggbs system, it is perhaps best to
first identify and discuss the separate reactions that are
obviously identifiable and that would inevitably take place
between any two of the material participants
(a) clay–lime reactions
(b) lime–ggbs reactions
(c) clay–ggbs reactions.
2.1. Clay–lime reactions
When quicklime is added to a clay soil in an aqueous medium,
the following lime hydration reactions occurs
1 CaO z H2 O ? CaðOHÞ2 z: Heat
The heat produced during this reaction is sufficient to drive off
some of the moisture within the soil as water vapour, which
reduces the moisture content of the soil. With either quicklime
or hydrated lime, the following lime dissociation further takes
place
2 CaðOHÞ2 ? Ca2z z 2OH{
After the formation of the positively charged calcium cations
from either quicklime or hydrated lime, an immediate cation
exchange begins to take place between the metallic ions
associated with the surfaces of the clay particles and the calcium
cations from the lime. The effects of the cation substitution
result in a compressed double layer of cations on the clay
particle surfaces (Kinutha et al., 1999). This subsequently leads
to a reduction in plasticity as the clay particles flocculate and
then agglomerate. These processes occur rapidly after addition
of lime to the clay, within the initial 24 h. The flocculation–
agglomeration process is responsible for the modification of the
engineering properties of a treated soil (Kinutha et al., 1999;
Sherwood, 1993).
After modification, the more long-term soil stabilisation takes
place. Lime stabilisation can be defined as the reaction between
silica and alumina within the clay structure on the one hand,
and lime and water on the other, to form C–S–H, C–A–H and C–
A–S–H gels, which subsequently crystallise to bind the entire
matrix together. As a result of these reactions the stabilised
material becomes stronger and more brittle (Arabi and Wild,
Nidzam N Kinuthia 159
 1989; Croft, 1964; Diamond and Kinter, 1966; Rogers et al.,
1997; Sloane, 1965). Strength gain increases with the length of
the curing period, and is attribut the progressive dissolution of
SiO2 and Al2O3 as the reactions continue to form C–S–H gel and
a wide range of compounds of varying crystallinity (Croft,
1964).
The reactions between lime, water, silica and alumina in clay to
form cementitious products are referred to as the soil–lime
pozzolanic reactions. The actual reaction products formed vary
depending on the type of clay, temperature, lime content and
curing period, although it is generally recognised that of all the
hydration products formed, the principal cementitious product is
a C–S–H gel (Arabi and Wild, 1989; Bell and Tyrer, 1987; Cobbe,
1988). Bell (1988) hypothesised that the long-term strength
development of lime–clay material may be attributed to the
gradual crystallisation of C–S–H gel forming an interlocking
structure, although this is disputed by Wild et al. (1998) who
argued that an impermeable pore filling gel is formed rather
than a crystalline material. However, it is indisputable that small
amounts of crystalline C–A–H phases (such as C4AH13 and
C3AH6) and C–A–S–H phases (for example C2ASH8) form in
some circumstances, particularly in lime-stabilised kaolinite
which has a high alumina content (Arabi and Wild, 1989). These
hydration compounds are similar to those observed in hydrated
PC-based systems. Cation exchange and chemical combination
(silicate bonding or pozzolanic reaction) commence at the same
time, the former being an immediate effect whereas the latter
takes a considerable time to complete (Bell, 1988; Croft, 1964).
As was seen earlier for the case of hydrated PC systems, ettringite
can also form in lime–clay systems in the presence of sulfate. The
sulfate may be naturally occurring in the clay soil, or may emanate
from other sources such as contaminated groundwater or artificial
dosage either intentionally or inadvertently before, during or after
lime-stabilisation. Wild et al. (1996, 1998), Higgins et al. (1998),
Kinuthia et al. (1999) and Oti et al. (2008a, 2008b) have all
researched into lime-stabilisation of clay soils in the presence of
both naturally occurring sulfate (sulfate-bearing clay soils) and
artificially added sulfates (sulfates of potassium, sodium, magne-
sium and calcium). In all cases, ettringite formation was witnessed,
and its formation was suppressed or totally eliminated by the
presence of ggbs. In the presence of sulfate, especially calcium
sulfate, ggbs activation was significantly enhanced, resulting in
improved strength and volume stability.
2.2. Lime–ggbs reactions
As neither lime nor ggbs can hydrate on their own to any
significant degree, lime–ggbs reaction in the ternary clay–lime–
ggbs system essentially comprises the hydration of ggbs, upon
activation by lime. ggbs is a latently hydraulic material and
although it may be cementitious on its own, the process is very
slow at room temperature unless there is some form of
activation. If ggbs is hydrated at elevated temperature and for a
long time, it requires no additives other than water for hydration
and hardening to take place (Song et al., 2000). Higgins (1998)
observed that ggbs on its own has only mild cementitious
properties and, in conventional concrete, it is used in
combination with PC, the alkalinity of which provides the
catalyst to activate the cementitious properties of the ggbs. He
also reported that lime (calcium hydroxide) could provide the
necessary alkali for activation.
160 Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review
Upon activation of ggbs, the lime–ggbs reaction produces C–A–
S–H gel and hydrotalcite type phases containing magnesium
(Meng et al., 1998). This reaction is known to consume lime. In
contrast to the pozzolanic reaction of clay with lime, which is
slow, ggbs hydration upon activation by lime is much quicker
(Wild and Tasong, 1999).
2.3. Clay–ggbs reactions
Wild et al. (1998), in their work on the effect of partial
substitution of lime with ggbs on the strength properties of lime-
stabilised clay soil, reported that if only ggbs is present and no
lime is present in the mixture, there is no strength development
and the strengths are extremely low. In contrast, strength
development of stabilised clay mixes using lime–ggbs blends of
low lime : ggbs ratio is substantial, thus confirming the much
more rapid reaction in the hydration of activated ggbs relative
to ggbs hydration without an activator. In more recent research
at the University of Glamorgan, UK, significant strength gain
has been observed when lower Oxford clay (LOC) was stabilised
using ggbs on its own. In the stabilisation of LOC with high
quantities of ggbs (up to 20%), Oti et al. (2008a, 2008b) observed
significant strength development of the order of 1730 kN/mm2.
Using the same quantity of ggbs, the researchers found no
strength development with an industrial (pure) kaolinite clay.
This is in agreement with the findings of Wild et al. (1998) who
used the same type of kaolinite clay. Oti et al. (2008a, 2000b)
attributed the strength development in LOC–ggbs mixtures to
the presence of gypsum in the Oxford clay. The gypsum is
thought to have contributed to the activation of ggbs, thus
gaining cementation potential to become a formidable binder.
At the higher ggbs content of 20%, ggbs outperformed 20%
lime. The high stabiliser dosages were used in a bid to achieve
high strengths for brick making, as opposed to the typical 3–8%
stabiliser dosages used for soil stabilisation in highway
construction.
In view of the above, the clay–ggbs reaction is therefore
essentially a cementation process, in a similar manner to the
cementation of clay soil particles by PC. The cementation takes
place when the gel formed by the hydration of the activated
ggbs coats and binds the clay particles together.
3. EFFECT OF GGBS ON ENGINEERING PROPERTIES
3.1. Consistency (Atterberg) limits
Wild et al. (1996) studied the effect of lime–ggbs blends on the
consistency limit of kaolinite clay. In the absence of ggbs, they
observed that the addition of 6 wt.% lime resulted in a large
increase in the liquid limit (LL), and a relatively smaller increase
in the plastic limit (PL). This resulted in a small increase in the
plasticity index (PI). Using the same type of kaolinite clay
Kinuthia et al. (1999) studied the effect of a wide range of lime
additions on the consistency limits of the clay. They observed
that the addition of 3 wt.% lime produced a marked increase in
all the Atterberg limits: LL, PL and PI. Upon further addition of
lime, there was a gradual decrease in the LL with a continued
gradual increase in the PL, the overall effect of which was a
gradual decline in the PI. The results are therefore in agreement
with Wild et al. (1996) on the effects of lime on LL and PL, but it
is apparent that the differences in the magnitudes of these
changes determine whether there is an increase or decrease in
the PI. The introduction of ggbs by Wild et al. (1996) in the
Nidzam N Kinuthia
stabilisation of kaolinite clay using various lime–ggbs blends of
decreasing lime : ggbs ratio resulted in significant reductions in
the LL, relative to the lime-stabilised kaolinite clay, and more
marked decreases in the PL, thus producing increases in the PI.
Akinmurusu (1991) studied the effect of ggbs addition on the
consistency, compaction and strength characteristics of a
lateritic soil. The ggbs content varied from 0 to 15% by weight
of the dry soil. He observed a decrease in both the LL and PL,
and an increase in the PI with increasing ggbs addition. These
observations appear to be in agreement with Wild et al. (1996).
Thus, there is research evidence to suggest that the effects of
ggbs on the consistency limits of a clay soil are directly opposite
those of lime, by reducing the LL and PL and increasing the PI.
Various researchers have used a variety of activators on ggbs for
soil stabilisation. The changes in the consistency limits of
different clay soils upon the introduction of ggbs, as observed
by a few of these researchers, are summarised in Table 1.
3.2. Compaction characteristics
Many researchers have studied the effects of lime on the
compaction properties of clay soils. However, not many researchers
have incorporated ggbs in their studies. Wild et al. (1996) also
observed that the addition of lime to kaolinite clay soil
significantly decreased the maximum dry density (MDD) and
increased the optimum moisture content (OMC) of the clay.
However, when they used lime–ggbs blends, a progressive decrease
in lime–ggbs ratio (increase in ggbs) produced a gradual increase in
the MDD and a slight and non-systematic decrease in the OMC.
Akinmurusu (1991) observed that the addition of ggbs to a lateritic
soil increased the MDD up to a ggbs dosage of 10%, beyond which
further addition of ggbs resulted in a reduction in the MDD. The
researcher attributed this behaviour to the increase of fine powder
in the mixture leading to a decrease in the proportion of the coarse
material thus making it difficult to attain good compaction.
However, in combination with the observations by Wild et al.
(1996) the overall effect of the presence of ggbs appears to be an
increase in the MDD of the stabilised material, and a reduction in
the OMC as summarised in Table 2.
3.3. Compressive strength
Relative to lime-stabilisation of soils, the strength gain in clay soils
stabilised with activated ggbs has not been investigated by many
researchers. Akinmurusu (1991) studied the effects of ggbs addition
on the shear strength parameters of a lateritic soil. The California
bearing ratio (CBR) strength index was found to increase with
increasing ggbs content up to a dosage of 10% and then started to
decrease. Wild et al. (1998) studied the influence of ggbs and
gypsum on kaolinite and Kimmeridge clay when stabilised with
lime. They found that partial substitution of lime with ggbs
produced improved strength for both kaolinite and Kimmeridge
clay, particularly at the low lime : ggbs ratios. In the case of
kaolinite, the effects were more significant in the presence of
artificially introduced gypsum. The strength enhancement was
attributed to the contribution of the gypsum to the kaolinite–lime
reaction via ettringite formation, and also to enhanced lime-
activated ggbs hydration. They also found that there was no
enhancement of either 7 and 28 day strength of kaolinite clay when
ggbs alone was added (without lime). Higgins et al. (1998) observed
that the optimum lime : ggbs ratio to achieve maximum UCS was
1 : 5 for kaolinite clay and about 1 : 1?5 for Kimmeridge clay. The
enhancement of strength by the presence of activated ggbs found by
these and a few other researchers and reported in the literature is
summarised in Table 3.
3.4. Swelling potential
Higgins et al. (1998, 2002) studied the effects of ggbs on the
swelling properties of lime-stabilised kaolinite in the presence of
artificially added gypsum (up to 8% by weight of the dry soil).
They found that kaolinite clay stabilised with lime in the
presence of gypsum produced large expansions when saturated
with water. The addition of ggbs to the clay–lime–gypsum
system resulted in a great reduction in expansion.
Higgins (1998) and Higgins and Kennedy (1999) have compre-

Stabiliser

Ratio

Target clay Type (%) LL PL PI

Kaolinite* None — 61 32 29
Lime 6 73 41 32
Lime 10 73 41 32
Lime 20 70 43 29
Kaolinite{ Lime : ggbs 6:4 83 41 42
Lime : ggbs 4:6 81 38 43
Lime : ggbs 2:8 79 35 44
Kaolinite{ WSA 20 83 49 40
WSA : ggbs 14:6 79 40 39
WSA : ggbs 10:10 76 39 37
LOC{ None — 66 35 31
WSA 20 78 50 28
WSA : ggbs 14:6 74 44 30
WSA : ggbs 10:10 70 41 29

WSA – Wastepaper sludge ash (contains 5% free lime for ggbs activation).
*Kinuthia et al. (1999); {Wild et al. (1996); {Nidzam (2004).

Table 1. Effects of ggbs on the Atterberg limits of some stabilised clay soils

Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review Nidzam N Kinuthia 161
 Maximum dry Optimum moisture
Stabiliser density content

Target clay Type Ratio (%) (Mg/m3) (%)

Kaolinite* None — 1?54 24

Lime 6 1?49 25
Lime 10 1?48 26
Lime 20 1?45 28
Kaolinite{ Lime : ggbs 6:4 1?38 29
Lime : ggbs 4:6 1?39 30
Lime : ggbs 2:8 1?40 28
Kaolinite{ WSA 20 1?26 30
WSA : ggbs 14:6 1?32 27
WSA : ggbs 10:10 1?37 26
LOC{ None — 1?36 25
WSA 20 1?20 32
WSA : ggbs 14:6 1?24 30
WSA : ggbs 10:10 1?28 29
English china clay1 Lime 1?5 — —
Lime 5 1?53 26
Lime : ggbs 1?5:10 1?58 23
Mercia mudstone1 Lime 1?5 — —
Lime 5 1?94 13
Lime : ggbs 1?5:10 2?00 12

*Kinuthia et al. (1999); {Wild et al. (1996); {Nidzam (2004); 1Poh et al. (2006).

Table 2. Effects of ggbs on the Proctor compaction properties of some stabilised clay soils

hensively reported on laboratory and subsequent full-scale site
trials using ggbs and lime, on the A421 Tingewick bypass in
Buckinghamshire, UK. They demonstrated that lime–ggbs blends
can successfully be used for soil stabilisation in the presence of
sulfates with a view to enhancing strength development and also
preventing swelling. The soil on this site comprised a sulfate-
bearing boulder clay. ggbs activated using lime was used for the
treatment of particular sections, and lime combined with PC used
in other sections. No swelling problems were observed in the
sections that were treated with ggbs activated using lime whereas
expansion was detected in the sections that were treated with lime
and cement without using ggbs. The effects of the presence of ggbs
on swelling may be summarised by Wild et al. (1998) who
concluded that in order to eliminate problems of expansion
associated with lime-stabilisation of sulfate-bearing clay soils, 60
to 80% of the lime used for stabilisation may be replaced with ggbs.
Table 4 summarises the reductions in linear expansion reported by
some researchers in this area, upon incorporating ggbs in lime- or
PC-stabilisation of clay soils in the presence of sulfates.
4. DISCUSSION
Some of the advantages of the stabilisation of clay soils using
lime include the early property changes taking place in the clay,
commonly referred to as modification. This, as explained earlier,
results from the initial cationic exchange taking place at the clay
particle surfaces, leading to flocculation of the clay particles.
The ‘floccs’ further agglomerate, leading to further material
changes. The open structure reduces the density of the stabilised
mixture, and takes in more water to reach optimal compaction.
The practical implications of this fundamental change in the
clay include the fact that the modified material has a higher
load-bearing capacity in far wetter conditions. This is proven by
the tendency of the material to show an increased LL and a
relatively much higher increase in PL, as discussed earlier and
also comprehensively reported by Kinuthia et al. (1999). The
overall effect is a lowered plasticity for the material in question.
Although this is a good outcome in terms of early enhancement
of load-bearing capacity in wet conditions, the capacity to hold
large amounts of water can become a liability in some cases, for
example in the presence of sulfate ion species. The free-flowing
conditions allow the interaction of calcium, aluminium and
sulfate ion species, allowing for the unhindered formation of the
expansive calcium sulfo-aluminate phases and compounds such
as ettringite. These phases are known to imbibe large amounts of
water during their formative stages, leading to large expansion
magnitudes. Once the voids are filled with the fast-growing
hydration products, the pressure generated may disrupt the
coherence of the matrix, leading to both expansion and loss of
strength. This has proven catastrophic, as witnessed in the
extensive road failures reported by numerous researchers and
practising engineers alike (Higgins et al., 1998, 2002; Hunter,
1988; Mitchell, 1986; Mitchell and Dermatas, 1990; Snedker,
1996; Snedker and Temporal, 1990; Wild et al., 1996, 1998). The
open structure of lime-stabilised material has yet another
disadvantage, in that it would require large amounts of stabiliser
in order to generate a large volume of hydration products to fill
up the voids. Thus, the open structure creates a capping to the
magnitude on strength development. This is not only uneco-
nomical, but research has shown that continued addition of lime
in a bid to gain more strength, for example beyond 8–10%, is
not beneficial as far as strength development, as well as
resistance to wetting, are concerned.
In summary, when ggbs is introduced in a lime–clay stabilised
mixture, the robustness of the system is enhanced by numerous
factors.
(a) Density: the reduction in the flocculation and agglomeration
of the stabilised clay mixture by the addition of ggbs leads to
a denser particle packing of lower permeability.

162 Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review Nidzam N Kinuthia
 28-day strength: Highest linear
Clay Stabiliser Ratio kN/m2 expansion:
Clay Stabiliser Ratio: % mm
Kaolinite* Lime 20 350
Lime : ggbs 14 : 6 375 Kaolinite* Lime 6 4?00
Lime : ggbs 8 : 12 750 Lime : ggbs 6:4 0?24
Lime : ggbs 4 : 16 375 Lime : ggbs 4:6 0?20
Lime : ggbs 0 : 20 150 Lime : ggbs 2:8 0?14
Kaolinite{ Lime : ggbs 6:4 850 Lime 6 0?98
Lime : ggbs 4:6 1400 Lime : ggbs 1?5 : 8 0?32
Lime : ggbs 2:8 1700 LOC{ WSA 20 0?25
Kaolinite* PC 20 1650 WSA : ggbs 14 : 6 0?05
PC : ggbs 14 : 6 900 WSA : ggbs 10 : 10 0?01
PC : ggbs 8 : 12 700 English china Lime
PC : ggbs 4 : 16 350 clay{ 1?5 3?20
PC : ggbs 0 : 20 150 Lime : ggbs 1?5 : 10 3?00
Kaolinite{ WSA 20 1440 Mercia Lime
WSA : ggbs 14 : 6 1370 mudstone{ 1?5 0?35
WSA : ggbs 10 : 10 690 Lime : ggbs 1?5 : 10 0?10
LOC* Lime 20 700
Lime : ggbs 14 : 6 1250 *Wild et al. (1999).
Lime : ggbs 8 : 12 1800 {Nidzam (2004).
Lime : ggbs 4 : 16 2000 {Poh et al. (2006).
Lime : ggbs 0 : 20 1750
LOC1 Lime : ggbs 10 : 10 1630
Lime : ggbs 6 : 14 2130 Table 4. Effects of ggbs on the linear expansion of some
LOC* PC 20 1500 stabilised clay soils after moist curing for 7 days and then
PC : ggbs 14 : 6 1800 soaking in water for 100 days
PC : ggbs 8 : 12 1850
PC : ggbs 4 : 16 1950
PC : ggbs 0 : 20 1825 system further enhances the robustness of the system. Wild
LOC1 PC : ggbs 12 : 8 1363 et al. (1996, 1998) and Wild and Tasong (1999) observed
PC : ggbs 8 : 12 1427 better activation of ggbs when gypsum was artificially
LOC{ WSA 20 664
dosed into the lime-stabilised kaolinite clay mixture. Oti et
WSA : ggbs 14 : 6 907
WSA : ggbs 10 : 10 950 al. (2008a, 2008b) have also observed sulfate-activated ggbs
English china Lime 1?5 1400 as being the cause for strength enhancement in LOC, even in
clay" the absence of lime.
Lime 5 1300
Lime : ggbs 1?5 : 10 1500
There are possibly other mechanisms in operation, the combined
Mercia Lime 1?5 1300
mudstone" action of which gives ggbs a multi-faceted operation, thus
Lime 5 1200 offering a very formidable contribution to the development of
Lime : ggbs 1?5 : 10 8000 the stabilisation and binding of soils, wastes and contaminants.
This is a very good development, especially after the many years
*Oti et al. (2008a).
{Wild et al. (1996). of reliance on the traditional stabilisers of lime and/or PC for
{Nidzam (2004). stabilising clay soils. The manufacture of these traditional
1Ismail et al. (2009). stabilisers, especially of PC, contributes an enormous proportion
"Poh et al. (2006). of the carbon dioxide emissions into the atmosphere. With
experience on the application of blended soil stabilisers using
Table 3. Effects of ggbs on the compressive strength of some some of the more-researched industrial by-product materials
stabilised clay soils such as ggbs, some of the early problems experienced with their
use may be overcome. In the use of lime–ggbs blends on clays of
high plasticity for example, sufficient lime and time should be
(b) Strength: the reduced pore volume facilitates rapid filling up ensured in order to achieve material modification to enable
with hydration products, thus enabling a far better strength sufficient workability to facilitate construction. This was
development relative to the use of lime alone, which would successfully carried out on the A421 Tingewick bypass site trials
result in a very open micro-structure and reduced volume of mentioned earlier, by staged application of lime and ggbs
hydration products. Wild et al. (1998) observed strength (Higgins, 1998; Higgins and Kennedy, 1999). The lime was
enhancement by the presence of ggbs, of the order of two to applied 3 days prior to the addition of ggbs. In this manner, the
three times for kaolinite clay at 7 and 28 days of moist material modification was allowed to take place unhindered.
curing, respectively, and a better enhancement of two to
four times for the two curing periods, respectively, for the 5. CONCLUSIONS
sulfate-bearing Kimmeridge clay. The substitution of conventional stabilisation methods with the
(c) Durability: the reduction in void volume by the addition of innovative use of ggbs possesses numerous benefits as demon-
ggbs limits the free flow of ion species, enhancing resistance strated in this paper. The consistency of the soil was changed, and
to sulfate attack. In addition, the improved activation of the engineering properties of compaction, strength development,
ggbs by the presence of certain ion species that otherwise, if durability and resistance to swelling were all significantly
present, would sometimes be deleterious to the lime–clay improved when additions of ggbs activated by relatively small

Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review Nidzam N Kinuthia 163
 amounts of an activator such as lime were incorporated in the soil
stabilisation process. Within the context of growing environ-
mental awareness, ggbs may be the preferred choice, when and
where it is available. The blended stabiliser is much more
favourable when compared with conventional additives such as
lime and PC. The potential for its utilisation has all the benefits of
economy, technology and caring for the environment, hence
accomplishing robust sustainability criteria that would enthuse
civil engineers, engineering geologists and other practitioners in
the general area of soil engineering and chemistry.
REFERENCES
Akinmurusu JO (1991) Potential beneficial uses of steel slag
wastes for civil engineering purposes. Resources
Conservation and Recycling 5(1): 73–80.
Arabi M and Wild S (1989) Property changes induced in clay soils
when using lime stabilization. Proceedings of the Institution of
Civil Engineers, Municipal Engineer 6(2): 85–99.
Bell FG (1988) Stabilisation and treatment of clay soils with
lime: Part 1 – Basic principles. Proceedings of the
Institution of Civil Engineers, Ground Engineering (2): 127–
153.
Bell FG and Tyrer MJ (1987) Lime-stabilisation and clay
mineralogy. Proceedings of Foundations and Tunnels,
London, vol. 11, pp. 1–7.
Bijen JG (1996) Blast Furnace Slag Cement. Association of the
Netherlands Cement Industry (VNC) the Netherlands.
Brook-Bradley HE (1952) Soil–cement roads in Worcestershire.
Surveyor 3: 571–573.
Caijun S and Day RL (1993) Chemical activation of blended
cements made with lime and natural pozzolans. Cement and
Concrete Research 23(6): 1389–1396.
Cobbe MI (1988) Lime-modification of kaolinite–illite clays.
Civil Engineering Technology 10(7): 9–15.
Croft JB (1964) The pozzolanic reactivities of some New South
Wales fly-ashes and their application to soil stabilization.
Proceedings of the ARBB Australia, Australian Road
Research Board (ARRB), Australia, vol. 2, part 2, paper 120,
pp. 1114–1167.
Daimon M (1980) Mechanism and kinetics of slag cement
hydration. Proceedings of the 7th International Congress of
the Chemistry of Cement, Paris, Sub-Theme III-2, vol. 1, III-
2/1–III-2/9.
Diamond S and Kinter EB (1966) Absorption of calcium
hydroxide by montmorillonite and kaolinite. Journal of
Colloid and Interface Science 22(3): 240–249.
Douglas E, Bilodeau A and Brandster J (1991) Alkali activated
ground blast-furnace slag concrete: preliminary investiga-
tion. Cement and Concrete Research 21(1): 101–108.
Gjorv OE (1989) Alkali activation of Norwegian granulated blast
furnace slag. Proceedings of the 3rd International
Conference, Trondheim, Norway, vol. 2, SP114-73, American
Concrete Institute, Detroit, MI, pp. 1501–1517.
Gupta S and Seehra SS (1989) Studies on lime-granulated blast
furnace slag as an alternative binder to cement. Highway
Research Board Bulletin 38: 81–97.
Hakkinen T (1993) The influence of slag content on the
microstructure, permeability and mechanical properties of
concrete. Part 1: Microstructural studies and basic mechan-
ical properties. Cement and Concrete Research 23(2): 407–
421.
164 Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review
Higgins DD (1998) What’s new in ggbs? Concrete Magazine,
May.
Higgins DD and Kennedy J (1999) Lime and ground granulated
blastfurnace slag stabilization of boulder clay on the A421
Tingewick bypass. Proceedings of the 3rd European
Symposium on the Performance and Durability of
Bituminous Material and Hydraulic Stabilized Composites,
Leeds, UK, 8–9 April 1999.
Higgins DD, Kinuthia JM and Wild S (1998) Soil stabilization
using lime-activated ggbs. Proceedings of the 6th CANMET/
ACI International Conference on Fly Ash, Silica Fume, Slag
and Natural Pozzolans in Concrete), Bangkok, 31 May–5
June, (Malhotra VM (ed.)), vol. 2, pp. 1057–1074.
Higgins DD, Thomas B and Kinuthia J (2002) Pyrite oxidation,
expansion of stabilised clay and the effect of GGBS.
Proceedings of the 4th European Symposium, Bitmat4, on
Performance of Bituminous and Hydraulic Materials in
Pavements, Nottingham (Zoorob SE, Collop, AC and Brown
SF (eds)) A.A. Balkema, the Netherlands, p. 348.
Hunter D (1988) Lime-induced heave in sulphate-bearing clay
soils. ASCE Journal of Geotechnical Engineering 114(2):
150–167.
Ismail N, Kinuthia JM and Robinson RB (2009) The development
of sustainable lime-clay building components. Proceedings
of 4th Research Student Workshop, University of Glamorgan.
Glamorgan, UK, ISBN 978-1-84054-224-0.
Kinuthia JM (1997) Property Changes and Mechanisms in Lime-
Stabilised Kaolinite in the Presence of Metal Sulphates. PhD
thesis, University of Glamorgan, UK.
Kinuthia JM, Wild S and Jones GI (1999) Effects of monovalent and
divalent metal sulphates on consistency and compaction of
lime-stabilised kaolinite. Applied Clay Science 14(1-3): 27–45.
Lee AR (1974) Blastfurnace and Steel Slag. Edward Arnold
Publishers Ltd, London, UK.
Macphee DE, Atkins M and Glasser FP (1989) Phase develop-
ment and pore solution chemistry in ageing blastfurnace
slag-Portland cement blends. Materials Research Society
Symposium Proceedings 127: 475–480.
Meng C, Weins U and Schiessl P. (1998) Fly ash, silica fume, slag
and natural pozzolans in concrete. Proceedings of the 6th
CANMENT/ACI International Conference on Fly Ash, Silica
Fume, Slag, and Natural Pozzolans in Concrete, vol. 1,
American Concrete Institute, Bangkok, pp. 109–128.
Mitchell JK (1986) Practical problems from surprising soil
behaviour. The twentieth Karl Terzaghi Lecture. ASCE
Journal of Geotechnical Engineering 112(3):274–279.
Mitchell JK and Dermatas D (1990) Clay soil heave caused by
lime-sulphate reactions. ASTM Symposium on Innovations
and Users of Lime, San Francisco, 1990.
Nidzam RM (2004) Soil Stabilisation utilising Waste paper
sludge Ash (WSA). PhD thesis, University of Glamorgan, UK.
Nixon PJ and Gaze ME (1981) The use of fly ash and granulated
blastfurnace slag to reduce expansion due to the alkali-aggregate
reaction. Proceedings of the 5th International Conference on
Alkali-Aggregate Reaction in Concrete, Cape Town, Pretoria,
National Building Research Institute, paper S 252/32.
Oti JE, Kinuthia JM and Bai J (2008a) Using slag for unfired clay
masonry-bricks. Proceedings of the Institution of Civil
Engineers, Construction Materials 161(4): 147–155.
Oti JE, Kinuthia JM and Bai J (2008b) Developing unfired
stabilized building materials in the UK. Proceedings of the
Nidzam N Kinuthia
 Institution of Civil Engineers, Engineering Sustainability
161(4): 211–218.
Ouf ME (2001) Stabilisation of Clay Sub-Grade Soils using
Ground Granulated Blastfurnace Slag. PhD thesis, Univeristy
of Leeds, UK.
Poh D, Ghataora G and Gazireh N (2006) Soil stabilisation using
basic oxygen steel (BOS) slag fines. Journal of Materials in
Civil Engineering 18(2): 229–240
Regourd M (1980) Structure and behaviour of slag Portland
cement hydrates. Proceedings of the 7th International
Conference on the Chemistry of Cement, Paris, France, Sub-
Theme III-2. vol. I, pp. III-2/10–III-2/26.
Richardson IG, Brough AR, Groves GW and Dobson CH (1994)
The characterization of hardened alkali-activated blas-
tfurnace slag pastes and nature of the calcium hydrate (C–
S–H) phase. Cement and Concrete Research 24(5): 813–
829.
Rogers CDF, Glendinning S, Roff TEJ (1997) Lime modification
of clay soils for construction expediency. Proceedings of the
Institution of Civil Engineers, Geotechnical Engineering
125(4): 242–249.
Sherwood PT (1993) Soil Stabilisation with Cement and Lime.
Transport Research Laboratory/HMSO, London.
Sloane RL (1965) Early reaction in the kaolinite-hydrated lime-
water system. Proceedings of the 6th International
Conference on Soil Mechanics and Foundation Engineering,
Montreal, vol. 1, pp. 121–125.
Smolczyk HG (1980) Slag structure and identification of slags.
Proceedings of the 7th International Conference on the
Chemistry of Cement, Paris, France, pp. III-1/3–III-1/17.
Snedker EA (1996) M40 – Lime stabilisation experiences.
Proceedings of Seminar held at Loughborough University,
Civil Engineering Department, 25 September. Thomas
Telford, London, pp. 142–158.
Snedker EA and Temporal J (1990) M40 Motorway Banbury IV
Contract – lime stabilisation. Highways and Transportation,
December, 7–8.
Song S, Sohn D, Jennings HM and Mason TO (2000) Hydration
of alkali-activated ground granulated blastfurnace slag.
Journal of Materials Science 35(1): 249–257.
Talling B (1989) Effect of curing condition on alkali-activated
slags. Trondheim Conferences, Elsevier, Trondheim, Norway,
SP 114-72, pp. 1485–1499.
Tasong WA, Wild S and Tilley RJD (1999) Mechanism by which
ground granulated blastfurnace slag prevents sulphate
attack of lime-stabilised kaolinite. Cement and Concrete
Research 29(7): 975–982.
Taylor HFW (1990) Cement Chemistry. Academic Press, London,
pp. 38–39.
Wild S and Tasong WA (1999) Influence of ground granulated
blastfurnace slag on the sulphate resistance of lime-stabilised
kaolinite. Magazine of Concrete Research 51(4): 247–254.
Wild S, Kinuthia JM, Robinson RB and Humphreys I (1996)
Effects of ground granulated blastfurnace slag (ggbs) on
strength and swelling properties of lime stabilised
kaolinite in the presence of sulphates. Clay Minerals 31(3):
423–433.
Wild S, Kinuthia JM, Jones GI and Higgins DD (1998) Effects of
partial substitution of lime with ground granulated blas-
tfurnace slag (ggbs) on the strength properties of lime–
stabilized sulphate bearing clay soils. Engineering Geology
51(1): 37–53.
Wild S, Kinuthia JM, Jones GI and Higgins DD (1999)
Suppression of swelling associated with ettringite formation
in lime stabilized sulphate bearing clay soils by partial
substitution of lime with ground granulated blastfumace
slag. Engineering Geology 51(4): 257–277.
Wu X, Jiang W and Roy DM (1990) Early activation and
properties of slag cement. Cement and Concrete Research
20(6): 961–974.

What do you think?

To discuss this paper, please email up to 500 words to the editor at journals@ice.org.uk. Your contribution will be forwarded to the
author(s) for a reply and, if considered appropriate by the editorial panel, will be published as discussion in a future issue of the
journal.
Proceedings journals rely entirely on contributions sent in by civil engineering professionals, academics and students. Papers should be
2000–5000 words long (briefing papers should be 1000–2000 words long), with adequate illustrations and references. You can submit
your paper online via www.icevirtuallibrary.com/content/journals, where you will also find detailed author guidelines.

Construction Materials 163 Issue CM3 Sustainable soil stabilisation with blastfurnace slag – a review Nidzam N Kinuthia 165