Soils and Foundations 2016;56(6):1021–1034

HOSTED BY The Japanese Geotechnical Society

Soils and Foundations

www.sciencedirect.com
journal homepage: www.elsevier.com/locate/sandf

A new low carbon cementitious binder for stabilising weak ground

conditions through deep soil mixing
P. Sargenta,n, P.N. Hughesb, M. Rouainiac
a
AECOM, First Floor, One Trinity Gardens, Broad Chare, Quayside, Newcastle upon Tyne NE1 2HF, UK
b
Christopherson Building, School of Engineering and Computing Sciences, Durham University, South Road, Durham, DH1 3LE, UK
c
Drummond Building, School of Civil Engineering and Geosciences, Newcastle University, Newcastle upon Tyne, Tyne and Wear NE1 7RU, UK
Received 21 September 2015; received in revised form 27 July 2016; accepted 13 August 2016
Available online 2 December 2016

Abstract

Soft alluvial soils present unfavourable conditions for engineering developments due to their poor bearing capacities and high potential for
experiencing shrinkage and swelling. This paper focusses on deep dry soil mixing (DDSM), which introduces cementitious binders to soft soils via a
rotating auger drill, thereby producing soil-cement columns. Ordinary Portland cement (CEM-I) is globally used across the construction industry and
is the most commonly used binder for DDSM applications due to its high strength performance. However, CEM-I production is one of the world's
most energy intensive and expensive industrial processes, contributing 5–7% of the world's total CO2. There is now significant pressure on the
cement and construction industries to greatly reduce their CO2 emissions by developing “greener” alternatives to CEM-I, which are both more
environmentally and financially sustainable in the long-term. Alkali activated industrial by-products (IBP's) such as ground granulated blast furnace
slag (GGBS), known as geopolymers have been identified as potential alternatives. These are advantageous due to negating the need to transfer IBP's
to landfill, their abundance, negligible or zero production costs. Geopolymers are capable of reducing greenhouse gas emissions by up to 64%.
Calcium-bearing slags have also been found to possess potential for carbon capture and storage (CCS). Comparisons with the strength and durability
of untreated and stabilised soils have been made in this study. Results indicate that stabilising an alluvial soil with sodium hydroxide (NaOH)
activated GGBS produced significant strength and durability improvements surpassing CEM-I. The addition of NaOH allowed pozzolanic reactions
to occur, leading to improved mechanical properties with time, with a particularly marked improvement in strength.
& 2017 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: Deep dry soil mixing; Sustainability; Low carbon; Alkali activation; Performance

1. Introduction swelling, settlement and durability properties. Various ground

improvement techniques can be adopted to enhance the
Soft soils, including alluvium, present problematic ground engineering performance of such soils, whereby the most
condition, as a result of their poor bearing capacity, shrinkage/ appropriate technique depends on the physico-chemical prop-
erties of the soil in question. Deep dry soil mixing (DDSM) is
becoming an increasingly popular and effective ground
n
Corresponding author. improvement technique in the UK for treating such soils by
E-mail addresses: [email protected] (P. Sargent),
[email protected] (P.N. Hughes),
creating cemented soil columns via auger mixing. Since the
[email protected] (M. Rouainia). development of the DDSM technique, lime and Ordinary
Peer review under responsibility of The Japanese Geotechnical Society. Portland Cement (CEM-I) have been used as the binders

http://dx.doi.org/10.1016/j.sandf.2016.11.007
0038-0806/& 2017 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1022 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

because of their ability to produce impressive strength, its higher levels of replacement and ability to produce higher
particularly in CEM-I. The presence of soil water and calcium strengths compared with PFA. However, the aim of this
silicates/aluminates within the cementitious binders injected research area is to design new low carbon cementitious binder
into the soil during mixing react to form hydration products, mixtures which negate the need for the use of any lime or
including calcium silica hydroxide (C-S-H) and calcium CEM-I.
aluminate hydroxide (C-A-H) gels. For DDSM, the physico- Laboratory and field-based research by workers including
chemical properties of a soil to be considered in selecting the Hughes and Glendinning (2004) and Sargent et al. (2013) has
most appropriate cementitious binder include the particle size demonstrated that GGBS has significant potential as a sustain-
distribution, plasticity, pH, moisture content, cation exchange able replacement for lime and CEM-I, Hughes and Glendin-
capacity (CEC), specific surface area, organic and sulphate ning (2004) implemented GGBS for stabilising peaty soils
contents. For lime and cement stabilisation to work effectively, through DDSM along the Channel Tunnel Rail Link. How-
a low organic content (o 1%), a low sulphate content ever, there are instances when soil pH is too acidic and the
(o 0.3%) and a clay contents of 10–50% are required moisture content is too high for strength gains to develop when
(Tutumluer, 2012). just using GGBS. The hydraulicity of the GGBS is latent; i.e. it
Unfavourable environmental and financial issues are asso- is confined within its glassy structure (Newman and Choo,
ciated with utilising CEM-I as a binder. Cement production 2003). Hence, the addition alkali-activators aims to release the
contributes 5–7% of global CO2 emissions (McLellan et al., GGBS reactivity by raising soil pH and ultimately increasing
2011). Increases in such emissions are anticipated to have long rates at which the mechanical properties of stabilised soils are
term adverse warming effects on the global climate, the improved (Palomo et al., 1999). Once the pH of the stabilised
consequences of which include changing weather patterns, soil reaches 10.5 (Davidson et al., 1965), the GGBS reactivity
melting polar ice caps, rising sea levels and ocean acidification will be triggered and the reaction will start and continue over
(EPA, 2015). These effects will potentially have major impacts long periods of time with the soil water via pozzolanic
on marine and land based life. Therefore, it has become reactions to produce cementitious gels. Such materials are
extremely important for the construction industry to become known as “Geopolymers”.
more sustainable by using materials with lower values of The continued use of lime as an alkali activator is
embodied CO2. The continued use of CEM-I is also financially environmentally unfavourable due to their high carbon costs.
unsustainable, since its manufacture is energy intensive and According to Shi et al. (2006), solid NaOH flakes/pellets and
expensive. The electrical energy consumed per ton of CEM-I Na2SiO3 solution are becoming two of the most widely
produced is 75 kWh (Madlool et al., 2011). The UK's price available and popular activators for use in concretes and
indices of electricity and coal as fuels rose by 75% and 63% cements. Cristelo et al. (2011) used NaOH and Na2SiO3 to
respectively between 2005 and 2011. Similar trends were activate IBP's (class F PFA) for use in DDSM treatment of a
observed during the same period in the United States, with low plasticity sandy clay. Field testing results demonstrated
electricity and coal prices having increased by 19% and 47%, that the activated geopolymer produced higher strengths than
respectively (Imbabi et al., 2012). Since the Fukushima nuclear CEM-I.
disaster in 2011, global coal prices have reduced by half to Habert et al. (2011) reported that lower financial and
approximately $70 per ton (Reuters, 2015), and oil prices have environmental costs are associated with the production of
fallen from $120 to approximately $45 per barrel (LSE, 2015). Na2SiO3 rather than lime and CEM-I. However, NaOH flake/
However, according to DECC (2015) and EIA (2015), pellet production has a lower environmental impact (3.71

electricity prices in the UK and US have remained relatively 10-1 kg CO2 eq, human toxicity level of 15.84 kg 1,4-DB eq
constant since 2011. and a freshwater ecotoxicity of 3.98 kg 1,4-DB eq) than
Hence, there is a need to identify more environmentally and Na2SiO3 solution, which has a high global warming potential
financially sustainable replacement binders. These binders of 117.8 kg CO2 eq, a high human toxicity level of 82.75 kg
should satisfy the following criteria: 1) they should provide 1,4-DB eq and a high freshwater ecotoxicity of 21.84 kg 1,4-
engineering performances comparable to or surpass those of DB eq.
CEM-I within similar curing times, 2) they should be Further justification behind using NaOH over Na2SiO3 as an
commercially available at comparable/lower prices than activator is that the latter has a higher accelerated carbonation
CEM-I, and 3) they should be produced in such a way that depth when used within geopolymer pastes. This has been
fuel costs are markedly lower than those affiliated with CEM-I attributed to the composition and structure of the C-S-H gels
production whilst incurring lower CO2 emissions. A modern formed. The NaOH activated slags have been reported to have
popular route for selecting new binders has been to recycle a higher Ca/Si ratio (1.2) than Na2SiO3 activated slags (Bernal,
alumino-silicate based pozzolanic industrial by-products 2014). However, since the Ca/Si ratio and reduced silicate
(IBP's) such as ground granulated blastfurnace slag (GGBS) chain length of NaOH activated slags is so much higher, these
and pulverised fly ash (PFA). Such materials are already being properties may consequently favour the precipitation of
used as additives to cement mixtures: CEM-II is used in the increasing quantities of carbonates to fill pore spaces. The
production of PFA, and CEM-II/III is being used in GGBS. MgO content of slags has been identified to have a significant
According to Hanson (2014), GGBS is preferred over PFA in role in the mechanism and extent of carbonation, and slags
the UK as a replacement for CEM-I in cement mixtures due to with higher MgO content are more likely to be less affected by
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1023

carbonation. Considering this and the typical MgO content of

various slag types as reported by Sanna et al. (2012),
carbonation is expected to extends less within GGBS-based
geopolymers than in steel slags, argon-oxygen decarburisation
(AOD) process slag and most ash wastes (e.g. PFA).
Under accelerated carbonation testing (i.e. 4 1%), if the
carbonation depth in alkali-activated slag pastes is similar to
that in CEM-I pastes, Bernal (2014) suggested that the alkali-
activated slag would be much more durable under natural
(atmospheric) carbonation conditions. C-S-H gels within
alkali-activated slag mixtures are known to be prone to turning
into amorphous silicate gels post-carbonation, which conse-
quently reduces the strength of the material (Song et al., 2014).
Whilst using higher dosages of alkali-activated slag binders
significantly reduces the carbonation depth (Bernal, 2014), Fig. 1. PSD curves for Lanton alluvium soil (red) and GGBS (blue) with
simply adjusting the slag/alkali activator ratio of the binder in Tsuchida (1970) liquefaction criteria superimposed (red area “1” denotes zone
for most liquefiable soils, green area “2” denotes zone for potentially liquefiable
favour of the activator causes an increased rate of reaction, soils). (For interpretation of the references to color in this figure legend, the
produces more C-S-H and therefore increases strength and reader is referred to the web version of this article.)
carbonation resistance (Song et al., 2014). Experiments under-
taken by Song et al. (2014) demonstrated that by changing the
GGBS/alkali activator ratio from 94/6 to 86/14, the 28 day Table 1
compressive strengths (no carbonation) increased from 14.9 to Summary of the GGBS properties. Courtesy of Hanson (2016).
42.5 MPa, the 21 day carbonated compressive strengths
Property Value
increased from 7.6 to 28.9 MPa and the carbonation depth
reduced from 25 to 10 mm, which is typical for pastes using Granulated grain size (mm) 0–5
CEM-I. Colour Light grey
This paper investigates whether NaOH can successfully Odour N/A
pH (DEV-S4-eluate according EN 12457-4 10 - 12
activate GGBS to ultimately produce engineering perfor- Density at 20 1C (Mg/m3) 2.4–3.0
mances similar to CEM-I and lime when incorporated within Water solubility (g/l) o1
a soft alluvial soil whilst minimising carbonation potential. Melting/freezing point (1C) 41000
Ultimately, the issue of whether the binder is suitable for use in Flash point N/A (inorganic)
DDSM treatment is discussed. Composition
EG-Nr 266-002-0
CAS-Nr 65996-69-2
2. Soil and binder materials CaO (%) 40
SiO2 (%) 35
2.1. Lanton alluvium Al2O3 (%) 12
MgO (%) 10
Fe2O3 (%) 0.2
For this paper, an alluvial soil was sourced from the flood
plain of the River Glen in Lanton, 4 km north west of Wooler
in Northumberland, UK. Disturbed and undisturbed U100
samples from depths of 1.5–2.4 m were obtained.
3. Methodologies

2.2. Cementitious binder 3.1. Soil index properties

The IBP binder used for this research was ground granulated A series of geotechnical index tests were conducted accord-
blast-furnace slag (GGBS) supplied by Hanson Cements Ltd. ing to BS 1377 (BSI, 1990). These included PSD through wet
A particle size distribution (PSD) curve for the GGBS is sieving and sedimentation (Fig. 1), Atterberg limits
provided within Fig. 1. For comparison purposes, non- (LL ¼ 35.7%, PL ¼ 20.7%, PI¼ 15.0), particle density
activated GGBS (Table 1), CEM-I (supplied by Lafarge) and
(2.61 Mg/m3), the natural in-situ moisture content (25%) and
lime (supplied by Fisher Scientific UK Ltd.) were also used.
compaction testing to determine the optimum moisture content
To ensure that the binder to be utilised for stabilising Lanton
alluvium in this paper had low embodied carbon in terms of of the soil (14.7%), and the bulk (2.0 Mg/m3) and dry
energy and production costs, lower financial cost and reduced (1.74 Mg/m3) densities. According to BS 5930 (2015), Lanton
potential carbonation depth, the Na2SiO3 solution activator alluvium may be categorised as a silty SAND in terms of dry
was not used. Instead, solid NaOH pellets supplied by Fisher density. A summary of these geotechnical index properties is
Scientific UK Ltd. were used as the activator. provided in Table 2.
1024 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

Table 2 BS 7755 (1995). The CEC-related charge density (sCEC) for

Summary of the soil's index properties. the soil was obtained as follows:
Property Value e ðCECx10  2 Þ
sCEC ¼
In-situ moisture content (%) 25.0
2 ab
Plasticity Index 14.95 where e is the elementary charge (1.6022
10–2 C), a and b are
Liquid limit (%) 35.66 unit cell parameters for clay minerals in the x–y plane (Meunier,
Saturated unit weight (kN/m3) 18.44
2005). The soil had an average CEC of 11.45 cmol/kg and sCEC of
Bulk density (Mg/m3) 1.99
Dry density (Mg/m3) 1.74 0.0019.
Cation exchange capacity (cmol/kg) 11.45 The Brunauer Emmett Teller (BET) nitrogen absorption
Specific surface area (m2/g) 6.45 method (Brunauer et al., 1938) was adopted to determine the
Total organic content (%) 0.76 surface area of the Lanton alluvium using a Micromeritics
Sulphate content (mg/kg soil) 49.0
Tristar 300. An average surface area of 6.45 m2/g was recorded
BS 5930 classification[15] Silty sand
for the soil. Total organic carbon (TOC) and total organic
matter (TOM) testing revealed low values of 0.42% and
0.76%, respectively. Per Nair and Little (2009), sulphates
Modern day loadings imposed on geotechnical structures within soils can have detrimental effects on their engineering
and foundation soils are generally complex, comprising a performances when stabilised with cementitious binders. This
mixture of monotonic, cyclic and dynamic loadings occur- arises from the likely formation of ettringite and/or thaumasite,
ring at different magnitudes over various timescales. With which swell upon contact with water. The sulphate content of
reference to the cyclic and dynamic loads, these may be the soil was determined through the use of a Dionex Ion-
imposed by human-related activities such as high-speed Chromatography ICS-1000, which revealed a low sulphate
trains and highway loads and/or natural phenomena such as content of 49 mg per kg. Finally, 100 mm diameter remoulded
earthquakes. Depending on the soil type, liquefaction can and undisturbed samples of Lanton alluvium were placed
result from the moisture content and loading conditions, within a triaxial cell under isotropically consolidated condi-
including both human-related activities and earthquakes, tions to determine their coefficients of vertical permeability
and can severely compromise the integrity of geotechnical (kv), per BS 1377 (BSI, 1990). At 22 1C, the kv values obtained
structures, along with the surrounding infrastructure. Thus, for the soil in its undisturbed and remoulded states were 1.41
the potential of the Lanton alluvium to undergo liquefaction
10-8 m/s and 5.58
10-9 m/s, respectively.
upon cyclic/dynamic loading was assessed according to
Japanese criteria defined by Tsuchida (1970) for port and 3.2. Engineering performance testing
airport design. This criteria was selected on the basis that
Japan is a seismically active region with a well established 3.2.1. Preparation of testing samples
understanding of the liquefaction behaviour of soils and the Numerous binder combinations were added to the soil at
remediation measures available. Additionally, the UK does dosages of 2.5–10% by dry weight to identify the most
not currently have any such comparable British Standards, effective binder and corresponding dosage. Once soil samples
and the ASTM only has a standard for the liquefaction were obtained from the field, they were oven dried at 110 1C
potential of a soil through the normalisation of penetration for 24 h and crushed into a fine powder (particle size
resistance, specifically for sands. r 1 mm). This provided a starting point for achieving a
According to Tsuchida's (1970) soil grading criteria, the specific moisture content of 25%, ensured the repeatability of
PSD curve for Lanton alluvium (Fig. 1) lies within the zone for the experiments and also increased practicality when mixing
soils which may potentially liquefy. This raises concerns about the dry powdered binders in the soil in the laboratory.
the soil's strength, and emphasises the need for treatment to Stabilised samples were mixed via a rotary mixer for 10 min
ensure the suitability of ground conditions for future engineer- to maximise homogeneity with regard to the number of
ing developments, particularly those involving variable cyclic reactive sites available for cementitious bond formation. Water
and dynamic loadings. was then added to the mixtures to achieve the required pre-
The soil was also subjected to physico-chemical tests to treatment (in-situ) soil moisture content.
determine its capacities to react with binders and produce Samples were created by tamping and compressing the
cementitious gels. The pH of the soil in its untreated and samples into split-sample moulds, which were then placed into
stabilised states after various curing periods was measured per a hydraulic press to compact the material to the required
BS 1377 part 3 (BSI, 1990). CEC and specific surface area dimensions and the compacted conditions induced upon
testing were also conducted. The former provides an estimate DDSM. For compressive strength tests, samples 38 mm in
of the number of sites on clay minerals where cation exchange diameter and 76 mm long were formed. For oedometer testing,
may occur. Soils with high surface area values (i.e. higher clay separate samples with dimensions of 76 mm in diameter and
contents) result in higher CEC values. These properties 18 mm in thickness were prepared. Finally, samples 105 mm
provide insights into the potential for the soil to undergo in diameter and 115 mm in length were produced and cured for
cementitious reactions. The method used for CEC analysis was durability testing. All stabilised samples for each type of test
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1025

Table 3
Summary of the binder compositions used for the laboratory stabilisation of
Lanton alluvium.

Soil–binder mixture Composition

Lanton Lanton (25% grav. water)

Lanton þ 5% Lime Lanton (25% grav. water) þ 5% lime
Lanton þ 10% Lime Lanton (25% grav. water) þ 10% lime
Lanton þ 5% CEM-1 Lanton (25% grav. water) þ 5% CEM-I
Lanton þ 10% CEM-1 Lanton (25% grav. water) þ 10% CEM-I
Lanton þ 5% GGBS Lanton (25% grav. water) þ 5% GGBS
Lanton þ 10% GGBS Lanton (25% grav. water) þ 10% GGBS
Lanton þ 2.5% GGBS- Lanton (25% grav. water) þ 1.67% GGBS þ
NaOH 0.83% NaOH
Lanton þ 5% GGBS- Lanton (25% grav. water) þ 3.33% GGBSþ
NaOH 1.67% NaOH Fig. 2. pH for all binder mixtures and dosages used to stabilise Lanton
Lanton þ 7.5% GGBS- Lanton (25% grav. water) þ 5% GGBS þ 2.5% alluvium over 56 days.
NaOH NaOH
Lanton þ 10% GGBS- Lanton (25% grav. water) þ 6.67% GGBS þ
NaOH 3.33% NaOH
required to raise soil pH to Z 10.5 to promote pozzolanic
conditions. pH values recorded for Lanton alluvium and non-
were prepared with a bulk density of 1.9 Mg/m3, based on activated GGBS samples were all o 10, highlighting the need
optimum compaction criteria and a moisture content of 25%. for activation (Fig. 2). For GGBS-NaOH specimens, all four
Table 3 summarises the soil–binder mixtures examined and dosages successfully raised the soil pH to at least 11.7. The
their respective concentrations. maximum pH values recorded ranged between 12.6 for the
For all compressive strength, durability, pH and moisture 7.5% dosage and 12.9 for 10%. These values were comparable
content testing, samples were cured within wax-sealed PVC with those recorded for samples stabilised with 10% lime or
sample moulds for periods of 0, 7, 14, 21, 28, 35, 42, 49 and 56 10% CEM-I, confirming that pH conditions were ideal for
days and stored within a temperature controlled room (55% pozzolanic reactions.
relative humidity, 20 1C ambient air temperature). In the case of Whilst stabilisation using lime and CEM-I at 5% raised and
compressibility testing, however, samples were prepared and maintained a pH 410.5 for 56 days, poor strength and
cured within oedometer rings and tightly sealed using three durability performances were recorded. A similar observation
layers of cling film. For compressibility and durability testing, was made for 2.5% GGBS-NaOH samples, which may be
samples were cured for 28 days before testing based on the attributed to the following: 1) insufficient binder dosage; 2)
construction specifications established by Hansson et al. (2001). insufficient curing time for cementitious gels to form, 3) the
Once cured, all samples were extruded, trimmed and tested. high water content and therefore low calcium concentration of
the soil. Should oxidising reactions and subsequent pH
3.2.2. Laboratory tests reductions occur within stabilised samples, hydration and
A suite of laboratory tests was conducted according to BS pozzolanic reactions may become inhibited and result in poor
1377 (BSI, 1990) to assess the performance of the various or severely delayed strength developments. Thus, it is essential
binders. Strength and stiffness were assessed via unconfined to assess the quantities of activator required within samples to
compressive strength (UCS). Samples were tested either to failure ensure that pH values 4 10.5 are maintained for the long-term.
or a maximum axial strain of 15%. The compressibility of
samples was assessed by oedometer testing, involving a series of
load-unload-reload loops up to vertical stresses of 1600 kPa. Per 4.2. Durability
Davidson et al. (1965), pozzolanic reactions require a minimum
soil pH of 10.5. Thus, the pH of all the stabilised mixtures was 4.2.1. Wetting–drying
tested prior to and after each curing period. The durability of the All sample mixtures were subjected to wetting–drying
stabilised mixtures was assessed by wetting–drying and freeze– durability testing to determine soil-cement loss and changes
thaw tests according to ASTM D559/559M and D560/560M in moisture content and sample volume. The results are
(ASTM, 2015a,2015b), especially given that frequent wetting– displayed in Fig. 3. For the purposes of this study, samples
drying can be expected in the UK's temperate maritime climate. are considered to have failed in wetting–drying and freezing–
thawing testing upon splitting and/or complete disaggregation.
4. Results and discussion In addition, samples were discontinued if it was determined
that volume measurements had become inaccurate due to soil-
4.1. pH cement loss (ASTM D559-559M, 2015). For the traditional
binders, a minimum dosage of 10% was required to maximise
While the strength development of stabilised soils can be the chances of samples surviving all 12 testing cycles. Samples
greatly influenced by their pH, alkali activation may be stabilised with CEM-I survived the 12 cycle testing period
1026 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

Fig. 3. Volumetric (a, b) and moisture content (c, d) changes experienced by all stabilised Lanton alluvium mixtures during wetting–drying testing.

Fig. 4. Volumetric (a, b) and moisture content (c, d) changes experienced by all stabilised Lanton alluvium mixtures during freezing–thawing testing.

with a gradual reduction in volume, whereas 10% lime activated GGBS samples. However, impressive performances
samples only survived 3 cycles exhibiting larger volumetric were recorded for specimens containing a dosage of 7.5% or
losses. According to Sargent et al. (2013), most non-activated 10%, as they displayed negligible changes in volume or
geopolymers fail after the first cycle. However, the 5% and moisture content over the 12 cycles. Whilst the Lanton
10% GGBS samples survived 2 and 3 cycles, respectively. þ 7.5% and 10% GGBS-NaOH samples demonstrate higher
GGBS-NaOH dosages of 2.5% and 5% displayed similar levels of durability against wetting–drying compared with
disappointing performances to those observed for non- Lanton þ 10% GGBS samples, the decreasing water content
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1027

of such non-activated samples after developing latent hydrau- wetting–drying and freezing–thawing, surpassing that of 10%
licity inhibits the hydration of cementitious gels and therefore CEM-I. NaOH activation proved extremely effective in produ-
long-term strength development. cing dense high strength samples, which possessed low
permeabilities and were deemed less likely to experience water
absorption and subsequent sample deterioration through
4.2.2. Freezing–thawing shrinkage, swelling and physical weathering.
Freezing–thawing testing was conducted to assess the ability
of the binder mixtures to resist repeated freezing–thawing 4.3. Compressive strength
cycles by measuring changes in sample volume, moisture
content and soil-cement losses. In the UK, freeze–thaw occurs A summary of the average maximum compressive strengths
up to depths of 2 m, subject to variations between locations achieved by the various stabilised mixtures is presented in
due to in-situ soil and vegetation conditions (Clarke and Fig. 5. With a few exceptions, sample strengths increased over
Smethurst, 2010). Given the grading of Lanton alluvium and the 28 day curing period for all mixtures. However, the degree
that water may easily percolate through the soil, shrinkage, of strength development observed within samples varies
swelling and physical weathering are likely to occur within the significantly between binder dosages. The use of 2.5 or 5%
shallow sub-surface. This raises concerns as to whether the for all binder types proved unsuccessful in producing the 28
binder within DDSM columns is capable of resisting such day UCS values of 300kPa required by EuroSoilStab (2002).
processes. While the ASTM D560/560M procedure may be The strengths recorded for 5% and 10% GGBS were also
considered extreme for simulating typical freeze–thaw condi- low, reaching 56 day strengths of 77 kPa and 157 kPa respec-
tions in the UK, the results are conservative and will prove tively: both failed to match the strengths achieved by lime and
useful when assessing similar alluvial soils in other countries CEM-I at equivalent dosages and curing periods. These
with harsher winters. The results from the freezing–thawing samples were characterised by highly ductile behaviour at
testing are presented in Fig. 4. failure, indicating the need for alkali activation to produce
The dosage of 5% for lime and CEM-I was too low to considerable strength development. Four dosages (2.5%, 5%,
produce any marked strength or durability enhancements, as 7.5% and 10%) of NaOH activated GGBS were assessed to
indicated by the failure of the samples after their first cycle of
testing. Although the 10% lime specimens only survived
2 cycles with significant volumetric reduction, the 10%
CEM-I samples survived all 12 cycles. However, the samples
did experience a volumetric reduction of up to 60%. The
NaOH activation of GGBS was essential to promote pozzo-
lanic conditions and ultimately ensure that samples survived
beyond 3cycles. The 10% GGBS-NaOH mixture was the most
effective binder tested, followed by the 7.5% dosage, as they
provided high levels of resistance to harsh freezing–thawing
cycles. This was confirmed by specimens surviving all testing
cycles and the negligible changes in volume and the moisture
content.
To summarise, GGBS-NaOH at dosages 4 7.5% (particu- Fig. 6. Relationship between unconfined compressive strength and GGBS-
larly 10%) produced the best durability performances for both NaOH dosage, with trend lines shown for 0, 7, 14, 28 and 56 days curing.

Fig. 5. UCS performances for all binder mixtures and dosages used to stabilise Lanton alluvium after all curing periods.
1028 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

identify the most sustainable and optimum dosage in terms of gel formation for hydration reactions in the sand and silt
strength gain. There was no advantage in using either the fractions of the soil is better than in soils with a higher clay
2.5 or 5% dosages due to the strengths achieved for each contents.
curing period (70 kPa after 28 days) closely resembling the The GGBS-NaOH dosage influences strength development,
performance of untreated Lanton alluvium. The 10% GGBS- as can clearly be seen in Fig. 6. Whilst further testing is
NaOH samples exhibited the highest and most rapid strength required to more definitively identify the optimum dosage for
enhancements with curing, with the effects of NaOH activation use in stabilising Lanton alluvium, Fig. 6 uses the strength
first seen after 14 days curing. After 28 and 56 days of curing, results achieved with increasing curing time for the four
impressive UCS values of 2.08 and 2.73 MPa were recorded, GGBS-NaOH dosages in an attempt to approximate an
respectively. optimum dosage. The performance of this new binder sur-
Although the 7.5% GGBS-NaOH binder dosage produced passed that of the CEM-I and lime binders. According to the
lower strengths than the 10% dosage after 7 and 14 days, 28 day strength development trend line presented in Fig. 6, the
significant strength enhancements initiated after 21 days of GGBS-NaOH dosage which satisfies the EuroSoilStab (2002)
curing, producing UCS values of 1.3–1.4 MPa, exceeding the strength criterion would be approximately 5.5%. However, this
strengths for samples stabilised with 10% lime or CEM-I by dosage would be insufficient to produce the 28 day strength of
factors of 3 and 2, respectively. The strengths also met the 778 kPa recorded for 10% CEM-I, in which case an approx-
EuroSoilStab (2002) 28 day 300 kPa requirement. The beha- imate GGBS-NaOH dosage of 6.5% would be required. Thus,
viour of the 7.5% and 10% GGBS-NaOH samples after 28 to satisfy both strength criteria, an optimum GGBS-NaOH
days was brittle upon failure. The 10% GGBS-NaOH samples dosage of 6.5% is envisaged for inclusion within the Lanton
exhibited higher levels of brittle behaviour due to their higher alluvium.
strength, even after 7 days of curing. This contrasts with the Moisture content and porosity also influence strength
more ductile behaviour observed for the other mixtures tested development. Mixtures with higher moisture contents have
after 7 days. lower calcium concentrations (pertaining to GGBS) for cement
Using NaOH with GGBS at dosages Z 7.5% was necessary formation, resulting in lower strengths. Consoli et al. (2011)
to produce significant strength gains. For the 5% dosage, suggested porosity also influences strength by modifying the
strength development commenced after 42 days, reaching 56 number of contact points between soil particles, suggesting
day strengths comparable to 10% CEM-I. However, such that delayed strength gains can be anticipated in soils of higher
delayed strength gains are unacceptable given that construction porosities as well as soils which contain less cement and
specifications require strength enhancements after 28 days therefore are characterized by a less dense network of
(Hansson et al., 2001). cementitious gels.
Considering the low permeability of Lanton alluvium, there Suction may have been generated within the UCS samples
are two factors which may explain why such impressive during their preparation, which may have resulted in higher
strength gains were achieved within stabilised samples. Firstly, sample strengths than anticipated. The DDSM treatment of
the activated GGBS binder was thoroughly mixed into the soil sands using CEM-I has been successfully put into practice
to achieve a high level of homogeneity, in effect facilitating throughout the world for many years, with high strengths
hydration and pozzolanic reactions throughout samples which generally achieved early (within 14 days) and very limited
resulted in a wide distribution of cementitious gels. The second long-term (months-years) strength development. In the case
factor may the even distribution of water throughout samples, of clayey soils, a clay content of 10–50% is considered
partly due to the thorough sample mixing and the particle size necessary in order to achieve sufficient CEC and meet the
distribution of the soil. The access to water for cementitious requirements for surface area characteristics for long-term

Table 4
Average changes in water content observed over 28 days for all untreated and stabilised mixtures.

Soil–binder mixture Water content (%) with increasing curing time Reduction in water content (%)

0 days 28 days

Lanton 24.6 23.9 0.7

Lanton þ 5% Lime 24.1 23.3 0.8
Lanton þ 10% Lime 25.1 21.6 3.5
Lanton þ 5% CEM-I 24.2 22.7 1.5
Lanton þ 10% CEM-I 24.5 19.3 5.3
Lanton þ 5% GGBS 24.8 24.1 0.7
Lanton þ 10% GGBS 24.9 24.6 0.4
Lanton þ 2.5% GGBS-NaOH 24.6 23.5 1.0
Lanton þ 5% GGBS-NaOH 25.6 24.5 1.1
Lanton þ 7.5% GGBS-NaOH 25.1 23.6 1.5
Lanton þ 10% GGBS-NaOH 24.7 22.9 1.8
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1029

Fig. 7. Oedometer compression curves of Lanton alluvium in (a) its untreated remoulded and undisturbed state, and (b) treated with 2.5–10% GGBS-NaOH after 28
days curing. Note: the curves are based on averages taken from triplicate testing.

Table 5
mv, Cc and Cs values for treated and untreated Lanton alluvium after 0 and 28 days.

Soil–binder mixture Coefficient of compressibility (mv), m2/MN Compression index (Cc) Swelling index (Cs)

0 days 28 days 0 days 28 days 0 days 28 days

50 kPa 100 kPa 200 kPa 400 kPa 50 kPa 100 kPa 200 kPa 400 kPa

Lanton remoulded 0.703 0.359 0.221 0.116 0.703 0.359 0.221 0.116 0.132 0.013
Lanton undisturbed 0.799 0.542 0.440 0.255 0.799 0.542 0.440 0.255 0.309 0.030
Lanton þ 2.5% GGBS- 0.238 0.074 0.098 0.055 0.276 0.153 0.073 0.036 0.065 0.039 0.015 0.004
NaOH
Lanton þ 5% GGBS- 0.231 0.150 0.071 0.058 0.230 0.109 0.070 0.029 0.087 0.034 0.011 0.004
NaOH
Lanton þ 7.5% GGBS- 0.222 0.142 0.071 0.046 0.025 0.035 0.016 0.011 0.077 0.016 0.011 0.004
NaOH
Lanton þ 10% GGBS- 0.214 0.133 0.071 0.034 0.025 0.019 0.015 0.012 0.055 0.014 0.009 0.003
NaOH

strength improvements through DDSM (Tutumluer, 2012). 41.4% were observed in samples containing dosages
Conversely, a higher clay contents reduces the permeability 47.5%. The most significant reductions (up to 5.25%) were
of the soil, as does the addition of GGBS. The stabilising of observed in samples stabilised with 10% lime and CEM-I.
soils with a clay content 4 30% may not be as successful as Sub-samples were retained after UCS testing at 28 days and
stabilizing soils with o 30% clay, because the process would tested according to BS 1377: Part 3 (BSI, 1990). Average pH
result in a reduction in the volume of water available to be values recorded for Lanton alluvium ranged between 6.7 and
utilized by the hydration and long-term pozzolanic reactions 7.2 over 56 days, indicating the need for an alkali activator or a
to form C-S-H/C-A-H gels (Hughes et al., 2011). high alkaline binder. Stabilisation using 5–10% CEM-I or lime
produced pH values 4 11.2 which were sustained for the 56
4.4. Moisture content day testing period.
Reductions in the moisture content between 1% and 1.8%
To assess whether samples contained adequate supplies of were noted in samples which achieved high UCS values and
water to allow for continued hydration and cementitious bond durability performances (e.g. 7.5–10% GGBS-NaOH). This
formation, the water content was determined according to BS suggests that cementitious bonds formed through hydration
1377: Part 2 (BSI, 1990). The water content of most samples and pozzolanic reactions, and implies that all sample mixtures
was in the range of 22–25%. According to Table 4, a slight contained sufficient volumes of water for cementitious bonding
reduction in moisture content over 28 days occurred for each reactions during and beyond 56 days.
of the mixtures. Negligible reductions were recorded within all As was reported by Duxson et al. (2007), a relationship may
non-activated samples. For GGBS-NaOH samples, the moist- exist between the decreasing moisture content and strength
ure content was reduced by 1–1.8%, whereas reductions enhancement when Lanton alluvium is stabilised with lime or
1030 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

GGBS-NaOH. As soil water is consumed during the dissolu- embankment fill unit weight of 20 kN/m3 and worst case
tion of geopolymeric reactions and the breakdown of solid shallow groundwater levels (5 mbgl), may be as high as
aluminosilicates to give silicate and aluminate species through 300–400 kPa. Hence, mv values recorded for each sample
alkaline hydrolysis (Duxson et al. (2007), it is not surprising mixture under vertical stresses up to 400 kPa during compres-
that reduction in the water content of GGBS-NaOH samples sibility testing were considered. It is under the vertical stresses
was greater than in non-activated GGBS. By increasing the of less than or equal to 400 kPa where specimens also
GGBS-NaOH dosage within the alluvium, the calcium con- experience the most compressional strain during testing, and
centration within the stabilised material is increased due to the therefore no mv values obtained for vertical stresses (sv) of
GGBS. As a result, the water in the soil is utilised at an 800–1600 kPa have been included.
accelerated rate to form cementitious gels, providing higher Compression curves for the soil in its untreated and GGBS-
strength gains over 28 days than samples with lower GGBS- NaOH stabilised states after 0 and 28 days curing are displayed
NaOH dosages (i.e. o 5%). This observation is in good in Fig. 7. The coefficients of compressibility (mv) values
agreement with the strength results presented in Fig. 5 and obtained for each sample after 0 and 28 days curing are given
the degrees of moisture content reduction within the four in Table 5.
GGBS-NaOH dosage sample mixtures. The compression curves in Fig. 7a illustrate a reduction of
Since the natural water content of Lanton alluvium is 25%, 0.23 in the initial void ratio for the remoulded Lanton
further hydration and pozzolanic reactions can be expected alluvium, compared with its undisturbed counterpart. This
beyond 56 days, requiring further water consumption and reduction can be attributed to the removal of the soil's natural
resulting in the further development of cementitious gel. In the structure, which provides a degree of additional strength to the
long term, Lanton alluvium stabilised with binder dosages material due to fabric and/or inter-particle bonding as a result
o 7.5% may outperform the performances of stabilised soils of the soil's geological history – i.e. depositional and post-
containing water contents of o 25%. Locat et al. (1990) depositional processes (Gasparre and Coop, 2008). A more
attributed such behaviour to solutes being able to migrate significant reduction in the initial void ratio can be seen in
through the soil's pore spaces more easily. This may be the Fig. 7a for all 28 day cured stabilised samples compared with
case for the 5% GGBS-NaOH sample, which started to display the undisturbed (untreated) Lanton alluvium. This is an
encouraging strengths after 42 days. inherent result of densification due to the nature of the DDSM
Locat et al.’s (1990) research implies that stabilising Lanton process, with the orientation of the auger mixer blades
alluvium with 10% GGBS-NaOH closely resembles the rapid providing a degree of compaction. Additionally, by introdu-
strength gain behaviour of soils with a low water content. The cing a fine grained (silty) cementitious GGBS binder into the
use of 7.5% GGBS-NaOH resulted in slower rates of strength alluvium, the PSD curve for the material is altered and the void
development, with 28 day strengths of 1.3 MPa. Compared to spaces within the soil are filled: this has the effect of reducing
the behaviour described by Locat et al.’s (1990) model, these the initial void ratio. However, the increase in material
results correspond well with the slower rates of strength stiffness due to vthe oid ratio reduction is small compared
development typically observed for soils with high water with that provided by the formation of cementitious gels.
contents. For 5% GGBS-NaOH, negligible strength develop- After 0 days, all stabilised mixtures experienced void ratio
ments were observed until 42 days. These findings prove that reductions of 0.1, whereas the void ratios of untreated samples
further study is required over a longer time period to gain a decreased by 0.25 and 0.4 (remoulded and undisturbed
better understanding of the progressive strength development samples, respectively). Although it is known that introducing
of GGBS-NaOH stabilised Lanton alluvium at various a binder via DDSM disturbs the soil, the initial void ratios of
dosages. stabilised samples were remarkably similar to the remoulded
soil. After 28 days of curing, samples containing 2.5% and 5%
4.5. Compressibility and stiffness GGBS-NaOH exhibited slightly improved consolidation beha-
viour compared with that observed after 0 days curing, as
Based on the relatively high strength and durability perfor- proven by the compression indices (Cc) and swelling indices
mances of the GGBS-NaOH binder, compressibility testing (Cs) values in Table 5. Fig. 6b shows samples containing 7.5%
was focussed on the Lanton alluvium when untreated and or 10% GGBS-NaOH displayed a significantly stiffer response
stabilised with GGBS-NaOH. In the UK, Network Rail after 28 days. Cc values for these mixtures decreased
standard RT/CE/S/065 Issue 4 2014 "Examination of Earth- considerably compared with their 0 day values. Cs values for
works" Risk Matrix (Network Rail, 2014) considers railway all four dosages were almost identical. Cc values for 7.5 and
embankment height ranges of: a) o 3 m b) 3–6 m and c) 10% dosage samples are at least one third of the values
6–10 m. Most failures occur at 3–6 m high embankments with obtained for samples containing 2.5 or 5% binder, 9 and 20
slope angles 4 25 degrees. On the UK's railway network, times smaller than the Cc values obtained for the remoulded
embankment heights are known to typically vary between and undisturbed soil, respectively.
3 and 10 m. However, there are locations where embankment The mv results complement the compression curves: all of the
heights are known to exceed this height range, with embank- binder dosages reduced the compressibility of the untreated
ments reaching up to 30 m high. Vertical effective stresses Lanton alluvium. The initial mv values of the untreated undis-
associated with 30 m high embankments, assuming an turbed and disturbed alluvium ranged between 0.7 – 0.8 m2/MN,
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1031

which is typical of normally consolidated alluvium and therefore contributes up to 7% of global CO2 emissions (McLellan
highly compressible (Tomlinson, 2001). The 0 day mv values for et al., 2011), with 900–1100kg of CO2 emitted for every ton of
each dosage ranged between 0.21 m2/MN for vertical stresses cement produced (EPA, 2005). An estimated 4100 million
(sv) of 50 kPa and 0.03 m2/MN for sv ¼ 400kPa. These values tonnes of cement was produced in 2015 (Statista, 2016). The
indicate that each mixture had a medium to very low compres- cement and civil engineering industries are under significant
sibility. For dosages of 2.5 and 5%, negligible changes in mv pressure to reduce CO2 permissions and contribute towards
values were achieved after 0 and 28 days curing. However, there atmospheric carbon capture storage (CCS).
was a significant reduction in mv values after 28 days for samples Using IBPs as alternatives or partial replacements to CEM-I
containing 7.5% or 10% GGBS-NaOH, where the effect of or lime has real potential for reducing the environmental
cementation is apparent. Considerably lower mv values of impact of using cements and lime. According to Hanson
0.01–0.03 m2/MN and 0.01 – 0.02 m2/MN were calculated for (2014), GGBS has impressive environmental sustainability
samples containing 7.5 and 10% binder dosages, respectively, credentials. Compared with the production of lime and
and were deemed to be of very low compressibility. CEM-I, embodied CO2 emissions are reduced by 2 million
These results suggest that the most effective dosage of the tonnes per annum, primary energy usage is reduced during
GGBS-NaOH binder in reducing the initially high compressi- manufacture/processing by 2000 million kWh, 3 million
bility of untreated Lanton alluvium was 10%. However, the tonnes of quarrying is saved and landfill is reduced by almost
7.5% dosage was equally as effective when samples were 2 million tonnes. GGBS has advantages over other IBPs due to
subjected to sv 4200 kPa. Hence, there is little benefit in using the quantity of CEM-I it can replace within a given concrete or
a dosage of 10% over 7.5%, given the similarly small stabilised soil, whilst achieving engineering performances
reductions in void ratio after two loading-unloading cycles. comparable with pure CEM-I (Connell, 2014). 50–70% of
The undisturbed soil possessed a fairly high level of CEM-I may be replaced by GGBS whereas only 20–25% may
sedimentation structure. A slight disturbance or increase in be replaced by PFA. While approximately 75% of CaO in the
loading up to 100 kPa resulted in a sudden and considerable CEM-I component of concretes is able to sequester atmo-
reduction in the void ratio, and therefore significant structural spheric CO2 via carbonation (DTI, 2005), mixing other high-
collapse. This highlights the need for treatment to improve the calcium binders including GGBS and steel slag within urban
high levels of compressibility of the soil. soils has recently proved to have excellent atmospheric CO2
In summary, a significant improvement in compressibility sequestration potential in the UK (Sanna et al., 2012). GGBS
and therefore stiffness was observed after 28 days for 7.5 and and steel slag have a theoretical global CO2 potential uptake of
10% GGBS-NaOH specimens, with mv values decreasing to approximately 80 million tonnes per year (Yan, 2015). CCS in
o 0.03 m2/MN with increasing sv, indicating very low levels urban soils through the incorporation of such slag materials has
of compressibility. The use of 4 7.5% GGBS-NaOH in recently been demonstrated by Washbourne et al. (2012) on
Lanton alluvium produced brittle failure. This behaviour could the Science Central site in Newcastle upon Tyne, UK.
be unfavourable under dynamic loading conditions due to Washbourne et al. (2012) demonstrated that the top metre of
passing rail traffic and earthquakes. However, the strengths soil for the 10 ha site has the potential to capture and store
achieved were lower and more favourable than those recorded 64,000 Mg of atmospheric CO2 in the form of carbonate
by Sargent et al. (2013) for a GGBS-NaOH-Na2SiO3 stabilised minerals, half of which was captured within 3 years at an
artificial alluvium. annual rate of 325 Mg C ha-1. Similar figures were also
calculated for similarly mixed quarry soils (Barrasford, UK);
5. Sustainability assessment namely 300 Mg C ha-1, which was observed down to depths of
three metres over ten years (Manning et al., 2013). Such
New geopolymers must demonstrate commercial and prac- differences in sequestration rates were attributed to different
tical viability as alternatives to CEM-I and lime. Per Global soil mineralogies: urban soils at Science Central contained
Cement (2011), demonstrating GGBS-NaOH for use in DDSM cementitious rubble material (calcium silicates and calcium
on an industrial field scale is essential to proving its practi- hydroxide) which produce greater rates of carbonation than
cality and for developing knowledge regarding the cost Barrasford quarry soil (Manning et al., 2013). Globally,
implications of non-equitable economies of scale. Assurance construction and development site soils are considered able
of this binder's potential for commercialisation should be taken to capture and store 290 million Mg of atmospheric carbon
from the commercialisation of similar alkali-activated GGBS- within soil matrices per year (Renforth et al., 2009).
based geopolymer concretes (“E-Crete”) in Australia (Global However, strength gains within GGBS-stabilised soft
Cement, 2011). alluvial soils require alkali activation. Most alkali materials
for industrial purposes are synthetically sourced with high
5.1. Environmental impact costs and negative environmental impacts. Although the use
of NaOH in this paper promoted pozzolanic conditions and
Although the use and practicality of cement and lime as high strength gains within stabilised samples, its production
binders is well proven in DDSM, their continued usage is involves the electrolysis of brine, which is energy intensive
environmentally unsustainable. Cement and lime production and produces considerable CO2 emissions. During its
is associated with high carbon costs. Cement production production, there is also the risk of contaminating local
1032 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

aquatic environments. A more sustainable NaOH source the international market price of CEM-I (£50 per ton), it is
could be the by-product from existing chlor-alkali plants only marginally more expensive than the UK price of £70 per
(Kumar et al., 2012). ton. In order to make the GGBS-NaOH break even with CEM-
Significant CO2 emission reductions of up to 64% are I in the UK, a price reduction of only 7% is required, which
estimated for Australian geopolymers compared with CEM-I would equate to a binder dosage of 5.5% in Lanton alluvium.
(McLellan et al., 2011). There are carbon footprint implica- This advocates that GGBS-NaOH is potentially a financially
tions associated with the brine electrolysis (membrane) more sustainable alternative to CEM-I. McLellan et al. (2011)
technique used to produce NaOH. However, the quantities demonstrated that for a number of NaOH-bearing geopolymers
of NaOH required in this research were small (0.83–3.3% by developed in Australia, the total financial cost of these
dry weight or 12.4–49.4 kg m-3) compared with the quan- mixtures was not much greater than that for CEM-I. However
tities of CEM-I typically used in more traditional binder their transportation costs were considerably higher than those
mixtures for DDSM treatment of soft soils. According to associated with CEM-I, due to the significant transportation
Topolnicki (2004), cement dosages of 120 – 300 kg m-3 may distances between binder source and the treatment site.
be used in most soils, with the exception of sludges, peats Turner and Collins (2013) stated that for an Australian
and organic silts/clays, where dosages of 250 – 400 kg m-3 geopolymer concrete comprising PFA, NaOH, sodium silicate
are recommended. and a superplasticiser, a carbon foot print reduction of 9% was
The GGBS CO2 reductions far outweigh the carbon measured against a comparable strength CEM-I concrete.
footprint produced by manufacture of NaOH. There is also There is a confliction between the reduced carbon footprint
potential for GGBS-NaOH to sequester atmospheric carbon estimates provided by McLellan et al. (2011) and Turner and
within the stabilised soil, which assists in driving the binder Collins (2013). However, the geopolymers assessed by Turner
towards commercialisation. Further work is required on the and Collins (2013) contained considerable quantities of sodium
effects of carbonation on the long-term durability, strength and silicate, which is expensive and has significant negative
mineralogical development of stabilised soft alluvial soils and environmental impacts (Habert et al., 2011). Thus, the reduc-
how much atmospheric CO2 such mixtures can capture. tions in the GGBS-NaOH carbon footprint compared with
CEM-I is more comparable with estimates provided by
5.2. Financial costs McLellan et al. (2011). Based on the available evidence,
GGBS-NaOH has a smaller carbon footprint and lower
Although the engineering performances observed for financial cost then CEM-I, which highlights its potential as a
GGBS-NaOH meet EuroSoilStab (2002) strength criteria and commercially viable alternative to CEM-I.
surpass those exhibited by CEM-I and lime, it is also crucial to
demonstrate that its overall cost is financially competitive. 6. Engineering practicality
There are considerable stockpiles of GGBS in the UK, given
the country's active pig iron and steel manufacturing industry. A key factor to consider in assessing the potential for the
This makes GGBS sustainable in the UK for use in geopoly- GGBS-NaOH binder to be used instead of lime or CEM-I for
mers as stockpiles are continuously replenished and thus future DDSM projects is engineering practicality. The applica-
avoids expensive overseas sourcing. GGBS also requires little tion of chemical treatment in stabilising soft soils is generally
post-production processing (i.e. ground granulation) to make it limited to those characterised by low organic and sulphate
suitable for mixing. contents and relatively high clay contents. These criteria have
Solid NaOH is available as flakes, pellets or pearls and is been defined to ensure that cation exchange and long term
commercially available in raw bulk production and scientific cementitious reactions occur within the stabilised material.
forms, the difference being that the latter has been purified. Although the Lanton alluvium possessed unfavourably high
Hence, whilst the cost of scientific NaOH is much higher, its moisture and silt-sand contents, using a GGBS-NaOH binder
effectiveness as an alkali activator is equal to that of industrial dosage Z 7.5% proved effective in producing high levels of
grade. For laboratory studies involving various chemicals, strength and durability within the soil post 21 days of curing.
scientific NaOH is recommended as it is in a much purer form. The enhanced mechanical performances of the GGBS-NaOH
For large scale applications, raw bulk production NaOH is stabilised Lanton soil were at least 2–3 times better than those
more financially feasible. Additionally, the electrical energy observed for samples mixed with traditional binders, whilst
consumed per ton of NaOH produced is 1.285 kWh, which is meeting strength criteria defined by EuroSoilStab (2002).
almost 60 times lower than that required per ton of CEM-I Another component which needs to be considered in
(Madlool et al., 2011). assessing the engineering practicality of the GGBS-NaOH
The international market prices of these materials at the time binder for use in DDSM is whether any modifications to
of writing (July 2014 US$:£ currency exchange rates) were existing plant and equipment would be required. GGBS comes
£25–60 per ton of GGBS and £175 per ton of NaOH (Alibaba, in the form of a dried fine powder and NaOH pellets/flakes are
2014). The use of the lowest available prices produces a solid particulates. Hence, it is anticipated that the GGBS-
significantly lower price of £75 per ton for the GGBS-NaOH NaOH binder could be substituted for CEM-I and lime with
binder when using the 2/3 GGBS: 1/3 NaOH ratio, the same as relative ease with only minimal plant and equipment modifica-
was used in this paper. Whilst this price is more expensive than tions required to ensure that the efficiency of the binder
 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034 1033

delivery during deep mixing is maintained. However, the use the binder becoming less environmentally and financially
of NaOH on site can pose some health and safety risks for sustainable. The use of GGBS-NaOH has the potential of
DDSM operatives, such as skin irritation (dermatitis) and eye becoming a more sustainable alternative than the continued use
damage. Thus, operatives should wear appropriate personal of lime and CEM-I, suggesting it has significant commercial
protective gear, including disposable rubber gloves, safety potential. Although the current UK and international market
goggles/glasses, disposable overalls and dust masks. prices for the GGBS-NaOH binder mixture are higher than
Based on the aforementioned factors, the GGBS-NaOH CEM-I, the financial costs in terms of raw materials, energy
binder potentially has an impressive level of engineering consumption, transport and the associated CO2 emissions
practicality, exceeding that of lime, CEM-I and other waste- incurred by CEM-I production far outweigh those incurred
based binders. In addition to being used as a binder to stabilise by GGBS-NaOH. Ultimately, this new binder aims to achieve
soils, it may also be used as a partial/total substitute to CEM-I a balance between the desire for reducing costs and a desire for
in concrete mixtures. However, the binder's engineering a binder with low embodied carbon.
practicality in terms of soil stabilisation requires further The cost of any DDSM project involving alkali-activated
research. The performance of the GGBS-NaOH binder needs mixtures can be quite high. It is common for geopolymer
to be assessed when incorporated in soils characterised by a components to come from various locations. The distances
higher organic and sulphate content than Lanton alluvium. between sourcing plants and stabilisation sites can be con-
siderable. As modern transportation costs are high, careful
7. Conclusions planning must be conducted to minimise delivery distances:
this will help geopolymers to become more sustainable and
Alkali-activated IBPs show great potential for use as competitive than lime or CEM-I.
replacements for CEM-I and lime in stabilising alluvial soils.
The strengths and stiffnesses achieved by GGBS-NaOH
significantly improved with curing, particularly when using
References
dosages 4 7.5% by dry weight (4 107 kg m-3), which com-
fortably met or surpassed those exhibited by samples stabilised Alibaba, 2014. International market prices for sodium hydroxide. [online].
with equivalent quantities of lime/CEM-I and met criteria Available at: 〈http://www.alibaba.com/premium/sodium_hydroxide.html?
defined by EuroSoilStab (2002). For any of the binder Uptime¼20140708&ptsid¼ 1012000053229821&crea ¼47501223981&pl
mixtures tested, a dosage o 7.5% is not recommended given ac¼ &netw¼ s&device¼ c&ptscode¼0110202010030001〉 (last accessed
the natural chemical and geotechnical properties of Lanton 19.07.14).
American Society for Testing and Materials (ASTM), 2015a. Designation: D
alluvium, along with the aggressiveness of the ASTM's 559/D559M – 15 Standard Test Methods for Wetting and Drying
durability testing programmes, which may occur regularly in Compacted Soil-Cement Mixtures. Annual Book of ASTM Standards.
some countries. 4.08. West Conshocken, PA.
Some engineering specifications require high strengths and American Society for Testing and Materials (ASTM), 2015b. Designation: D
560/560M – 15 Standard Test Methods for Freezing and Thawing
stiffnesses and such cases, using GGBS-NaOH would be the
Compacted Soil-Cement Mixtures. Annual Book of ASTM Standards.
most logical replacement for CEM-I or lime. However, in 4.08. West Conshocken: PA.
addition to engineering scenarios characterised by static loads, Bernal, S.A., 2014. The resistance of alkali-activated cement-based binders to
there are those such as high-speed rail lines where the high- carbonation. In: Pacheco-Torgal, F., Labrincha, J., Leonelli, C., Palomo,
frequency dynamic loading of the ground requires slightly A., Chindaprasit, P. (Eds.), Handbook of Alkali Activated Cements,
lower stiffnesses to prevent brittle failure. In such cases, the Mortars and Concretes. Woodhead Publishing.
Brunauer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multi-
use of GGBS-NaOH at lower dosages would be preferred. molecular layers. J. Am. Chem. Soc. 60, 309–319.
NaOH is very effective in activating GGBS and producing BSI, 1990. BS: 1377 Incorporating Amendment No. 1, Methods of test for
long term high strength gains, and proved to be equally as Soils for Civil Engineering Purposes. British Standards Institution, Milton
effective as the sodium silicate-sodium hydroxide activator Keynes.
BSI, 1996. BS: 7755-3.12, ISO 13536:1995. Soil Quality. Chemical methods.
used by Sargent et al. (2013) in raising soil pH, promoting
Determination of the potential cation exchange capacity and exchangeable
pozzolanic conditions and long term strength gains. The cations using barium chloride solution buffered at pH ¼ 8.1. British
GGBS based binders and solid NaOH activator pellets could Standards Institution, Milton Keynes.
be substituted for CEM-I and lime in DDSM with minimal BSI, 2015. BS: 5930, Incorporating Amendment No. 1, Code of Practice for
modification to pre-existing plants and equipment, or a Site Investigations. British Standards Institution, Milton Keynes.
reduction in the efficiency of mixing. Clarke, D., Smethurst, J.A., 2010. Effects of climate change on cycles of
wetting and drying in engineered clay slopes in England. Q. J. Eng. Geol.
The dosage of GGBS-NaOH for use within soils will differ Hydrol. 43, 473–486.
between DDSM projects. For soils more problematic than Connell, M., 2014. Hanson Cements, UK. Personal communication.
Lanton alluvium, higher dosages would be required to achieve Consoli, N.C., Rosa, D.A., Cruz, R.C., Rosa, A.D., 2011. Water
high strengths. Additionally, the ratio between GGBS and content, porosity and cement content as parameters controlling
NaOH in the binder will also require customisation, with strength of artificially cemented silty soil. Eng. Geol. 122,
328–333. http://dx.doi.org/10.1016/j.enggeo.2011.05.017.
higher NaOH concentrations required to stabilise soils with a Cristelo, N., Glendinning, S., Teixeira Pinto, A., 2011. Deep soft soil improvement
low pH. Care must be taken in designing the GGBS-NaOH by alkaline activation. Proceedings of the Institution of Civil Engineers, Ground
binder mixture, as high NaOH concentrations would result in Improvement. 164 (2), 73–82. http://dx.doi.org/10.1680/grim.900032.
1034 P. Sargent et al. / Soils and Foundations 56 (2016) 1021–1034

Davidson, L.K., Demirel, T., Handy, R.I., 1965. Soil pulverization and lime help-europes-ailing-economies-but-the-wider-implications-of-the-price-dro
migration in soil lime stabilisation. Highw. Res. Rec. 92, 103–126. p-remain-to-be-seen/〉 (last accessed 02.08.15).
DECC – Department of Energy & Climate Change, 2015. Statistical press Madlool, N.A., Saidur, R., Hossain, M.S., Rahim, N.A., 2011. A critical
release: UK Energy Statistics, Q1 2015, ref: 2015/020. [online]. Available review on energy use and savings in the cement industries. Renew. Sustain.
at: 〈https://www.gov.uk/government/uploads/system/uploads/attachment_ Energy Rev. 15, 2042–2060. http://dx.doi.org/10.1016/j.rser.2011.01.005.
data/file/437747/Press_Notice_June_15.pdf〉 (last accessed 02.08.15). Manning, D.A.C., Renforth, P., Lopez-Capel, E., Robertson, S., Ghazireh, N., 2013.
DTI – Danish Technical Institute, 2005. Nordic Innovation Centre Project: NI- Carbonate precipitation in artificial soils produced from basaltic quarry fines and
project 03018 “CO2 uptake during the concrete life cycle”. Guidelines: composts: an opportunity for passive carbon sequestration. Int. J. Greenh. Gas.
Uptake of carbon dioxide in the life cycle inventory of concrete. October Control 17, 309–317. http://dx.doi.org/10.1016/j.ijggc.2013.05.012.
2005. Prepared by Pommer, K. and Pade, C. ISBN: 87-7756-757-9. McLellan, B.C., Williams, R.P., Lay, J., van Riessen, A., Corder, G.D.,
Duxson, P., Fernandez-Jimenez, A., Provis, J.L., Lukey, G.C., Palomo, 2011. Costs and carbon emissions for geopolymer pastes in compar-
A., Deventer, J.S.J., 2007. Geopolymer technology: the current state of ison to ordinary portland cement. J. Clean. Prod. 19,
the art. J. Mater. Sci. Adv. Geopolymer Sci. Technol. 42, 1080–1090. http://dx.doi.org/10.1016/j.jclepro.2011.02.010.
2917–2933. http://dx.doi.org/10.1007/s10853-006-0637-z. Meunier, A., 2005. Clays. Springer Publishing, Berlin, Heidelberg, New York
EIA – Energy Information Administration, 2015. Growth in residential (ISBN 3-540-21667-7).
electricity prices highest in 6 years, but expected to slow in 2015. [online]. Nair, S. and Little, D.N., 2009. Water as the Key to Expansion of Ettringite in
Available at: 〈http://ww.eia.gov/todayinenergy/detail.cfm?Id ¼20372〉 (last Cementitious Materials. Transportation Research Record: Journal of the
accessed 2 August 2015). Transportation Research Board, No. 2104, Transportation Research Board
EPA – United States Environmental Protection Agency, 2005. AP 42 – of the National Academies, Washington DC, 55-62. doi:10.3141/2104-06.
Compilation of Air Pollutant Emission Factors, Volume 1: stationary Point Network Rail (UK), 2014. Network Rail standard RT/CE/S/065 Issue 4 2014
and Area Sources. Environmental Protection Agency, Washington DC. "Examination of Earthworks".
EPA - United States Environmental Protection Agency, 2015. Increasing Newman, J., Choo, B.S., 2003. Advanced Concrete Technology. Elsevier
greenhouse gas concentrations will have many effects. [online]. Available Butterworth Heinemann publishing (ISBN 0-7506-5103-2).
at: 〈http://www.epa.gov/climatechange/science/future.html〉 (last accessed Palomo, A., Grutzeck, M.W., Blanco, M.T., 1999. Alkali-activated fly ashes. A
01.06.15). cement for the future. Cem. Concr. Res. 29, 1323–1329.
EuroSoilStab, 2002. Development of design and construction methods to Renforth, P., Manning, D.A.C., Lopez-Capel, E., 2009. Carbonate precipitation
stabilise soft organic soils. Design Guide Soft Soil Stabilisation (CT97- in artificial soils as a sink for atmospheric carbon dioxide. Appl. Geochem.
0351. Project No.: BE 96-3177). 24 (2009), 1757–1764.
Gasparre, A., Coop, M.R., 2008. Quantification of the effects of structure on Reuters, 2015. Is coal’s decline permanent? By Mike Corones – 6 January
the compression of a stiff clay. Can. Geotech. J. 45, 1324–1334. 2015. [online]. Available at 〈http://blogs.reuters.com/data-dive/2015/01/06/
Global Cement, 2011. Geopolymer concrete – A commercial reality (2 is-coals-decline-permanent/〉 (last accessed 2 August 2015).
February 2011). [online]. Available at: o 〈http://www.globalcement.com/ Sanna, A., Dri, M., Hall, M.R., Maroto-Valer, M., 2012. Waste materials for
magazine/articles/316-geopolymer-concrete-a-commercial-reality〉 (last carbon capture and storage by mineralisation (CCSM) – a UK perspective.
accessed 10.07.14). Appl. Energy. 99 (2012), 545–554.
Habert, G., d’Espinose de Lacaillerie, J.B., Roussel, N., 2011. An Sargent, P., Hughes, P.N., Rouainia, M., White, M., 2013. The use of alkali
environmental evaluation of geopolymer based concrete prod- activated waste binders in enhancing the mechanical properties and
uction: reviewing current research trends. J. Clean. Prod. 19, durability of soft alluvial soils. Eng. Geol. 152, 96–108.
1229–1238. http://dx.doi.org/10.1016/j.clepro.2011.03.012. Shi, C., Krivenko, P.V., Roy, D., 2006. Alkali-Activated Cements and
Hanson Cements (Heidelberg Cement Group), 2014. Regen GGBS. [online]. Concretes. Taylor & Francis Publishing (ISBN 0-415-70004-3).
Available at: 〈http://www.heidelbergcement.com/uk/en/hanson/products/ Song, K.-I., Song, J.-K., Yeon Lee, B., Yang, K.-H., 2014. (Article ID
cements/ggbs_and_related_products/regen_ggbs.htm〉 (last accessed 326458). Carbonation characteristics of alkali-activated blast-furnace slag
10.07.14). mortar. Advances in Materials Science and Engineering, vol. 2014.
Hanson Cements (Heidelberg Cement Group), 2016. GGBS Material Safety Hindawi Publishing Corporationhttp://dx.doi.org/10.1155/2014/326458.
Data Sheet (MSDS), updated May 2015. Available at: 〈http://www.hanson. Statista, 2016. . United States and world cement production in 2010–2015.
co.uk/en/products/regen-ggbs〉 (last accessed 08.07.16). [online]. Available at: 〈http://www.statista.com/statistics/219343/cement-
Hansson, T., Parry, L., Graham, M., Troughton, V. Eriksson, H., 2001. Limix: production-worldwide/〉 (last accessed 14 February 2016).
a deep dry mixing system used at Channel Tunnel Rail Contract 440. In: Tomlinson, M.J., 2001. Foundation Design and Construction seventh edition
Proceedings of Underground Construction 2001 Symposium and Exhibi- Pearson Education, Prentice Hall (ISBN 0-13-031180-4).
tion, London. Institute of Materials, Minerals and Mining, London. Topolnicki, M., 2004. In situ soil mixing. In: Moseley, M.P., Kirsch, K. (Eds.),
Hughes, P.N., Glendinning, S., Manning, D.A.C., White, M.L., 2011. Use of Within “Ground Improvement” 2nd edition Spon Press, pp.
red gypsum in soil mixing engineering applications. Proc. Inst. Civil. Eng., 331–428 (ISBN 0-415-27455-9).
Geotech. Eng. 164, 223–234. http://dx.doi.org/10.1680/geng.10.00061. Tsuchida, H., 1970. Prediction and countermeasure against the liquefaction in
Imbabi, M.S., Carrigan, C., McKenna, S., 2012. Trends and developments in sand deposits. Abstract of the seminar in the Port and Harbor Research
green cement and concrete technology. Int. J. Sustain. Built Environ. 1, Institute (in Japanese).
194–216. http://dx.doi.org/10.1016/j.ijsbe.2013.05.001. Turner, L.K., Collins, F.G., 2013. Carbon dioxide equivalent (CO2-e) emissions: a
Kumar, S., Drozd, V., Saxena, S.K., 2012. Catalytic studies of sodium comparison between geopolymers and OPC cement concrete. Constr. Build.
hydroxide and carbon monoxide reaction. Catalysts. 2, Mater. 43, 125–130. http://dx.doi.org/10.1016/j.conbuildmat.2013.01.023.
532–543. http://dx.doi.org/10.3390/catal2040532. Tutumluer, E., 2012. Short course notes for Geotechnical Aspects of Pavement
Locat, J., Berube, M.A., Choquette, M., 1990. Laboratory investigations on the Design and Construction – ASCE GeoCongress 2012. “State of the Art and
lime stabilization of sensitive clays: shear strength development. Can. Practice in Geotechnical Engineering” 25 – 29 March. .
Geotech. J. 27, 294–304. Washbourne, C.-L., Renforth, P., Manning, D.A.C., 2012. Investigating
LSE – London School of Economics, 2015. EUROPP – Falling oil prices carbonate formation in urban soils as a method for capture and storage
should help Europe’s ailing economies, but the wider implications of the of atmospheric carbon. Sci. Total Environ. 431 (2012), 166–175.
price drop remain to be seen, by Bert Scholtens. [online]. Available Yan, J., 2015. Handbook of Clean Energy Systems: vol. 5 – Energy Storage.
at 〈http://blogs.lse.ac.uk/europpblog/2015/01/21/falling-oil-prices-should- John Wiley & Sons Publishing Ltd. (ISBN: 978-1-118-38858-7.s).