Minerals Engineering 29 (2012) 89–104

Contents lists available at SciVerse ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Technical and commercial progress in the adoption of geopolymer cement

Jannie S.J. van Deventer a,b,⇑, John L. Provis a, Peter Duxson b
a
Department of Chemical & Biomolecular Engineering, University of Melbourne, Victoria 3010, Australia
b
Zeobond Pty. Ltd., P.O. Box 210, Somerton, Victoria 3062, Australia

a r t i c l e i n f o a b s t r a c t

Article history: If formulated optimally, geopolymer cement made from fly ash, metallurgical slags and natural pozzolans
Available online 12 October 2011 could reduce by 80% the CO2 emissions associated with the manufacturing of cement. However, almost all
standards and design codes governing the use of cementitious binders and concrete in construction are
Keywords: based on the use of Portland cement. The 100+ year track record of in-service application of Portland
Fine particle processing cement is inherently assumed to validate the protocols used for accelerated durability testing. Moreover,
Environmental the entire supply chain associated with cementitious materials is based on the production of Portland
Recycling
cement. The geopolymerisation of aluminosilicates constitutes a radical change in construction materials
Waste processing
Particle size
chemistry and synthesis pathways compared with the calcium silicate hydrate chemistry which under-
pins Portland cement. Consequently, there are regulatory, supply chain, product confidence and technical
barriers which must be overcome before geopolymer cement could be widely adopted. High profile dem-
onstration projects in Australia have highlighted the complex regulatory, asset management, liability and
industry stakeholder engagement process required to commercialise geopolymer cement. While the
scale-up from the laboratory to the real-world is technically challenging, the core challenge is the
scale-up of industry participation and acceptance of geopolymer cement. Demand pull by a carbon con-
scious market continues to be the key driver for the short term adoption of geopolymer cement. In the
absence of an in-service track record comparable in scale and longevity to Portland cement, research is
essential to validate durability testing methodology and improve geopolymer cement technology. Colloid
and interface science, gel chemistry, phase formation, reaction kinetics, transport phenomena, comminu-
tion, particle packing and rheology, which are familiar concepts to minerals engineers, are also key build-
ing blocks in the development of geopolymer knowledge. Analysis of the nanostructure of geopolymer
gels has enabled the tailored selection of geopolymer precursors and the design of alkali activator com-
position, aiding in establishing the relationship between geopolymer gel microstructure and durability.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Under the 2009 Copenhagen Accord, more than 120 countries
have agreed to keep global average temperature increase below
2 °C. This maximum acceptable temperature increase is based on
recommendations from numerous scientific studies, warning that
increases in excess of 2 °C can trigger dangerous anthropogenic
interference with the climate system, including climate-tipping
points, with unmanageable consequences to water supply, agricul-
tural productivity, sea-level rise, human habitability, and global
security (Rockström et al., 2009). This is an urgent goal, because
many scientists estimate that the concentrations of CO2 and other
climate-forcing substances in the atmosphere already exceed the
⇑ Corresponding author at: Department of Chemical & Biomolecular Engineering,
University of Melbourne, Victoria 3010, Australia. Tel.: +61 3 9303 7777; fax: +61 3
9303 7644.
E-mail address: jannie@zeobond.com (J.S.J. van Deventer).
safe level (Richardson et al., 2009). In contrast, there remain many
sceptics, including in the minerals industry, who do not accept that
climate change is a result of human activity (Plimer, 2009). This pa-
per does not aim to contribute to such debate, but instead assumes
that there is economic and social benefit in reducing CO2 emissions
and valorising waste materials at the same time, and develops dis-
cussions from this viewpoint.
Concrete made from Ordinary Portland Cement (OPC), including
its blends with mineral admixtures, is second only to water as the
commodity most used by mankind today. Global cement produc-
tion in 2008 was around 2.6 billion tonnes (Freedonia Group,
2009), contributing conservatively 5–8% of global anthropogenic
CO2 emissions (WWF-Lafarge Conservation Partnership, 2008;
Scrivener and Kirkpatrick, 2008), and the rapidly increasing de-
mand for advanced civil infrastructure in China, India, the Middle
East and the developing world is expected to expand the cement
and concrete industries significantly (Taylor et al., 2006). CO2 emis-
sions are due mainly to the decomposition of limestone and

0892-6875/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2011.09.009
90 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104

combustion of fossil fuels during cement production; grinding and ing synchrotron-based and other nanostructural characterisation
transport are also significant contributors to the environmental of geopolymers to durability and engineering properties, identify-
footprint of the cement industry. ing the technical, commercial and regulatory barriers to industrial
The cement industry has started to acknowledge the role of adoption, and reviewing progress made in Australia along the path
alternative binders in a carbon constrained industry, given that to commercialisation. This will be addressed from a joint research-
there are significant reductions in CO2 emissions and also some commercialisation viewpoint, with a particular focus on the areas
advantages in performance only offered by these alternative bind- of research which are specifically beginning (or continuing) to im-
ing systems (Damtoft et al., 2008). Historically, the driver for com- pact developments in the commercial arena. The underlying
petition in the construction materials industry has been cost assumption in research papers and grant applications is that good
reduction, in which case alternative binders, starting from a low research will necessarily lead to adoption in industry, which is far
volume basis, could never compete against large-scale OPC produc- from the situation in reality. This paper will use an experiential and
tion. However, CO2 abatement and technical features are now play- tutorial style to show that the interplay between research problem
ing a major role in growth of alternative binder systems. Juenger identification, technical development work and commercial strat-
et al. (in press) recently presented a review of potential alterna- egy is as important as high-quality fundamental research, and that
tives to OPC technology, including calcium sulphoaluminate ce- both are required in the implementation of a new class of binders
ments, magnesium cements, the magnesium phosphate system on a large scale.
and alkali activated materials (AAM) or geopolymers.
Calcium sulphoaluminate cements are made from clinkers that
include ye’elimite as a primary phase, which requires a lower
2. The role of chemical research in the commercialisation of
amount of limestone and lower fuel consumption. Their commer-
geopolymers
cial use in expansive cements and ultra-high early strength ce-
ments has been pioneered in China. Magnesium-based cements
The commercial implementation of geopolymer technology in
and magnesium phosphate cements have been used in niche appli-
Australia is currently being driven by multiple teams operating
cations and can also give superior fire resistance, with much lower
in different parts of the country. Curtin University and the Centre
CO2 emissions than OPC. The use of magnesium silicate hydrates
for Sustainable Resource Processing, based in Perth, have made ad-
together with magnesium carbonate as a ‘carbon negative con-
vances in this area over the past decade and conducted a number
crete’ is also attracting commercial attention at present (Gartner
of trial pours in recent years. The Melbourne-based company, Zeo-
and Macphee, 2011). However, many of these alternative binders
bond, developed its own pilot-scale production facilities in 2007
require a new supply chain for raw materials, the development
and now supplies its concrete, E-Crete™, to major civil infrastruc-
of new chemical admixtures, regulatory approval, the development
ture projects including freeway expansion works and bridge con-
of new durability testing protocols, and an in-service track record
struction and repairs under licence. E-Crete™ utilises for its
before they would be adopted widely by industry.
binder a blend of fly ash and ground blast furnace slag, with com-
Geopolymer cement faces the same obstacles, but has a longer
binations of proprietary alkaline activating components which are
in-service track record (Shi et al., 2006; Van Deventer et al.,
tailored for specific raw materials and products.
2010a; Xu et al., 2008), supported by an expanding body of funda-
There are various technical and commercial factors driving the
mental research relating gel chemistry and nanostructure to dura-
commercial adoption of geopolymer technology, and demand pull,
bility. In geopolymer chemistry, the reactive aluminosilicate
led by a carbon-conscious end-user market, continues to be the key
phases present in precursors including fly ash from coal combus-
driver for the short term adoption of geopolymer concrete. Coun-
tion, metallurgical (including blast furnace) slag, calcined clays
ter-balancing this demand driver is the inherent resistance of the
and/or volcanic ash are reacted with alkaline reagents including al-
civil construction industry to the large-scale adoption of new prod-
kali metal silicates, hydroxides, carbonates, and/or aluminates
ucts, where time and demonstration on relevant industrial scales
(Provis and Van Deventer, 2009) to form aluminosilicate gel phases
are prerequisites for practical credibility, as well as the cost impli-
with varying (but generally low) degrees of crystalline zeolite for-
cations of the absence of economies of scale.
mation. Geopolymer concrete has been widely reported to display
A detailed chemical understanding of the properties of geopoly-
high resistance to fire and acids, and does not produce the high
mer binders, in particular in areas such as control of setting time,
evolution of reaction heat associated with OPC concrete, reducing
workability and durability, also plays an enabling role in the com-
cost and potential cracking issues when the material is placed in
mercialisation process. There is a growing volume of scientific lit-
large volumes. The benefits of geopolymerisation when compared
erature exploring the properties of geopolymeric materials on the
with OPC technology are largely based around the ability to valo-
laboratory scale. Unfortunately, much of this information has lim-
rise high-volume industrial waste streams into high-performance
ited direct value in commercial adoption; geopolymer concrete
concretes, with a highly significant reduction in CO2 emissions
that performs adequately according to all standards can be syn-
(Duxson et al., 2007a). Fly ash and slag appear at present to be
thesised easily in a laboratory, while it is much more difficult to
the most promising precursors for large-scale industrial produc-
reproduce such performance in a commercially and practically fea-
tion of geopolymer cement due to the more favourable rheological
sible form in real-world applications. Indeed, closing this apparent
properties and lower water demand achievable when compared to
disparity and gap between the laboratory and the real-world has
mixes based on calcined clays (Provis et al., 2010).
been the focus of much research conducted by the authors. Being
The history, chemical principles, reaction phenomena and engi-
able to achieve this challenging goal has thus unlocked the com-
neering properties of geopolymer concrete have been reviewed
mercial value of geopolymer technology.
extensively (Davidovits, 2009; Duxson et al., 2007b; Komnitsas
and Zaharaki, 2007; Pacheco-Torgal et al., 2008a,b; Provis and
Van Deventer, 2009; Shi et al., 2011). With the core focus of these
recent reviews being laboratory research on geopolymer binders 3. Developments in geopolymer gel phase chemistry
rather than the production of concrete, it is not the aim of this re-
view paper to duplicate such analysis or discussion. Instead, this This section summarises selected research advances in geopoly-
paper will fill a gap in the current literature by explaining the rel- mer gel chemistry and also identifies topics that require further
evance of particle technology in geopolymer concrete design, link- work, as depicted in Fig. 1.
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Fig. 1. Conceptual diagram showing geopolymer gel phase processes.
3.1. Precursor design
Based on data obtained from the literature, Duxson and Provis
(2008) proposed an ideal composition range for glassy aluminosil-
icate precursors containing network-modifying cations (particu-
larly calcium, magnesium, sodium and potassium) in order to
give sufficiently high solubility to supply the necessary aluminium
into the growing geopolymer gel. This control of aluminium avail-
ability has been highlighted as being necessary to enable the
growth of geopolymer gel particulates, leading to cross-linking,
hardening and strength development (Fernández-Jiménez et al.,
2006; Provis and Van Deventer, 2007). The need for addition of a
separate alkali source may be either greatly reduced or even elim-
inated if the correct glass, or combination of glasses, can be selec-
tively synthesised (Buchwald and Wiercx, 2010). Duxson and
Provis (2008) postulated that this may be achieved by addition of
components into pulverised coal prior to combustion, or by the
selective manufacture of a highly reactive raw material (Van
Deventer et al., 2010b) that can be blended with less-reactive
raw materials to provide a geopolymer cement to a given
specification.
Keyte (2008) determined that the aluminosilicate glass particles
present within Class F fly ashes predominantly consisted of two
intimately intermixed amorphous phases, one silica-rich and one
very alumina-rich, with composition close to Al6Si2O13 and resem-
bling mullite. By applying devitrification to geopolymers, Keyte
(2008) showed that the poor performance of certain fly ashes in
geopolymer synthesis is related to the low aluminium content of
the amorphous phases present in the precursor, resulting in a high
silicon to aluminium ratio in the formed geopolymer. The coordi-
nation environment of aluminium in geopolymer precursors is a
complex issue, and plays a key role in controlling the availability
of Al in geopolymerisation; this is an area of active research and
debate (White et al., 2011).
91
Clearly, there is a need for a workable one-part (‘just add
water’) mix design if geopolymeric cements and concretes are to
achieve widespread market penetration, as this would greatly sim-
plify the process of handling and distributing the components of a
geopolymer binder compared to the use of two-part systems
involving large volumes of alkaline liquids. Various methodologies
for designing the necessary precursors have been proposed
(Koloušek et al., 2007; Duxson and Provis, 2008; O’Connor and
MacKenzie, 2010; Van Deventer et al., 2010b), but its successful
commercial implementation still remains to be demonstrated.
3.2. Binder phase chemistry
By studying the remnant fly ash and slag particles left embed-
ded in a hardened geopolymer binder using scanning electron
microscopy (SEM), Lloyd et al. (2009a,b) observed that calcium is
active during alkali activation, and that, if discrete high calcium
binder regions are formed (rather than the calcium being incorpo-
rated into the main aluminosilicate gel binder), it is on a nanome-
tre rather than a micron length scale. Yip et al. (2005, 2008)
suggested that both geopolymeric (alkali aluminosilicate) gel and
calcium silicate hydrate (C–S–H) gel co-exist at low alkalinities,
while geopolymeric gel appeared to be the dominant product at
high alkalinities, while Buchwald et al. (2007) also showed gel
coexistence at relatively high alkalinity. Recent work from
García-Lodeiro et al. (2011) shows different trends with respect
to alkalinity, with C–A–S–H gel proposed to be more stable than
N–A–S–H at high pH and further work is certainly required in this
area to resolve the remaining disagreements. Provis et al. (2009)
did identify discrete Ca-rich regions within a hydroxide-activated
Class F fly ash binder using 80 nm-resolution X-ray fluorescence
microscopy; these regions were not observed in silicate-activated
samples, and were proposed to be related to the precipitation of
Ca(OH)2 in poorly crystalline form during the very early stages of
reaction of the fly ash particles in highly alkaline environments,
in contrast to its wider distribution throughout the gel formed at
lower alkalinity.
Using the results of synchrotron X-ray diffraction, Oh et al.
(2010) proposed that the geopolymeric gel contains zeolitic pre-
cursors related to disordered forms of the ABC-6 family of zeolites,
depending on the pH environment of the paste. Despite a high cal-
cium content, only very weak C–S–H(I) peaks and no Ca(OH)2
phase were found in their Class C fly ash paste when activated with
sodium hydroxide solution. This implies that the calcium in class C
fly ash did not dissolve in the activator solution as readily as the
calcium in the slag; slag-based systems yielded much higher levels
of crystalline C–S–H(I). The high calcium content in the Class C fly
ash seemed to reduce the strength of the matrix, while the calcium
in the slag appeared to increase strength. Oh et al. (2010) postu-
lated that the calcium in the slag is available to form C–S–H(I)
while it is not available in the Class C fly ash ‘‘due to the different
chemical forms of calcium in these raw materials.’’ In summary,
the precise role of calcium in geopolymer formation remains
poorly defined, although it is clearly pivotal in determining the
engineering properties and durability of the product.
Iron is also likely to be important; Lloyd et al. (2009a) observed
that a high iron content appears to render precursor particles rel-
atively unreactive, and noted that phase segregation between
iron-rich and iron-poor glasses within particles can mean that
compositional information obtained even at an individual particle
level is not necessarily able to describe reactivity in geopolymer
formation. During alkali activation, iron does not appear to move
much from its original position within fly ash particles (Lloyd
et al., 2009b; Provis et al., 2009). However, the effect of iron on
the lability of precursor glassy phases and the role that iron may
play during geopolymer reactions remain largely undescribed;
92 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Mössbauer spectroscopy does appear to be giving some initial an-
swers in this area, at least for the case of very Fe-rich precursors
(Gomes et al., 2010).
The mechanistic distinction between silicate-activated and
hydroxide-activated geopolymer formation can be attributed to
some extent to the differences in the sites at which gel precipita-
tion takes place, with hydroxide-activated gels forming predomi-
nantly on fly ash particle surfaces rather than by polymerisation
in the bulk region (Lloyd et al., 2009b). In silicate-activated fly
ash, the gel consisted of roughly similar colloidal-sized, globular
units closely bonded together at their surfaces. The gel appeared
to be homogenous throughout the sample, whether at the surface
of the ash particles or relatively far away in the interstitial space. It
is, however, possible to influence this through nanoparticle seeding
of hydroxide-activated binders to manipulate the nucleation loca-
tion and enhance gel growth (Rees et al., 2008).
The reason why this geopolymer gel is formed instead of large
zeolitic crystallites is related to the degree of polymerisation of sil-
icate species, as well as the influence of temperature and water
content (Provis et al., 2005; Provis and Van Deventer, 2007,
2009; Fernández-Jiménez et al., 2008). White et al. (in press a) used
neutron pair distribution function (PDF) analysis to show that,
with increasing reaction time, the geopolymer gel derived from
metakaolin transitions to a more ordered state via an increase in
cross-linking. Following the initial nuclear magnetic resonance
(NMR) and thermal analysis data of Duxson et al. (2005a) and
Duxson et al. (2007c), a neutron PDF study of heated geopolymer
gels (White et al., 2010) confirmed that the water in metakaolin
geopolymers was present mainly as free water in large pores, with
only a small percentage (approximately <5%) of water either phys-
ically bound in small pores or chemically bound as hydroxyl
groups attached to the aluminosilicate framework structure.
Coarse-grained Monte Carlo simulations, based on interaction
energies derived from density functional modelling, showed that,
as the activator silica content of a metakaolin geopolymer system
is increased, more species (both aluminate and silicate) participate
in geopolymerisation, leading to a denser nanostructure and less
monomeric species existing in the pore solution (White et al., in
press b). The precipitates formed become larger with increasing sil-
ica content, which indicates that different structural transforma-
tion mechanisms occur depending on the type of activator used.
The results of these simulations have confirmed a previously pro-
posed hypothesis (Provis et al., 2005) regarding nucleation of pre-
cipitates in metakaolin-based systems, where in the presence of
silicate activating solutions the early release of aluminium from
metakaolin was suggested to lead to localised nucleation close to
the surface of the partially dissolved metakaolin particles; this
behaviour was reflected in the simulation results. On the other
hand, in hydroxide-activated systems there is no localised nucle-
ation taking place, and therefore the precipitates form throughout
the system. This differs from the nucleation behaviour observed in
fly ash systems, as outlined above; the aluminium release rates and
dissolution mechanisms differ significantly between the two
systems, and this leads to differences in the influence of dissolved
silica in the nucleation processes. The simulation results for
silicate-activated systems also provide direct evidence of Ostwald
ripening of geopolymer gel particles, which has never before been
explicitly shown to occur in geopolymer systems (White et al., in
press b).
These differences in particle dissolution mechanisms, and their
influence on geopolymer gel development, are essential in deter-
mining the processes which control the rate and location of geo-
polymer gel formation, and thus the microstructure and
performance of the final geopolymer binder. By using Attenuated
Total Reflectance Fourier Transform Infrared (ATR-FTIR) spectros-
copy, Rees et al. (2007a,b) observed that in a sodium hydroxide
activated Class F fly ash system selective leaching of Al from the
fly ash produces in the first instance a loose, Al-rich, ‘primary’
gel, as proposed earlier by Fernández-Jiménez and Palomo
(2005). The Al-deficient surface layer on the fly ash particles then
dissolves, followed by later stoichiometric release of Al and Si spe-
cies. During an induction period, the gel slowly comes to pseudo-
equilibrium with the surrounding solution via depolymerisation/
repolymerisation reactions (White et al., in press b). Gel nuclei
(particles that are sufficiently stable to resist depolymerisation)
begin to form, and the growth of a new gel phase begins. This
new gel is the phase predominantly responsible for strength devel-
opment and durability in geopolymers.
When nanoparticle seeds are added to the geopolymer system,
no induction period occurs, as the nanoparticles immediately cata-
lyse the formation of nuclei (Rees et al., 2008). In this case, when
the first Al-rich species are released into the solution from the fly
ash particles, their immediate addition to the nuclei forms an Al-
rich gel. Structural reorganisation of this gel will later lead to the
formation of zeolites, which can differ in crystal structure from
the zeolites which develop in unseeded systems (Rees et al.,
2008). Dissolution of the remnant siliceous layer on the fly ash par-
ticles releases Si-rich species, which also add rapidly to the grow-
ing nuclei, forming a high-silica gel region. Congruent ash
dissolution then releases Al and Si species which add to the grow-
ing nuclei in a similar way to the unseeded system, creating a bulk
geopolymer gel with similar composition and structure to that ob-
served in the absence of seeds. Hajimohammadi et al. (2010,
2011a,b) then followed this work by applying synchrotron radia-
tion-based FTIR microscopy to the spatially-resolved analysis of
both one-part and two-part geopolymer systems, showing that
the release rates of both Si and Al are critical in determining
strength development and microstructural evolution in growing
geopolymer gels. The enhancement of the nucleation process
through seeding led to the formation of an additional silica-rich
phase in the early stages of the reaction, which left more Al avail-
able to contribute to bulk gel formation and improved the early
strength development of geopolymer binders. The later release of
more silica from the Si-rich phase then also enhanced final
strength.
4. Role of particle technology in the optimisation of geopolymer
paste and concrete
Fig. 2 depicts a conceptual model for the interrelationship be-
tween geopolymer concrete mix design, the behaviour of wet con-
crete, and the performance of in-service concrete regarding
engineering properties and durability. In this section the principles
of particle technology will be used to analyse the behaviour of
fresh geopolymer cement paste and wet concrete, while in Section
5 recent progress on the interrelationship between binder micro-
structure and durability will be reviewed.
4.1. Particle shape effects in fresh pastes
The grinding of clinker and blast furnace slag into cementitious
powders results in non-spherical particle morphologies. It has been
hypothesised that the spherical particle shape of fly ash reduces vis-
cosity and yield stress of fresh paste when fly ash is added to OPC
(Laskar and Talukdar, 2008). Provis et al. (2010) used a packing
model to demonstrate the ‘ball-bearing’ effect of spherical ash par-
ticles in a paste, by the reduction of particle interlocking when the
paste is sheared. Palomo et al. (2005) observed that ash chemistry
and variability largely affected the rheology of fly ashes activated
by sodium hydroxide solutions. Various investigations have also
studied the effect of ash fineness on early-age properties of OPC
and geopolymer pastes. However, many of these investigations
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Fig. 2. Conceptual diagram showing interrelationships between mix design, gel chemistry, matrix characterisation, engineering properties and durability of geopolymer
concrete.
used milling to obtain different particle size distributions (Bou-
zoubaâ et al., 1997, 1999; Kumar et al., 2007), which adds the com-
plication of a change in particle shape as the glassy spherical ash
particles are shattered into fragments during the milling process.
Kumar et al. (2007) showed that slightly higher geopolymer
strengths were achievable by use of the finer particle size fractions
obtained by classification or attrition milling, when compared to
the use of raw fly ash, but that vibratory milling was able to give
an improvement in strength of as much as 50%. These effects could
not be attributed solely to particle size reduction, given that the
vibratory-milled fly ash was not the finest of the ash samples stud-
ied, and a mechanochemical activation process was postulated to
enhance reactivity. In contrast, Keyte (2008) did not show a signif-
icant effect of either classification or ring milling in the compres-
sive strength of geopolymers.
Clearly, the effect of particle geometry on the behaviour of fresh
geopolymer pastes is complex, so there is a need to deconvolute the
effects of particle shape, particle size and mechanochemical phe-
nomena and fly ash chemistry. Some detailed studies in this area
have used synthetic aluminosilicate glasses (Hos et al., 2002; Keyte,
2008). However, to fully simulate the geopolymerisation of waste
materials, it is also necessary to incorporate non-framework cations
into these synthetic glasses to better reflect the main reactive
phases participating in alkaline activation (Duxson and Provis,
2008). Some of the knowledge of mixed alkali/alkaline earth alumi-
nosilicate glasses developed recently through the study of synthetic
slags (Rajaokarivony-Andriambololona et al., 1990; Shimoda et al.,
2008) should also be relevant in future developments in this area.
93
4.2. Water–binder ratio and rheology of geopolymer pastes
The pore volume of calcium silicate-based (Portland cement),
aluminate-based (high-alumina cement) or sulphate-based (super-
sulphated or calcium sulphoaluminate cement) binders decreases
significantly as they hydrate over the first few days to weeks after
mixing, with this reduction in pore volume attributable to the
effective consumption of water as it participates in hydration reac-
tions. A water to binder (w/b) ratio of approximately 0.2, where the
exact value depends on the cement composition (Brouwers, 2004),
is converted during the hydration of OPC to ‘non-evaporable’ water
in hydration products, in addition to the considerable amount of
water bound in ‘gel pores’ (less than 2.7 nm diameter), which is
also not readily removed from the gel. Thus, hydrated OPC with a
nominal water/binder ratio of 0.5 has a final pore volume which
is markedly less than the volume of the water which was initially
added into the mix (Lothenbach et al., 2008).
In contrast, geopolymeric binders are primarily aluminosilicate-
based and do not form hydrate products, as discussed in Section
3.2. Hence, geopolymers do not have the same pore volume reduc-
tion benefit through the conversion of water into a solid via its
incorporation into reaction products. This distinction has signifi-
cant implications for the development of geopolymer concrete.
While in geopolymer binders the pore size distribution is much
finer, and is greatly refined throughout the hardening process
(Duxson et al., 2005b; Sindhunata et al., 2008; Lloyd et al.,
2009c), the absolute pore volume is not affected by the formation
of hydration products and there is no significant amount of water
94 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104

bound into the solidified geopolymer gel (Provis et al., submitted
for publication). Consequently, the minimisation of w/b ratio in
geopolymer concrete is a particularly significant operational factor
in ensuring quality low permeability concrete.
4.3. Particle packing and mix design in geopolymer concretes
Besides particle shape, particle size distribution also has a
marked effect on the mixing, workability and rheology of wet con-
crete. This is an active area of research in which many fundamental
questions remain unanswered (Hunger and Brouwers, 2009), and
the development of advanced analytical and simulation techniques
is continuing to provide important advances. Because this is based
around the analysis of the locations and interactions of particles on
every length scale from a few nanometres (the basic building units
in C–S–H or geopolymer gel) to more than a centimetre (coarse
aggregate), it is not a problem which has an easy or straightfor-
ward answer. It is very difficult to develop a unified discussion of
particle–fluid interactions across such a range of length scales, as
the effects, relative importance and interactions of parameters
such as surface tension, chemical admixtures and solution ionic
strength will differ across length scales (Provis et al., 2010). There-
fore, most studies consider in detail either the packing of larger
particles (particularly the ‘fine’ and ‘coarse’ aggregates; usually
sand and crushed rock) (Amirjanov and Sobolev, 2008; Goltermann
et al., 1997) or the packing of binder components (Wong and Kwan,
2008), with the combination of these two types of components
presenting a much more challenging set of experimental and the-
oretical problems (Jones et al., 2002, 2003).
The term ‘mix design’ in concrete is used mainly to describe the
process of proportioning of binder constituents, chemical and min-
eral admixtures, rock and sand, with the aim of optimising some
specified combination of technical (mechanical and durability)
properties, placement and finishing, and cost (Ji et al., 2006). In
geopolymer concrete mix design, it is important to tailor the con-
tent and composition of the alkali activator and aluminosilicate
precursor materials, as well as the aggregate components. This
does not necessarily mean that the change from Portland cement
to a geopolymer binder requires wholesale changes to overall
mix designs, as the optimisation of the aggregate blends and aggre-
gate-binder ratio is driven by the order of magnitude cost differen-
tials between cementitious materials and aggregate. However,
with a fundamental understanding of aluminosilicate chemistry,
significant improvements in geopolymer concrete quality and
economy can be achieved, particularly with regard to control of
curing conditions, activator dosage and efflorescence. Successful
geopolymer concrete mix designs have been published by various
researchers; some of these mix designs incorporate chemical
admixtures (particularly superplasticisers), although analysis of
the effectiveness of many of the admixtures which are commonly
used in Portland cement mixes under the highly alkaline condi-
tions of geopolymerisation has shown mixed outcomes to date
(Puertas et al., 2003; Criado et al., 2009).
5. Linking geopolymer binder structure and durability
5.1. Factors affecting the service life of reinforced concrete
Weather-induced mechanisms of attack on the binder and concrete
(including freeze–thaw damage and/or salt scaling), and alkali-
aggregate reactions involving some siliceous or carbonate aggre-
gates, can also be problematic under given circumstances and in
some climatic conditions, but the focus of the review presented
here will be on mechanisms of degradation which are explicitly re-
lated to ionic transport. The loss of alkalinity (via leaching, carbon-
ation or other mechanisms) and the ingress of chloride are the
primary causes of steel corrosion within concrete, and the resis-
tance of the material to degradation by these mechanisms will
obviously depend intrinsically on its mass transport properties. In
particular, the width, connectivity and degree of water saturation
of the pores and cracks which are present throughout the binder
will play a highly significant role in determining the resistance to
ingress by aggressive agents. Cracks and pores with relevance to
binder structure and durability performance are present on length
scales from nanometres to millimetres in most concrete structures,
making this another true multiscale problem, similar to the prob-
lems of mix design and particle packing as discussed above. The
application of both traditional and advanced techniques in the anal-
ysis of the microstructure of the binder therefore again becomes
imperative to understanding and designing for durability, particu-
larly when introducing a new binder system such as the alkali-
activated family of binders.
Fig. 3 shows a schematic illustration of some of the physico-
chemical factors which can determine the service life of a concrete;
the complexity of this system is immediately apparent from this
diagram – which by no means provides an exhaustive listing of
the mechanisms by which concrete can be physicochemically dam-
aged in-service. The discussion in the subsequent sections of this
paper will present a brief overview of selected aspects of the sce-
nario depicted in Fig. 3, with a specific focus on the design and
characterisation of durable alkali-activated (including geopolymer)
binders and concretes.
5.2. Microcracking phenomena
In broad terms, cracking of concretes can be caused by chemo-
mechanical (shrinkage or expansion of gel phases, autogenous
heating during curing) or physicomechanical (applied load,
freeze–thaw) processes. Concrete is well known to be strong in
compression but weak in flexion and tension; the use of steel
reinforcing, often in combination with techniques such as preten-
Environment Embedded steel
reinforcing
Mg2+(aq) Gel binder
Cl-(aq)
SO42-(aq)
CO2(g)
H+(aq)
OH-(aq)
Na+(aq)
Ca2+ & Na+ in gel
Aggregate

The fundamental basis of the durability of reinforced concrete is

the ability to develop and maintain a dense, impermeable binder
gel which creates and stabilises a highly alkaline environment, with
appropriate chemical (in particular electrochemical) conditions to
enable the stabilisation of embedded steel in a passive state. Steel Fig. 3. Schematic diagram of the factors influencing the service life of concrete
corrosion is, on a worldwide basis, the predominant cause of pre- element (based on either alkali-activated or OPC binders) with embedded steel
mature failure of reinforced concrete elements (Hobbs, 2001). reinforcing.
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
sioning, and careful structural design, will often be intended spe-
cifically to compensate for this weakness by ensuring that the con-
crete itself bears minimal tensile load. However, an advantage of
geopolymer binders is that they tend to demonstrate a flexural
(and presumably also tensile, although this is rarely tested di-
rectly) strength significantly higher than is specified by the stan-
dard relationships which apply for OPC concretes of similar
compressive strength (Sofi et al., 2007). Mortar compressive
strengths in excess of 100 MPa, and concrete strengths higher than
70 MPa, have been relatively widely reported for laboratory
samples. Achieving such strengths consistently in large-scale pro-
duction is more complex, but the large-scale production of ‘high-
performance’ concretes by alkaline activation of suitably chosen
precursor blends is certainly possible, and is beginning to be dem-
onstrated on a commercial scale.
It is also important to distinguish load-induced macroscopic
cracking of the concrete from microcracking of the binder; both
are likely to be deleterious in terms of durability performance,
but the length scales on which the cracks form are very different.
In an early investigation, Byfors et al. (1989) observed significant
microcracking in samples of ‘F-concrete’ (superplasticised NaOH/
Na2CO3-activated slag). Häkkinen (1992a,b) observed a tendency
towards microcracking in alkali-activated slag concretes, with a
corresponding increase in the rate of carbonation and reduction
in strength, and found that a high activator content tended to cor-
relate with a high extent of microcracking. Collins and Sanjayan
(2001) conducted a detailed investigation in this area and con-
cluded that drying effects during curing were a primary cause of
microcracking of alkali-activated slags.
More recently, Bernal et al. (2011) proposed, via capillary suc-
tion measurements, that the extent of microcracking in silicate-
activated slag-based concretes depends significantly on the paste
content of the concrete mix design; excessive binder content was
suggested to lead to heat generation during curing of concrete
specimens, leading to thermally induced microcracking of the con-
crete at early ages. However, in this case, microcracking was not
observed to have any significant influence on the rate of carbon-
ation of the concretes, indicating that the higher binder content
was able to provide sufficient densification of the matrix to com-
pensate for additional carbonation taking place via transport along
cracks. The fundamental cause of microcracking is the partially re-
strained chemical shrinkage of binder phases after hardening,
which has been proposed from a thermodynamic basis to be intrin-
sic to the chemistry of alkali-activated binders and related (pozzo-
lanic) systems (Justnes et al., 1998; Chen and Brouwers, 2007).
Almost all cement-like binders show either shrinkage or expansion
during or after hardening as a result of the process of crystallo-
graphic (including gel) phase evolution which results in strength
generation and development (Brouwers, 2004). Maintaining
dimensional stability is thus a key challenge across the field of ce-
ment and concrete technology.
The observed trends in microcracking intensity in alkali-
activated concretes as a function of paste content and curing
regime thus highlight the value of understanding interactions be-
tween the binder and aggregate, and effects related to heat gener-
ation and heat and moisture transport during curing, in mitigating
the effects of microcracking on concrete performance and durabil-
ity. The key interactions take place in the region known as the
‘interfacial transition zone’, and the microstructural characteristics
of this region of the concrete are critical in terms of both strength
and durability performance.
5.3. Interfacial transition zone effects
In the area where the binder comes into contact with aggregate
particles, there are often microstructural differences when com-
95
pared with ‘bulk’ binder regions, meaning that these specific re-
gions can exert a disproportionately high (and usually negative)
influence on the mass transport, tensile and flexural properties of
the geopolymer concrete. This is an area in which geopolymer
binders are believed to provide significant advantages over OPC.
The chemistry of the high-calcium OPC systems tends to lead to
the formation of a porous zone containing large, mechanically
weak crystals surrounding aggregate particles (Ollivier et al.,
1995; Mehta and Monteiro, 2006), and this is a key pathway for
both mechanical failure and mass transport in a concrete. The
interfacial transition zone in geopolymer concretes has been iden-
tified as being dense and much less microstructurally distinct from
the bulk of the binder region (Shi and Xie, 1998; Lee and Van
Deventer, 2004, 2007; Provis et al., 2007; Zhang et al., 2009), which
provides much higher tensile and flexural performance, as well as
removing the possibility for these regions to lead to the formation
of a percolated porous pathway for mass transport through the
binder, thus enhancing the durability of the geopolymer material.
Fig. 4 shows an example, presented in detail by Lee and Van
Deventer (2007) and reproduced in simplified form here, of the ele-
mental compositions observed across the interface between a (K,
Na) silicate-activated fly ash geopolymer binder and a quartz par-
ticle within a section of siltstone. The elemental compositions were
determined by scanning electron microscopy–energy dispersive
spectroscopy (SEM–EDS), and show a uniform binder region span-
ning the distance from the bulk binder to the aggregate particle
surface. Similar results were also obtained for a range of phases
present within a basalt specimen; an iron-rich augite region in
the basalt did generate a distinguishable region less than 2 lm in
thickness and rich in Fe and Si at the interface, but there was not
a distinct and chemically/crystallographically different region at
the interface in any case when comparing to the bulk binder struc-
ture. The SEM–EDS data in Fig. 4 obviously do not provide detailed
microstructural information; however, imaging of the same speci-
mens did not show any differences in microstructure or any appar-
ent additional porosity in these regions (Lee and Van Deventer,
2007), consistent with the general consensus in the literature
regarding alkali-activated binders.
5.4. Microporosity in the bulk of the geopolymer binder
There are a variety of techniques available for the analysis of
micron-scale pores within a geopolymer (or any cement-type
binder), which can be classified roughly into the categories of
one-dimensional, two-dimensional and three-dimensional analy-
sis. Among these categories, there is no universally applicable tech-
nique which can provide a full multiscale characterisation of a
complex material such as a geopolymer binder; a fuller toolkit of
techniques is required to obtain detail across the length scales of
interest.
Fig. 4. Elemental concentrations (from scanning electron microscopy–energy
dispersive spectroscopy); data from Lee and Van Deventer (2007).
96 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
One-dimensional analysis of porosity in the range of nanome-
tres to microns is usually conducted by gas sorption (most com-
monly the Barrett–Joyner–Halenda, BJH, technique), direct
penetration measurements (air, water and chloride penetration
being the most commonly applied), or mercury intrusion porosi-
metry (MIP). The limitations and inaccuracies of MIP in application
to cements are well known (Diamond, 2000), but it is widely used
due to its apparent simplicity and the ready availability of instru-
mentation. MIP largely fails in the case of samples with complex
pore geometries (the ‘ink-bottle’ effect), as the full volume of a
complex-shaped pore is registered as having an effective pore size
equivalent to that of the narrowest part of its entry. This results in
an under-estimation of mean pore size. The high pressures used to
intrude mercury into the smallest pores in a sample are also prob-
lematic, as the applied pressure will often exceed the strength
(compressive or tensile) of the material to which it is being applied,
which can result in significant crushing effects. Multi-cycle MIP
has been proposed as a method by which the ink-bottle effect
can be minimised or avoided (Kaufmann et al., 2009), and appears
to provide some significant advances in this area with good agree-
ment with N2 sorption data. Gas sorption is believed to provide
good characterisation of very small pores, but cannot accurately
measure the larger pores which are essential to mass transport
in samples with pores spanning a very wide range of length scales.
Direct water and air permeability measurements of geopolymer
and other alkali-activated binders have shown a range of perfor-
mance, depending mainly on the mix designs tested. Binders with
well-cured alkali-activated binders with low water/binder ratios
perform acceptably in these tests (Shi, 1996; Olivia et al., 2008;
Sagoe-Crentsil et al., 2010), but generally do not provide results
which could be considered particularly outstanding, most likely
due to the low levels of space-filling bound water associated with
these gels. On the other hand, chloride permeability testing of alka-
li-activated mortars and concretes has shown a wide range of perfor-
mance, with the outcomes depending to a significant extent on the
details of the testing methodology selected. The ASTM C1202 test
methodology involves the application of a current to the sample
and uses measurement of the charge passed as a proxy for the move-
ment of chloride ions into the sample under the electrical field gra-
dient. The outcomes of this test are strongly dependent on pore
solution chemistry (Liu and Beaudoin, 2000; Shi, 2004), and so it
sometimes registers alkali-activated binders as showing very good
resistance to mass transport (Shi, 1996; Roy et al., 2000; Bernal
et al., 2011), and sometimes as performing rather poorly. Direct
measurement of alkali cation movements from the pore solutions
of alkali-activated binders (Lloyd et al., 2010) provides insight into
the cause of this behaviour. The mobility of alkalis will certainly dis-
play a strong influence on the total charge passed by the samples
during the test. It has been proposed that alternative methodologies
which provide a more direct measurement of the progress of chlo-
ride migration into the binder – for example ponding tests, or the
NordTest Build 492 accelerated test, will provide a more valid com-
parison which is relatively independent of the pore solution chemis-
try of the binder, and work in this area is ongoing.
Two-dimensional analysis of pore structures is predominantly
conducted using microscopic methods, in particular scanning elec-
tron microscopy (SEM). The primary challenge associated with the
observation of porosity by SEM (designed to detect solid matter) is
that the aim of the measurement is to detect the regions which
contain no solid matter. This process of ‘trying to see the parts
which are not there’ provides challenges in pore identification. It
can be overcome to some extent for calcium-rich systems, such
as OPC or alkali-activated slags, by the use of image analysis algo-
rithms (Brough and Atkinson, 2000; Wong et al., 2006). However,
for aluminosilicate geopolymer systems with lower electron
density contrast between the solid and pore regions, the applica-
tion of Wood’s metal intrusion prior to the collection and analysis
of SEM images has proven to be of value (Lloyd et al., 2009c). In this
technique, the pores of the sample are filled with a high elemental-
number, low melting point alloy, which is intruded into the pore
space of the sample under moderate pressure (lower than the pres-
sures used in MIP, to prevent damage to delicate parts of the
microstructure) and at temperatures below 100 °C but still above
the melting point of the alloy. The sample is then cooled, and the
alloy solidifies within the pore network, to provide a high degree
of elemental contrast and make the pores visible against the low-
elemental number binder regions (Willis et al., 1998; Nemati,
2000). In the recent application of this technique to alkali-activated
binders (Lloyd et al., 2009c), it was calculated that pores as small as
11 nm – well below the spatial resolution achievable by standard
SEM techniques – are able to be filled with the molten alloy, while
using an intrusion pressure lower than the compressive strength of
the binder material to minimise microstructural damage.
The primary three-dimensional characterisation technique
which is able to be applied to the analysis of cements (alkali-acti-
vated and traditional) is X-ray microtomography. There have been
a number of studies of OPC-based materials by this technique over
the past decade or more (Gallucci et al., 2007). The systematic anal-
ysis of alkali-activated binders by X-ray microtomography and
hard X-ray nanotomography has recently been presented for the
first time (Provis et al., submitted for publication, 2011a).
In the analysis of the alkali-activated binder sample set dis-
cussed here, the samples were analysed using beamline 2-BM at
the Advanced Photon Source synchrotron at Argonne National Lab-
oratory, using 22.5 keV X-ray radiation and achieving a 750 nm
voxel resolution. Porosity was calculated during segmentation of
the sample into ‘pore’ and ‘solid’ regions using slightly modified
versions of freely available data processing scripts (Nakashima
and Kamiya, 2007) based on the full greyscale tomographic recon-
structions, and tortuosity was computed by random walker simu-
lations utilising the segmented pore network, again based on the
algorithms of Nakashima and Kamiya (2007). An example showing
the greyscale and segmented images for a given region of a fly ash-
Na silicate geopolymer binder is given in Fig. 5.
Fig. 5 also demonstrates some of the challenges associated with
microtomographic characterisation of fly ash geopolymer samples.
The large dark regions represent the interior of hollow or partially
hollow (cenosphere or plerosphere) fly ash particles; these regions
are not connected to the pore volume within the gel, and so will
not contribute to the transport properties of the material. However,
in measurements or calculations of total porosity, these regions
must be excluded via the application of algorithms which considers
‘connected porosity’ only. As an additional complicating factor, the
tomographic reconstruction process (if sample alignment was not
on a single vertical axis during rotation for scanning in the instru-
ment) can introduce streaking artefacts that effectively break
through the shell of a fly ash particle and connect its interior to the
binder pore network. Thus, the collection of high-resolution, high-
quality tomography data is of particular importance in enabling
the correct analysis of pore network transport parameters for sam-
ples containing fly ash. This consideration is less problematic for
samples based on slag (or Portland cement) because the precursor
particles do not contain significant inaccessible pore volumes.
The data collected by Provis et al. (submitted for publication) as
shown in Fig. 6, demonstrate that the porosity of the more cal-
cium-rich geopolymer binders (the systems containing 50% or more
slag) decreases as a function of curing duration, indicating that there
is some chemical binding of water taking place in this system, con-
sistent with the known formation and coexistence of sodium alumi-
nosilicate hydrate (‘N–A–S–(H)’) and calcium (alumino)silicate
hydrate (‘‘C–(A)–S–H’’) gels in mixed fly ash/slag geopolymer
binders.
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
(a)
200 µm
Fig. 5. Binary thresholding of a microtomographic image: (a) original greyscale image, (b) binary segmented image. Scalebar represents 200 lm.
Fig. 6. Relationship between segmented porosity and curing duration for a range of
sodium silicate-activated fly ash/slag systems; data selected from Provis et al.
(submitted for publication).
The results of drying tests conducted on a series of mortar sam-
ples synthesised from the same binders as were studied by tomog-
raphy (Ismail et al., 2011) show further that the C–(A)–S–H phase is
binding some water – but to a lesser extent than in the case of OPC
hydrate products. As the gel evolves, and its porosity decreases over
time, the tortuosity of the pore network also increases. The sample
set presented in Fig. 6 shows close agreement with an inverse rela-
tionship between connected porosity and diffusion tortuosity
(Provis et al., submitted for publication). The diffusion tortuosity
describes the effect on mass transport of the constriction due to
the presence of the pore network. An increase in this parameter
by almost a factor of two between early age (<7 days) and 45 days
of age indicates that the rate of diffusion through a well-cured bin-
der would be halved when compared to a poorly-cured binder. This
further highlights the importance of adequate curing in achieving
high performance and durability in geopolymer concretes. A criti-
cism of alkali-activation technology is that when mixes are
poorly-designed, thermal curing is often required for adequate
strength development. This is not the case for a well-designed
mix with sufficiently well-controlled activation conditions, as has
been demonstrated for systems based on fly ash, on slag, and on
mixtures of these two precursors. However, it is clear from the
tomography results that, regardless of the rate of early strength
97
(b)
development, an extended period of curing will provide marked
advantages in-service life and overall durability performance once
the binder is placed in-service and exposed to aggressive
environments.
The substitution of fly ash for slag is seen in Fig. 6 to lead to
higher porosity. This is consistent with the results of nitrogen sorp-
tion analysis of a range of similar binders (Lloyd et al., 2009c), and
the porosities obtained from tomography are also within 2 vol.% of
the values obtained in that study for samples of similar mix design
and curing duration. The slag-rich systems predominantly form a
C–(A)–S–H gel, which appears to bind a larger amount of water
into its nanostructure, while the fly ash-rich systems form N–A–
S–(H) gels with a low bound water content (Juenger et al., in press).
The discussion presented here therefore highlights the impor-
tance of understanding and controlling the porosity of geopolymer
binders, particularly in the case of fly ash-rich systems. If the high-
er porosity of the geopolymer-type (sodium aluminosilicate) gel
when compared with calcium silicate gels such as those produced
by OPC hydration were to lead to high diffusivity (and thus rapid
mass transport) of aggressive ions through the binder to reach
the embedded steel reinforcing, this would mean that the durabil-
ity of these materials would in all likelihood be unacceptably poor.
Evidence from the in-service performance of geopolymer and other
alkali-activated binders (Deja, 2002; Shi et al., 2006; Xu et al.,
2008), as well as from laboratory chloride penetration testing as
discussed above, shows that the observed performance is signifi-
cantly better than would be expected from raw permeability or
carbonation rate data (Shi et al., 2006; Provis et al., 2007; Rodrí-
guez et al., 2008; Adam, 2009; Bernal et al., 2011). Carbonation
of alkali-activated binders is an open and active area of research,
and much remains to be explained in this area, particularly with
regard to the relationship between carbonation and steel degrada-
tion, which may or may not be similar to the corresponding rela-
tionship in Portland cement-based concretes. This suggests that
there are additional effects which compound with, or mitigate,
the direct influence of porosity on permeability (particularly ionic
permeability, which also relates to gel chemistry-specific effects
and interactions) and durability.
6. Technical challenges
In developing the technology of novel cements and concretes, it
is often important to apply expertise obtained from the general
98 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
field of mineral processing, and in particular from the knowledge
of fluid-particle interactions which has resulted from intensive glo-
bal research efforts in the optimisation of hydrometallurgical pro-
cesses. For example, the phase chemistry of cement and other
mineral components (particularly fly ash and slag) is critical in
controlling reactivity. So, as was discussed in detail in the preced-
ing sections of this paper, the accurate understanding of the differ-
ent glass phases present in geopolymer precursor materials is
required to determine the optimal processing method and compo-
sition of binder components to be applied to manipulate concrete
properties. Classified fly ash and ground slag available through
the cement supply chain are not always able to be used straightfor-
wardly with existing commercially available alkaline activators to
prepare optimal geopolymer binders, so special preparation of
alkaline activators may also be warranted. In addition, fly ash
and slag (being waste materials) are subject to far more variability,
especially when they are obtained from a range of sources, than
OPC as a quality controlled product. Therefore, online monitoring
and optimisation of activator types and binder components is nec-
essary in the production of consistent, high-quality geopolymer
concretes.
As was mentioned briefly above, it is important to control the
rheology of fresh concrete to enable it to be placed and finished,
and the ability to do this without adversely affecting the final prop-
erties of the hardened concrete depends primarily on the manipu-
lation of colloidal interactions by the use of chemical admixtures
such as superplasticisers (Puertas et al., 2003; Criado et al.,
2009). The existing range of commercially available superplasticis-
ers have been developed specifically to suit the complex series of
chemical reactions which take place in the OPC system, and are
usually not effective in the geopolymer system. Moreover, most
of the various admixtures used to control slump, air dispersion,
water retention, and other properties of the OPC system are less
effective in the geopolymer system. Consequently, there is a need
to develop a whole set of new admixtures for the geopolymer sys-
tem, which presents a significant challenge for an emerging indus-
try with a lack of scale, but which is still required to compete with
the well-established OPC industry.
Alkali activation also provides the potential for the utilisation of
non-blast furnace slags; alternative materials such as ferronickel,
steel and phosphorus slags have also been alkali-activated to form
usable binders and concretes, some on a laboratory scale and some
in larger-scale applications in China and the former Soviet Union
(Shi et al., 2006). The leaching of toxic metal components from
some of these slags during activation may prove to be a cause for
some concern, but – as is the case for fly ashes – the selection of
appropriate waste materials for use as geopolymer precursors is
both important and possible.
At present, geopolymer concrete is most commonly produced in
a ‘two-part’ mix format, by blending coal fly ash, slag and alkali
activators together in a concrete batching plant. This requires a
high level of skill from the operators of the batching plant, which
is rarely available in the premix concrete industry. Ash and slag
are waste materials of variable composition, so that the concrete
mix designs and the added alkalis need to be varied to compensate
for these variations in the ash and slag. Such a technology, where
the quality control is mainly at the batching plant level, is not scal-
able on an industry-wide level and has limited appeal in the mar-
ket, despite the growing market pull for ‘green’ construction
materials.
In contrast with this situation, Zeobond has developed a process
whereby the various solid materials and proprietary activators are
processed together to produce a dry cement binder that behaves in
a similar way to OPC (Van Deventer et al., 2010b). The quality con-
trol is hence centralised in the cement binder plant, so that the dry
powder can be distributed to various concrete batching plants, as is
the case with OPC. To a large extent this addresses the supply chain
challenges and difficulties in price competition which are inherent
when sourcing materials from OPC-related suppliers. However, the
process of establishing such dry binder processing facilities is cap-
ital intensive and requires significant market drivers for geopoly-
mer concretes (synthesised via the existing process) to justify
investment.
7. Reduction in carbon emissions
CO2 emissions from cement production are incurred through
the consumption of fossil fuels, the use of electricity and the chem-
ical decomposition of limestone during clinkerisation, which can
take place at around 1400 °C. The decarbonation of limestone to
give the calcium required to form silicates and aluminates in clin-
ker releases roughly 0.53 tonnes CO2 per tonne of clinker (Damtoft
et al., 2008). In 2005, cement production (total cementitious sales
including OPC and OPC blends) had an average emission intensity
of 0.89 with a range of 0.65–0.92 tonnes CO2 per tonne of cement
binder (International Energy Agency, 2007). Therefore, the decar-
bonation of limestone contributes about 60% of the carbon emis-
sions of OPC, with the remaining 40% attributed to energy
consumption, most of which is related to clinker kiln operations;
the WWF-Lafarge Conservation Partnership (2008) estimated that
the production of clinker is responsible for over 90% of total ce-
ment production emissions.
In view of the fact that the requirement for decarbonation of
limestone presents a lower limit on CO2 emissions in clinker pro-
duction, and that there exist technical issues associated with the
addition of supplementary cementitious material (SCM, including
fly ash and ground granulated blast furnace slag), which restrict
the viability of direct Portland cement supplementation by SCM
above certain limits, the possibility to reduce CO2 emissions using
OPC chemistry is limited. The WWF-Lafarge Conservation Partner-
ship (2008) expects that the emissions intensity of cement, includ-
ing SCM, could be reduced to 0.70 tonnes CO2 per tonne of cement
by 2030 – which still amounts to around 2 billion tonnes of CO2 per
annum worldwide, even if cement production does not increase
from its current level.
Fig. 7 shows the CO2 emissions of various binder designs as a
function of OPC content. There have been a limited number of life
cycle analyses of alkali activation technology. One reasonably
extensive research program carried out in Germany (Weil et al.,
2009) has provided information regarding the selection of precur-
sors and mix designs for a range of geopolymer-based materials.
However, geographic specificity plays a significant role in a full life
Business as usual
Best Practice 2011
Stretch/Aspirational
E-CreteTM since 2008
Fig. 7. CO2 emissions of various cement binders as a function of OPC content.
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
cycle analysis, so there is the need for further studies considering
different locations in addition to a wider range of mix designs
spanning the broader spectrum of alkali activated materials. The
most carbon-intensive and also the most expensive ingredient in
geopolymer cement is the alkali activator, which should be mini-
mised in mix design. Two more recent studies (McLellan et al.,
2011; Habert et al., 2011) have also provided additional insight
in this area, and further highlighted the importance of the activator
in determining the environmental impact of geopolymer concretes.
A commercial life cycle analysis was conducted by the NetBal-
ance Foundation, Australia, on Zeobond’s E-Crete™ geopolymer
cement, as reported in the ‘Factor Five’ report released recently
by the Club of Rome (Von Weizsäcker et al., 2009). This life cycle
analysis compared the geopolymer binder to the standard OPC
blends available in Australia in 2007 on the basis of both binder-
to-binder comparison and concrete-to-concrete comparison. The
binder-to-binder comparison showed an 80% reduction in CO2
emissions, whereas the comparison on a concrete-to-concrete ba-
sis showed slightly greater than 60% savings, as the energy cost
of aggregate production and transport was identical for the two
materials. However, this study was again specific to a single loca-
tion and a specific product, and it will be necessary to conduct fur-
ther analyses of new products as they reach development and
marketing stages internationally. Fig. 7 shows a comparison of
the CO2 emissions of four different E-Crete™ products against the
‘Business as usual’, ‘Best practice 2011’, and a ‘Stretch/Aspirational’
target for OPC blends. It is noted that in some parts of the world
(particularly Europe), some of the blends shown here in the
‘Stretch/Aspirational’ category are in relatively common use for
specific applications, particularly CEM III-type OPC/slag blends,
but this is neither achievable on a routine scale worldwide at pres-
ent, nor across the full range of applications in which Portland ce-
ment concretes are used in large volumes.
8. Standards framework
Fig. 8 shows that the strategic development of standards is piv-
otal to the commercialisation of geopolymer cement. The regula-
tory framework governing concrete to be utilised in various
applications relies on a typical cascade of standards, with applica-
tion standards referring to concrete standards, and concrete
standards referring to standards covering cement and other raw
materials. Hence, when considering the regulatory framework for
a concrete binder system such as geopolymer cement or concrete,
most of the attention to regulatory aspects should be focused on
the cement standards, although some aspects of concrete stan-
dards also need to be considered. In general, all of the world’s con-
crete application, concrete and cement standards are based on two
‘super’ standards, i.e., European Union EN 197 and United States
ASTM C150/C595/C1157. For instance, Chinese cement and con-
crete standards are based largely on European Union Standards,
while Australian standards are based mainly on the American
standards.
These standards have been developed over many years, and in
collaboration with input from OPC manufacturing companies, with
the chemistry and behaviour of OPC-based concretes intrinsically in
mind. However, prescriptive standards containing constraints such
as ‘minimum cement content’ are increasingly being viewed as
excessively prohibitive, even for OPC-based systems. Products such
as geopolymer concrete may not simply be an evolution of existing
OPC technology, but instead may require an entirely different chem-
ical paradigm to understand their behaviour, and may perform en-
tirely acceptably, but without conforming exactly to the established
regulatory standards – particularly with regard to rheology and
chemical composition (Hooton, 2008). This is a significant hin-
99
drance to the acceptance of geopolymer technology. However, by
working with all stakeholders (Fig. 8), these barriers can be over-
come, provided that the intent of regulatory standards is met.
VicRoads, the state roads authority in Victoria, Australia, re-
cently recognised geopolymer concrete as being equivalent to
OPC for non-structural applications in a 2010 update to their de-
sign specification Section 703 (VicRoads, 2010); this specification
requires acceptable binder content, strength development, water
content and quality, aggregate properties, consistency, mixing
and placement methodologies, curing, finishing, and material sup-
plier competency guarantees. Zeobond is working with VicRoads to
also recognise geopolymer concrete for structural applications in
Section 620. Zeobond’s E-Crete is already used in VicRoads struc-
tural projects, and by local councils and housing developers in
sub-divisional works and slabs, applications which represent
approximately 70% of all concrete usage. These large scale applica-
tions are pivotal to the process of gradually convincing standards
authorities to accept geopolymer concrete.
9. Testing for durability
The question of whether geopolymer concretes are durable re-
mains the major obstacle to recognition in standards for structural
concrete, and hence to their commercial adoption. A material such
as geopolymer, which has been subjected to detailed investigation
only recently, cannot possibly have the availability of decades of
in-service testing and durability data to prove its long-term stabil-
ity. Most standard methods of testing cement and concrete dura-
bility involve exposing small samples to very extreme conditions
– in particular highly concentrated acid or salt solutions, with or
without the application of electrical field gradients – for short peri-
ods of time. The data obtained from these tests are then used to
predict how the material will perform under normal environmen-
tal conditions over a period of decades or more. In some of these
predictive models, engineering concepts including mass transport
through porous media, reaction kinetics and particle packing are
used, although usually in a simplified and semi-empirical form to
enable utilisation of the derived equations by non-specialists in
the field. However, the key shortcoming of this approach to ‘prov-
ing’ durability is that it can only provide indications of the ex-
pected performance, rather than definitive proof. Therefore, there
has been a very slow process of adoption of new materials, as it
is considered necessary to wait up to 20–30 years for ‘real-world’
verification. Adoption of fly ash and slag in OPC concretes is the
prime example of this, where the use of these SCMs was resisted
for decades. It is the authors’ experience that asset owners and
their insurance companies are willing to use geopolymer concrete
in low risk applications based on accelerated durability testing.
Higher-risk applications such as high rise buildings, which consti-
tute a smaller fraction of the total concrete market, will follow only
when the market is comfortable with the real-world track record of
the material in low risk applications. Therefore, a staged approach
towards the development of standards and commercial adoption
needs to be followed, as outlined in Fig. 8.
Zeobond also works closely with various roads authorities in
monitoring the performance of in situ geopolymer concrete struc-
tures and also with researchers in developing appropriate test
methods for durability (Fig. 8). The work conducted by RILEM Tech-
nical Committee 224 on Alkali Activated Materials (AAM) provides
essential advice to standards authorities about the structure of per-
formance based standards for AAM and the associated testing
methods for durability (http://www.rilem.net/tcDetails.php?tc=
224-AAM).
As outlined by Provis et al. (2011b), there have been a number
of studies of chloride diffusion in alkali-activated slag concretes,
100 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Fig. 8. Flow diagram for commercialisation of geopolymer cement and concrete.
with the performance of these materials in accelerated chloride
penetration tests generally observed to be at least comparable to
that of OPC. Zeobond’s E-Crete has been shown repeatedly to have
significantly lower chloride diffusion and acceptable freeze–thaw
performance compared with OPC concrete. As outlined in Section
5.4, the excellent performance of geopolymer concrete is related
to the highly refined pore network forming a dense low-calcium
C–A–S–H phase (Provis et al., 2011b). It is envisaged that advanced
techniques such as synchrotron-based nanotomography will ulti-
mately be used to compute transport properties of the geopolymer
gel generated from various mix designs (such as water/binder ra-
tio), and hence changes in durability performance.
10. Supply chain risks
The presence of a competitive and efficient supply chain is
essential for the successful scale-up of geopolymer technology.
For instance, granulated blast furnace slag (GBFS) is produced dur-
ing the production of pig-iron. Although there is a risk that produc-
tion methods for pig-iron will change and reduce the availability of
GBFS, this is not currently viewed as a substantial risk. As demand
for GBFS increases, blast furnace operators will increasingly invest
in rapid cooling slag handling equipment in order to increase the
production of GBFS. The further development and industrial adop-
tion of the new CSIRO air-cooled slag granulation method (Jahan-
shahi and Xie, 2010) is important in regions with a shortage of
clean water, which is the standard chilling medium used at pres-
ent. The main risk facing GBFS supply for geopolymer use is pref-
erential utilisation of GBFS in OPC blends rather than in
geopolymers.
There is a risk that a substantial reduction in world-wide carbon
emissions will result in reduced coal fired energy production, and
hence reduced availability of coal ash. While there is a global
determination to reduce energy dependency on thermal coal, it is
more likely that coal combustion will continue to increase in the
coming decades, but with some level of carbon capture through
sequestration or mineralisation. In any event, only a very small
fraction of coal ash is currently used in concrete. A more substan-
tial risk is the well-intended control of coal ash as a hazardous
material by, for example the US Environment Protection Agency,
which leads to uncertainty in investment in the supply chain of
all SCMs.
In economically developed and structurally segmented markets
for cementitious materials, distribution networks for slag and fly
ash are established, making it uncompetitive or practically impos-
sible to invest in new channels to market. In some cases, even volu-
minous by-product streams such as slag and fly ash are largely
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Fig. 9. E-Crete™ in Port Melbourne, Australia: (a) 25 MPa footpath along Westgate Freeway extension. (b) 55 MPa precast panels across Salmon Street bridge.
utilised or earmarked via long-term option contracts. Due to the
rapid increase in demand and production in markets such as China
and India, new opportunities for both OPC and geopolymer supply
channels are foreseeable. However, in established markets such as
the UK, where fly ash production does not exceed its consumption
in cement, the existing market for OPC utilises all of the available
fly ash. Despite the increasing number of research papers on geo-
polymers, there has been little development of the supply chain
channels necessary for scale production, which will continue to
limit the wider adoption of geopolymer technology until this bot-
tleneck is resolved.
11. The path of commercialisation
Fig. 9a shows a small section of E-Crete™ paving completed as
part of the Victorian Government upgrade of the Westgate Freeway
in Port Melbourne. While small in volume, this project required
that the concrete meet all of the technical specifications of the local
road authority, VicRoads, and the ultimate asset owner, the Port
Melbourne City Council, and was approved for use by a construc-
tion consortium which included numerous national and multi-
national construction companies and engineering firms. This small
project demonstrates the entire process of commercialisation of
geopolymer concrete (Fig. 8).
Fig. 9b shows the installation of 55 MPa E-Crete precast panels
for VicRoads. This concrete was required to meet the structural con-
crete code (VicRoads Section 610), which is far more stringent than
the requirements faced by the non-structural grade concrete in
Fig. 9a. Here the level of concrete specification and scrutiny was
consistent with bridge design, with the aim of using the highest
grade concrete specified by VicRoads as a ‘stretch target’ for trial
purposes.
As opposed to focusing on the technical detail of these projects,
attention is drawn to the fact that these case studies demonstrate
something that is rarely seen in geopolymer technology, i.e., the
vast regulatory, asset management, liability and industry stake-
holder engagement process that has been undertaken by Zeobond
to commercialise its E-Crete™ (Fig. 8). The shift from the labora-
tory to the real-world is not a process of scale-up of technology
alone. While such technical challenges are sufficient to be insur-
mountable to many, it is the ability to manage the scale-up of
industry participation and acceptance of geopolymer concrete that
provides the core challenge to the future world of a geopolymer ce-
ment/concrete industry.
12. Conclusions
Significant progress has been made in developing an under-
standing of the colloid and interface science, gel chemistry, phase
101
formation, reaction kinetics and transport phenomena underlying
geopolymerisation of aluminosilicates, including by the applica-
tion of advanced experimental and computational techniques.
However, despite much research, the role of calcium in geopolymer
gel processes and phase formation remains poorly understood.
Analysis of the nanostructure of geopolymer gels has enabled the
tailored selection of geopolymer precursors and the design of alkali
activator composition; such studies have established the relation-
ship between geopolymer gel microstructure and durability. Geo-
polymer concrete has performed very well in accelerated
durability tests, which is supported by its tortuous pore structure
and the specific details of its gel chemistry, although detailed ana-
lytical work in this area is still ongoing. In the absence of a long in-
service track record, chemical research is essential to validate
durability testing methodology and improve geopolymer cement
technology. Although the principles of particle technology have
been applied to the mix design and rheology of fresh geopolymer
concrete, this area is in its infancy and could benefit from the
knowledge base of mineral processing engineers. For example, no
suitable superplasticiser exists which can reliably enhance the rhe-
ology and reduce the required mix water content of fresh geopoly-
mer concrete.
Demand pull applied by a carbon conscious market continues to
be the key driver for the short term adoption of geopolymer ce-
ment. While the scale-up from the laboratory to the real-world is
technically challenging, the core challenge is the scale-up of indus-
try participation and acceptance of geopolymer cement. High pro-
file demonstration projects in Australia have highlighted the vast
regulatory, asset management, liability and industry stakeholder
engagement process required to commercialise geopolymer
cement.
The fact that all premixed and precast concrete standards are
based on an assumption of the use of OPC remains a major obstacle
to the commercial adoption of geopolymer cement. Even when as-
set owners and specifiers such as government, architects and de-
sign engineers accept the results of durability testing of
geopolymer concrete, the main barrier to entry of geopolymers
into an established market is access to a suitable supply of source
materials including fly ash, granulated blast furnace slag and alka-
line activators. Significant development of supply chains to accom-
modate geopolymer cement, predominantly in developing
markets, is essential in overcoming this barrier. This crucial aspect
of commercialisation is seldom appreciated by the research com-
munity, governments or the ultimate users of concrete. It is impor-
tant for commercial geopolymer concrete producers to work
closely with research partners to develop testing methods for
accelerated durability, especially as longer term in-service testing
data become available. Substantial progress has been made in Aus-
tralia, where the local road authority has recognised geopolymer
concrete for non-structural applications.
102 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Acknowledgements
This work has been funded by Zeobond Pty. Ltd. and the Austra-
lian Research Council (ARC) through an ARC Linkage Project grant
awarded jointly to JLP and PD. Authors JSJvD and PD hold financial
interest in Zeobond Pty. Ltd., a producer of geopolymer cements
and concretes.
References
Adam, A.A., 2009. Strength and Durability Properties of Alkali Activated Slag and Fly
Ash-based Geopolymer Concrete. Ph. D. Thesis, RMIT University, Melbourne,
Australia.
Amirjanov, A., Sobolev, K., 2008. Optimization of a computer simulation model for
packing of concrete aggregates. Particulate Science and Technology 26 (4), 380–
395.
Bernal, S.A., Mejía de Gutierrez, R., Pedraza, A.L., Provis, J.L., Rodríguez, E.D.,
Delvasto, S., 2011. Effect of binder content on the performance of alkali-
activated slag concretes. Cement and Concrete Research 41 (1), 1–8.
Bouzoubaâ, N., Zhang, M.H., Bilodeau, A., Malhotra, V.M., 1997. The effect of
grinding on the physical properties of fly ashes and a Portland cement clinker.
Cement and Concrete Research 27 (12), 1861–1874.
Bouzoubaâ, N., Zhang, M.H., Malhotra, V.M., Golden, D.M., 1999. Blended fly ash
cements – a review. ACI Materials Journal 96 (6), 641–650.
Brough, A.R., Atkinson, A., 2000. Automated identification of the aggregate-paste
interfacial transition zone in mortars of silica sand with Portland or alkali-
activated slag cement paste. Cement and Concrete Research 30 (6), 849–854.
Brouwers, H., 2004. The work of powers and Brownyard revisited: Part 1. Cement
and Concrete Research 34 (9), 1697–1716.
Buchwald, A., Wiercx, J., 2010. ASCEM cement technology – alkali-activated cement
based on synthetic slag made from fly ash. In: Shi, C., Shen, X. (Eds.), Advances
in Chemically Activated Materials CAM’2010. RILEM Publications, Jinan, China,
pp. 15–21.
Buchwald, A., Hilbig, H., Kapps, C., 2007. Alkali-activated metakaolin-slag blends –
performance and structure in dependence on their composition. Journal of
Materials Science 42 (9), 3024–3032.
Byfors, K., Klingstedt, G., Lehtonen, H.P., Pyy, H., Romben, L., 1989. Durability of
concrete made with alkali-activated slag. In: Malhotra, V.M. (Ed.), 3rd
International Conference on Fly Ash, Silica Fume, Slag and Natural Pozzolans
in Concrete. ACI SP114, pp. 1429–1444.
Chen, W., Brouwers, H., 2007. The hydration of slag, part 1: reaction models for
alkali-activated slag. Journal of Materials Science 42 (2), 428–443.
Collins, F., Sanjayan, J.G., 2001. Microcracking and strength development of alkali
activated slag concrete. Cement and Concrete Composites 23 (4–5), 345–352.
Criado, M., Palomo, A., Fernández-Jiménez, A., Banfill, P.F.G., 2009. Alkali activated
fly ash: effect of admixtures on paste rheology. Rheologica Acta 48 (4), 447–455.
Damtoft, J.S., Lukasik, J., Herfort, D., Sorrentino, D., Gartner, E., 2008. Sustainable
development and climate change initiatives. Cement and Concrete Research 38
(2), 115–127.
Davidovits, J., 2009. Geopolymer Chemistry & Applications. second ed. Institut
Géopolymère, Saint-Quentin, France, 590pp.
Deja, J., 2002. Carbonation aspects of alkali activated slag mortars and concretes.
Silicates Industries 67 (1), 37–42.
Diamond, S., 2000. Mercury porosimetry. An inappropriate method for the
measurement of pore size distributions in cement-based materials. Cement
and Concrete Research 30 (10), 1517–1525.
Duxson, P., Provis, J.L., 2008. Designing precursors for geopolymer cements. Journal
of the American Ceramic Society 91 (12), 3864–3869.
Duxson, P., Lukey, G.C., Separovic, F., Van Deventer, J.S.J., 2005a. The effect of alkali
cations on aluminum incorporation in geopolymeric gels. Industrial &
Engineering Chemistry Research 44 (4), 832–839.
Duxson, P., Provis, J.L., Lukey, G.C., Mallicoat, S.W., Kriven, W.M., Van Deventer, J.S.J.,
2005b. Understanding the relationship between geopolymer composition,
microstructure and mechanical properties. Colloids and Surfaces A 269 (1–3),
47–58.
Duxson, P., Provis, J.L., Lukey, G.C., Van Deventer, J.S.J., 2007a. The role of inorganic
polymer technology in the development of ‘Green concrete’. Cement and
Concrete Research 37 (12), 1590–1597.
Duxson, P., Fernández-Jiménez, A., Provis, J.L., Lukey, G.C., Palomo, A., Van Deventer,
J.S.J., 2007b. Geopolymer technology: the current state of the art. Journal of
Materials Science 42 (9), 2917–2933.
Duxson, P., Lukey, G.C., Van Deventer, J.S.J., 2007c. Physical evolution of Na-
geopolymer derived from metakaolin up to 1000 °C. Journal of Materials Science
42 (9), 3044–3054.
Fernández-Jiménez, A., Palomo, A., 2005. Mid-infrared spectroscopic studies of
alkali-activated fly ash structure. Microporous and Mesoporous Materials 86
(1–3), 207–214.
Fernández-Jiménez, A., Palomo, A., Sobrados, I., Sanz, J., 2006. The role played by the
reactive alumina content in the alkaline activation of fly ashes. Microporous and
Mesoporous Materials 91 (1–3), 111–119.
Fernández-Jiménez, A., Monzó, M., Vicent, M., Barba, A., Palomo, A., 2008. Alkaline
activation of metakaolin–fly ash mixtures: obtain of zeoceramics and
zeocements. Microporous and Mesoporous Materials 108 (1–3), 41–49.
Freedonia Group, 2009. World Cement to 2012. <http://www.freedoniagroup.com/
World-Cement.html>.
Gallucci, E., Scrivener, K., Groso, A., Stampanoni, M., Margaritondo, G., 2007. 3D
experimental investigation of the microstructure of cement pastes using
synchrotron X-ray microtomography (lCT). Cement and Concrete Research 37
(3), 360–368.
García-Lodeiro, I., Palomo, A., Fernández-Jiménez, A., Macphee, D.E., 2011.
Compatibility studies between N–A–S–H and C–A–S–H gels. Study in the
ternary diagram Na2O–CaO–Al2O3–SiO2–H2O. Cement and Concrete Research
41 (9), 923–931.
Gartner, E.M., Macphee, D.E., 2011. A physico-chemical basis for novel cementitious
binders. Cement and Concrete Research 41 (7), 736–749.
Goltermann, P., Johansen, V., Palbol, L., 1997. Packing of aggregates: an alternative tool
to determine the optimal aggregate mix. ACI Materials Journal 94 (5), 435–443.
Gomes, K.C., Lima, G.S.T., Torres, S.M., De Barros, S., Vasconcelos, I.F., Barbosa, N.P.,
2010. Iron distribution in geopolymer with ferromagnetic rich precursor.
Materials Science Forum 643, 131–138.
Habert, G., d’Espinose de Lacaillerie, J.B., Roussel, N., 2011. An environmental
evaluation of geopolymer based concrete production: reviewing current
research trends. Journal of Cleaner Production 19 (11), 1229–1238.
Hajimohammadi, A., Provis, J.L., Van Deventer, J.S.J., 2010. The effect of alumina
release rate on the mechanism of geopolymer gel formation. Chemistry of
Materials 22 (18), 5199–5208.
Hajimohammadi, A., Provis, J.L., Van Deventer, J.S.J., 2011a. The effect of silica
availability on the mechanism of geopolymerisation. Cement and Concrete
Research 41 (3), 210–216.
Hajimohammadi, A., Provis, J.L., Van Deventer, J.S.J., 2011b. Time-resolved and
spatially-resolved infrared spectroscopic observation of seeded nucleation
controlling geopolymer gel formation. Journal of Colloid and Interface Science
357 (2), 384–392.
Häkkinen, T., 1992a. The permeability of high strength blast furnace slag concrete.
Nordic Concrete Research 11 (1), 55–66.
Häkkinen, T., 1992b. The microstructure of high strength blast furnace slag
concrete. Nordic Concrete Research 11 (1), 67–82.
Hobbs, D.W., 2001. Concrete deterioration: causes, diagnosis, and minimising risk.
International Materials Reviews 46 (3), 117–144.
Hooton, R.D., 2008. Bridging the gap between research and standards. Cement and
Concrete Research 38 (2), 247–258.
Hos, J.P., McCormick, P.G., Byrne, L.T., 2002. Investigation of a synthetic
aluminosilicate inorganic polymer. Journal of Materials Science 37 (11),
2311–2316.
Hunger, M., Brouwers, H.J.H., 2009. Flow analysis of water–powder mixtures:
application to specific surface area and shape factor. Cement and Concrete
Composites 31 (1), 39–59.
International Energy Agency, 2007. Tracking Industrial Energy Efficiency and CO2
Emissions – Executive Summary. <http://www.iea.org/Textbase/npsum/
tracking2007SUM.pdf>.
Ismail, I., Provis, J.L., Van Deventer, J.S.J., Hamdan, S., 2011. The effect of water
content on compressive strength of geopolymer mortars. In: CD Proceedings of
7th AES-ATEMA 2011 International Conference on Advances and Trends in
Engineering Materials and their Applications, Milan, Italy, 4–8 July 2011.
Jahanshahi, S., Xie, D., 2010. A new integrated dry slag granulation and heat
recovery process. In: McCaffrey, R., Edwards, P. (Eds.), CD Proceedings of 6th
Global Slag Conference. PRo Publications Int. Ltd., Sydney, Australia.
Ji, T., Lin, T.W., Lin, X.J., 2006. A concrete mix proportion design algorithm based on
artificial neural networks. Cement and Concrete Research 36 (7), 1399–1408.
Jones, M.R., Zheng, L., Newlands, M.D., 2002. Comparison of particle packing models
for proportioning concrete constituents for minimum voids ratio. Materials and
Structures 35 (5), 301–309.
Jones, M.R., Zheng, L., Newlands, M.D., 2003. Estimation of the filler content
required to minimise voids ratio in concrete. Magazine of Concrete Research 55
(2), 193–202.
Juenger, M.C.G., Winnefeld, F., Provis, J.L., Ideker, J., in press. Advances in alternative
cementitious binders. Cement and Concrete Research. doi: 10.1016/
j.cemconres.2010.11.012.
Justnes, H., Ardoullie, B., Hendrix, E., Sellevold, E.J., Van Gemert, D., 1998. The
chemical shrinkage of pozzolanic reaction products. In: Malhotra, V.M. (Ed.),
6th CANMET Conference on Fly Ash, Silica Fume, Slag, and Natural Pozzolans in
Concrete. Bangkok, Thailand, pp. 191–205.
Kaufmann, J., Loser, R., Leemann, A., 2009. Analysis of cement-bonded materials by
multi-cycle mercury intrusion and nitrogen sorption. Journal of Colloid and
Interface Science 336 (2), 730–737.
Keyte, L.M., 2008. What’s Wrong With Tarong? The Importance of Coal Fly Ash Glass
Chemistry in Inorganic Polymer Synthesis, Ph.D. Thesis, University of
Melbourne, Australia.
Koloušek, D., Brus, J., Urbanova, M., Andertova, J., Hulinsky, V., Vorel, J., 2007.
Preparation, structure and hydrothermal stability of alternative (sodium
silicate-free) geopolymers. Journal of Materials Science 42 (22), 9267–9275.
Komnitsas, K., Zaharaki, D., 2007. Geopolymerisation: a review and prospects for
the minerals industry. Minerals Engineering 20 (14), 1261–1277.
Kumar, S., Kumar, R., Alex, T.C., Bandopadhyay, A., Mehrotra, S.P., 2007. Influence of
reactivity of fly ash on geopolymerisation. Advances in Applied Ceramics 106
(3), 120–127.
Laskar, A.I., Talukdar, S., 2008. Rheological behavior of high performance concrete
with mineral admixtures and their blending. Construction and Building
Materials 22 (12), 2345–2354.
 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
Lee, W.K.W., Van Deventer, J.S.J., 2004. The interface between natural siliceous
aggregates and geopolymers. Cement and Concrete Research 34 (2), 195–206.
Lee, W.K.W., Van Deventer, J.S.J., 2007. Chemical interactions between siliceous
aggregates and low-Ca alkali-activated cements. Cement and Concrete Research
37 (6), 844–855.
Liu, Z., Beaudoin, J.J., 2000. The permeability of cement systems to chloride ingress
and related test methods. Cement, Concrete and Aggregates 22 (1), 16–23.
Lloyd, R.R., Provis, J.L., Van Deventer, J.S.J., 2009a. Microscopy and microanalysis of
inorganic polymer cements. 1: remnant fly ash particles. Journal of Materials
Science 44 (2), 608–619.
Lloyd, R.R., Provis, J.L., Van Deventer, J.S.J., 2009b. Microscopy and microanalysis of
inorganic polymer cements. 2: the gel binder. Journal of Materials Science 44
(2), 620–631.
Lloyd, R.R., Provis, J.L., Smeaton, K.J., Van Deventer, J.S.J., 2009c. Spatial distribution
of pores in fly ash-based inorganic polymer gels visualised by Wood’s metal
intrusion. Microporous and Mesoporous Materials 126 (1–2), 32–39.
Lloyd, R.R., Provis, J.L., Van Deventer, J.S.J., 2010. Pore solution composition and
alkali diffusion in inorganic polymer cement. Cement and Concrete Research 40
(9), 1386–1392.
Lothenbach, B., Le Saout, G., Gallucci, E., Scrivener, K., 2008. Influence of limestone
on the hydration of Portland cements. Cement and Concrete Research 38 (6),
848–860.
McLellan, B.C., Williams, R.P., Lay, J., van Riessen, A., Corder, G.D., 2011. Costs and
carbon emissions for geopolymer pastes in comparison to ordinary Portland
cement. Journal of Cleaner Production 19 (9–10), 1080–1090.
Mehta, P.K., Monteiro, P.J.M., 2006. Concrete: Microstructure, Properties and
Materials, third ed. McGraw-Hill, New York, 659pp.
Nakashima, Y., Kamiya, S., 2007. Mathematica programs for the analysis of three-
dimensional pore connectivity and anisotropic tortuosity of porous rocks using
X-ray computed tomography image data. Journal of Nuclear Science and
Technology 44 (9), 1233–1247.
Nemati, K.M., 2000. Preserving microstructure of concrete under load using the
Wood’s metal technique. International Journal of Rock Mechanics and Mining
Sciences 37 (1–2), 133–142.
O’Connor, S.J., MacKenzie, K.J.D., 2010. Synthesis, characterisation and thermal
behaviour of lithium aluminosilicate inorganic polymers. Journal of Materials
Science 45 (14), 3707–3713.
Oh, J.E., Monteiro, P.J.M., Jun, S.S., Choi, S., Clark, S.M., 2010. The evolution of
strength and crystalline phases for alkali-activated ground blast furnace slag
and fly ash-based geopolymers. Cement and Concrete Research 40 (2), 189–196.
Olivia, M., Nikraz, H., Sarker, P., 2008. Improvements in the strength and water
penetrability of low calcium fly ash based geopolymer concrete. In: Uomoto, T.,
Nga, T.V. (Eds.), 3rd ACF International Conference – ACF/VCA 2008. Ho Chi Minh
City, Vietnam, pp. 384–391.
Ollivier, J.P., Maso, J.C., Bourdette, B., 1995. Interfacial transition zone in concrete.
Advanced Cement Based Materials 2 (1), 30–38.
Pacheco-Torgal, F., Castro-Gomes, J., Jalali, S., 2008a. Alkali-activated binders: a
review. Part 1. Historical background, terminology, reaction mechanisms and
hydration products. Construction and Building Materials 22 (7), 1305–1314.
Pacheco-Torgal, F., Castro-Gomes, J., Jalali, S., 2008b. Alkali-activated binders: a
review. Part 2. About materials and binders manufacture. Construction and
Building Materials 22 (7), 1315–1322.
Palomo, A., Banfill, P.F.G., Fernández-Jiménez, A., Swift, D.S., 2005. Properties of
alkali-activated fly ashes determined from rheological measurements.
Advances in Cement Research 17 (4), 143–151.
Plimer, I., 2009. Heaven and Earth. Global Warming: The Missing Science. Connor
Court Publishing, Ballan, Australia, 503pp.
Provis, J.L., Van Deventer, J.S.J., 2007. Geopolymerisation kinetics. 2. Reaction kinetic
modeling. Chemical Engineering Science 62 (9), 2318–2329.
Provis, J.L., Van Deventer, J.S.J. (Eds.), 2009. Geopolymers: Structures, Processing,
Properties and Industrial Applications. CRC Press, Cambridge UK, 54pp.
Provis, J.L., Lukey, G.C., Van Deventer, J.S.J., 2005. Do geopolymers actually contain
nanocrystalline zeolites? – a reexamination of existing results. Chemistry of
Materials 17 (12), 3075–3085.
Provis, J.L., Muntingh, Y., Lloyd, R.R., Xu, H., Keyte, L.M., Lorenzen, L., Krivenko, P.V.,
Van Deventer, J.S.J., 2007. Will geopolymers stand the test of time? Ceramic
Engineering and Science Proceedings 28 (9), 235–248.
Provis, J.L., Rose, V., Bernal, S.A., Van Deventer, J.S.J., 2009. High resolution nanoprobe
X-ray fluorescence characterization of heterogeneous calcium and heavy metal
distributions in alkali activated fly ash. Langmuir 25 (19), 11897–11904.
Provis, J.L., Duxson, P., Van Deventer, J.S.J., 2010. The role of particle technology in
developing sustainable construction materials. Advanced Powder Technology
21 (1), 2–7.
Provis, J.L., Rose, V., Winarski, R.P., Van Deventer, J.S.J., 2011a. Hard X-ray
nanotomography of amorphous aluminosilicate cements. Scripta Materialia
65 (4), 316–319.
Provis, J.L., Ismail, I., Myers, R.J., Rose, V., Van Deventer, J.S.J., 2011b. Characterising
the structure and permeability of alkali-activated binders. RILEM International
Conference on Advances in Construction Materials Through Science and
Engineering, Hong Kong, 4–8 September 2011.
Provis, J.L., Myers, R.J., White, C.E., Rose, V., Van Deventer, J.S.J., submitted for
publication. X-ray microtomography shows pore structure and tortuosity in
alkali-activated binders. Cement and Concrete Research.
Puertas, F., Palomo, A., Fernández-Jiménez, A., Izquierdo, J.D., Granizo, M.L., 2003.
Effect of superplasticisers on the behaviour and properties of alkaline cements.
Advances in Cement Research 15 (1), 23–28.
103
Rajaokarivony-Andriambololona, Z., Thomassin, J.H., Baillif, P., Touray, J.C., 1990.
Experimental hydration of two synthetic glassy blast furnace slags in water and
alkaline solutions (NaOH and KOH 0.1 N) at 40 °C: structure, composition and
origin of the hydrated layer. Journal of Materials Science 25 (5), 2399–2410.
Rees, C.A., Provis, J.L., Lukey, G.C., Van Deventer, J.S.J., 2007a. Attenuated total
reflectance Fourier transform infrared analysis of fly ash geopolymer gel aging.
Langmuir 23 (15), 8170–8179.
Rees, C.A., Provis, J.L., Lukey, G.C., Van Deventer, J.S.J., 2007b. In situ ATR-FTIR study of
the early stages of fly ash geopolymer gel formation. Langmuir 23 (17), 9076–9082.
Rees, C.A., Provis, J.L., Lukey, G.C., Van Deventer, J.S.J., 2008. The mechanism of
geopolymer gel formation investigated through seeded nucleation. Colloids and
Surfaces A 318 (1–3), 97–105.
Richardson, K., Steffen, W., Schellnhuber, H.J., Alcamo, J., Barker, T., Kammen, D.M.,
Leemans, R., Liverman, D., Munasinghe, M., Osman-Elasha, B., Stern, N., Waever,
O., 2009. Synthesis Report from Climate Change: Global Risks, Challenges &
Decisions. University of Copenhagen, 10–12 March 2009, Copenhagen,
Denmark <http://climatecongress.ku.dk/pdf/synthesisreport/>.
Rockström, J., Steffen, W., Noone, K., Persson, Å., Chapin III, F.S., Lambin, E.F., Lenton,
T.M., Scheffer, M., Folke, C., Schellnhuber, H.J., Nykvist, B., De Wit, C.A., Hughes,
T., Van der Leeuw, S., Rodhe, H., Sörlin, S., Snyder, P.K., Costanza, R., Svedin, U.,
Falkenmark, M., Karlberg, L., Corell, R.W., Fabry, V.J., Hansen, J., Walker, B.,
Liverman, D., Richardson, K., Crutzen, P., Foley, J.A., 2009. A safe operating space
for humanity. Nature 461, 472–475.
Rodríguez, E., Bernal, S., Mejía de Gutierrez, R., Puertas, F., 2008. Alternative
concrete based on alkali-activated slag. Materiales de Construcción 58 (291),
53–67.
Roy, D.M., Jiang, W., Silsbee, M.R., 2000. Chloride diffusion in ordinary, blended, and
alkali-activated cement pastes and its relation to other properties. Cement and
Concrete Research 30 (12), 1879–1884.
Sagoe-Crentsil, K., Brown, T., Yan, S., 2010. Medium to long term engineering
properties and performance of high-strength geopolymers for structural
applications. Advances in Science and Technology 69, 135–142.
Scrivener, K.L., Kirkpatrick, R.J., 2008. Innovation in use and research on
cementitious material. Cement and Concrete Research 38 (2), 128–136.
Shi, C., 1996. Strength, pore structure and permeability of alkali-activated slag
mortars. Cement and Concrete Research 26 (12), 1789–1799.
Shi, C., 2004. Effect of mixing proportions of concrete on its electrical conductivity
and the rapid chloride permeability test (ASTM C1202 or ASSHTO T277) results.
Cement and Concrete Research 34 (3), 537–545.
Shi, C., Xie, P., 1998. Interface between cement paste and quartz sand in alkali-
activated slag mortars. Cement and Concrete Research 28 (6), 887–896.
Shi, C., Krivenko, P.V., Roy, D.M., 2006. Alkali-Activated Cements and Concretes.
Taylor & Francis, Abingdon, UK, 376pp.
Shi, C., Fernández-Jiménez, A., Palomo, A., 2011. New cements for the 21st century:
the pursuit of an alternative to Portland cement. Cement and Concrete Research
41 (7), 750–763.
Shimoda, K., Tobu, Y., Kanehashi, K., Nemoto, T., Saito, K., 2008. Total understanding
of the local structures of an amorphous slag: perspective from multi-nuclear
(29Si, 27Al, 17O, 25Mg, and 43Ca) solid-state NMR. Journal of Non-Crystalline
Solids 354 (10–11), 1036–1043.
Sindhunata, Provis, J.L., Lukey, G.C., Xu, H., Van Deventer, J.S.J., 2008. Structural
evolution of fly ash-based geopolymers in alkaline environments. Industrial &
Engineering Chemistry Research 47 (9), 2991–2999.
Sofi, M., Van Deventer, J.S.J., Mendis, P.A., Lukey, G.C., 2007. Engineering properties
of inorganic polymer concretes (IPCs). Cement and Concrete Research 37 (2),
251–257.
Taylor, M., Tam, C., Gielen, D., 2006. Energy Efficiency and CO2 Emissions from the
Global Cement Industry. International Energy Agency.
Van Deventer, J.S.J., Provis, J.L., Duxson, P., Brice, D.G., 2010a. Chemical research and
climate change as drivers in the commercial adoption of alkali activated
materials. Waste and Biomass Valorization 1 (1), 145–155.
Van Deventer, J.S.J., Feng, D., Duxson, P., 2010b. Dry Mix Cement Composition,
Methods and Systems Involving Same. US Patent 7691,198 B2.
VicRoads, 2010. Standard Specification Section 703, General Concrete Paving.
<http://www.vicroads.vic.gov.au/Home/Moreinfoandservices/TendersAndSuppliers/
StandardDocuments.htm>.
Von Weizsäcker, E., Hargroves, K., Smith, M.H., Desha, C., Stasinopoulos, P., 2009.
Factor Five: Transforming the Global Economy Through 80% Improvements in
Resource Productivity. Earthscan, London, UK.
Weil, M., Dombrowski, K., Buchwald, A., 2009. Life-cycle analysis of geopolymers.
In: Provis, J.L., van Deventer, J.S.J. (Eds.), Geopolymers: Structures, Processing,
Properties and Industrial Applications. Woodhead Publishing, CRC Press,
Cambridge, UK, pp. 194–212.
White, C.E., Provis, J.L., Proffen, T., Van Deventer, J.S.J., 2010. The effects of
temperature on the local structure of metakaolin-based geopolymer binder: a
neutron pair distribution function investigation. Journal of the American
Ceramic Society 93 (10), 3486–3492.
White, C.E., Perander, L.M., Provis, J.L., Van Deventer, J.S.J., 2011. The use of XANES to
clarify issues related to bonding environments in metakaolin: a discussion of
the paper S. Sperinck et al., ‘Dehydroxylation of kaolinite to metakaolin – a
molecular dynamics study,’ J Mater Chem 21: 2118–2125; 2011. Journal of
Materials Chemistry 21 (19), 7007–7010.
White, C.E., Provis, J.L., Llobet, A., Proffen, T., Van Deventer, J.S.J., in pressa. Evolution
of local structure in geopolymer gels: an in-situ neutron pair distribution
function analysis. Journal of the American Ceramic Society. doi: 10.1111/j.1551-
2916.2011.04515.x.
104 J.S.J. van Deventer et al. / Minerals Engineering 29 (2012) 89–104
White, C.E., Provis, J.L., Proffen, T., Van Deventer, J.S.J., in pressb. Molecular
mechanisms responsible for the structural changes occurring during
geopolymerization: multiscale simulation. AIChE Journal. doi: 10.1002/
aic.12743.
Willis, K.L., Abell, A.B., Lange, D.A., 1998. Image-based characterization of cement
pore structure using Wood’s metal intrusion. Cement and Concrete Research 28
(12), 1695–1705.
Wong, H.H.C., Kwan, A.K.H., 2008. Packing density of cementitious materials:
measurement and modelling. Magazine of Concrete Research 60 (3), 165–175.
Wong, H.S., Buenfeld, N.R., Head, M.K., 2006. Estimating transport properties of
mortars using image analysis on backscattered electron images. Cement and
Concrete Research 36 (8), 1556–1566.
WWF-Lafarge Conservation Partnership, 2008. A Blueprint for a Climate Friendly
Cement Industry: How to Turn Around the Trend of Cement Related Emissions.
<http://assets.panda.org/downloads/cement_blueprint_climate_fullenglrep_lr.
pdf>.
Xu, X., Provis, J.L., Van Deventer, J.S.J., Krivenko, P.V., 2008. Characterization of aged
slag concretes. ACI Materials Journal 105 (2), 131–139.
Yip, C.K., Lukey, G.C., Van Deventer, J.S.J., 2005. The coexistence of geopolymeric gel
and calcium silicate hydrate at the early stage of alkaline activation. Cement
and Concrete Research 35 (9), 1688–1697.
Yip, C.K., Lukey, G.C., Provis, J.L., Van Deventer, J.S.J., 2008. Effect of calcium silicate
sources on geopolymerisation. Cement and Concrete Research 38 (4), 554–564.
Zhang, J.X., Sun, H.H., Wan, J.H., Yi, Z.L., 2009. Study on microstructure and
mechanical property of interfacial transition zone between limestone aggregate
and sialite paste. Construction and Building Materials 23 (11), 3393–3397.